P001 A PROSPECTIVE RANDOMIZED DOUBLE BLIND PLACEBO

P001 A PROSPECTIVE RANDOMIZED DOUBLE BLIND PLACEBO CONTROLLED TRIAL ON THE
EFFICACY OF ETHANOL CELIAC PLEXUS NEUROLYSIS (ECPN) IN PATIENTS WITH OPERABLE
PANCREATIC AND PERIAMPULLARY ADENOCARCINOMA (PPA) Harish Lavu, MD1, Harry B Lengel, BS1,
Naomi M Sell, BS1, Joseph A Baiocco, BS1, Eugene P Kennedy, MD1, Theresa P Yeo, PhD1, Sherry A Burrell,
PhD2, Jordan M Winter, MD1, Sarah Hegarty, MPhil1, Benjamin E Leiby, PhD1, Charles J Yeo, MD1; 1Thomas
Jefferson University, 2Rutgers University, Philadelphia, US
Background: Cancer­related pain is a common symptom in patients with pancreatic and periampullary
adenocarcinoma (PPA). Ethanol celiac plexus neurolysis (ECPN) has been shown to be effective in reducing
cancer­related pain in patients with locally advanced PPA. This study examined its efficacy in patients undergoing
PPA resection.
Study Design: 485 patients consented to participate in this prospective, randomized, double blind placebo controlled
trial that examined the effectiveness of ECPN in patients undergoing surgical treatment for PPA. Patients were
stratified at the time of surgery by the presence or absence of preoperative pain and by disease resectability. They
received either ECPN using 50% ethanol or 0.9% normal saline as a placebo control. Pain and quality of life (QOL)
were assessed using the Brief Pain Inventory and FACT­Hep Version 4 surveys respectively. The primary aim of
the study was to determine if ECPN is beneficial for patients with resectable PPA in reducing short and long­term
pain. Secondary outcomes included assessments of postoperative morbidity, QOL and overall survival.
Results: Data from 467 patients were available for analysis: 119 patients in the resectable/pain (R/P) group, 268
patients in the resectable/no pain (R/NP) group, 31 patients in the unresectable/pain (UR/P) group, and 49 patients
in the unresectable/no pain (UR/NP) group. The demographic characteristics were similar between the groups.
Postoperative complications were also similar and were not attributable to the ECPN. The primary outcome
measure, the percentage of PPA patients experiencing a worsening of pain compared to preoperative baseline for
resectable patients, was not different between the ethanol and the saline groups at the 12 month time point in either
the R/P stratum (22% vs 18%, RR 1.23 (0.34, 4.46)), or the R/NP stratum (37% vs 34%, RR 1.10 (0.67, 1.81)). On
multivariate analysis of a subset of only resected pancreatic ductal adenocarcinoma (PDA) patients, there was a
statistically significant reduction in pain at the 3 month time point following surgical resection in the R/P group,
which maintained its magnitude in absolute pain reduction to 9 months post­surgery, suggesting that surgical
resection of the malignancy alone decrements pain scores to a significant degree among patients with preoperative
pain. This finding was independent of ECPN. Resected patients with preoperative pain averaged lower QOL scores
(116) than those without preoperative pain (141) (p<0.01), however, within each stratum, there was no difference
between the ethanol and saline groups. When examining only the subset of patients with resected PDA, the overall
survival between the groups was nearly identical at 18.3 months for the ethanol group, and 17.6 months for the
saline group. Conclusions: In this study, the world’s largest randomized controlled trial evaluating ECPN in patients with operable
PPA, we have demonstrated a significant reduction in pain following surgical resection of the malignancy. However
the addition of ECPN did not synergize to affect a further reduction in pain compared to saline placebo, and in fact
may have masked the effect. Given these findings, we cannot recommend the use of ECPN to mitigate cancer
related pain in resectable PPA patients.
P003 DEVELOPING A CORE SET OF PATIENT­REPORTED OUTCOMES IN PANCREATIC CANCER: A
DELPHI SURVEY Arja Gerritsen1, Marc Jacobs1, Inge Henselmans1, Jons van Hattum1, Geert­Jan Creemers2,
Ignace de Hingh2, Miriam Koopman3, Quintus Molenaar3, Hanneke Wilmink1, Olivier Busch1, Marc Besselink1,
Hanneke van Laarhoven1, For the Dutch Pancreatic Cancer Group1; 1Academic Medical Center, Amsterdam, the
Netherlands, 2Catharina Hospital, Eindhoven, the Netherlands, 3University Medical Center Utrecht, Utrecht, the
Netherlands, Amsterdam, NL
Introduction
Patient­reported outcomes (PROs) are amongst the most relevant outcome measures in pancreatic cancer care and
research. However, it is unknown which out of the numerous PROs are most important in this setting. The aim of
this study was to identify a core set of PROs to be incorporated in a nationwide prospective multidisciplinary
pancreatic cancer registry.
Methods
We performed a two­round Delphi survey among 150 patients diagnosed with pancreatic or periampullary cancer
(treated either in curative or palliative setting) and 78 health care professionals (HCPs: surgeons, oncologists,
gastroenterologists, radiotherapists, nurses, and dietitians) in the Netherlands. In round 1, participants were invited
to rate the importance of 53 PRO­topics, which were extracted from 17 different PRO measures and grouped into
global domains, on a 1­9 Likert scale. Topics rated as very important (score 7­9) by the majority (≥80%) of curative
and/or palliative patients as well as HCPs were considered sufficiently important to be incorporated in the core set.
Topics not fulfilling these criteria in round 1, were re­presented to the participants in round 2 along with individual and
group feedback.
Results
A total of 97 patients (94%) in curative setting, 38 patients (81%) in palliative setting and 73 HCPs (94%) completed
both round 1 and 2. After both rounds, 17 PRO domains were included in the core set: general quality of life, general
health, physical ability, ability to work/do usual activities, medication, pancreatic enzyme replacement therapy,
appetite, weight changes, defecation, fatigue, negative feelings, positive feelings, coping, fear of recurrence,
relationship with partner/family, satisfaction with caregivers, and satisfaction with services and care organization.
Conclusion
This study provides a core set of PROs selected by patients and HCPs which may be incorporated in pancreatic
cancer care and research. Validation outside the Dutch context is recommended. P004 INITIATION OF AN ANESTHESIA PROTOCOL REDUCES INTRAOPERATIVE CRYSTALLOID AND
BLOOD ADMINISTRATION DURING PANCREATICODUODENECTOMY: A SINGLE CENTER
RETROSPECTIVE STUDY Nathan Bolton, MD1, William Conway, MD1, Shoichiro Tanaka, MD1, Kara Roncin,
BS2, James Hyatt2, John Bolton, MD1; 1Ochsner, 2Medical University of the Americas, New Orleans, US
Introduction
Recent evidence points to improved outcomes after complex GI surgery when fluid and blood administration is
minimized, often as part of an enhanced recovery after surgery program (ERAS). Due to the potential for significant
fluid shifts and blood loss, Pancreaticoduodenectomy (PD) patients can be given excess volume during surgery. Herein we report outcomes after PD when a restrictive protocol is followed.
Methods
Data was collected retrospectively on patients who underwent PD from 01/2008 until 09/2013. Two groups were
defined based on initiation of an anesthesia protocol in October 2010 which dictated conservative crystalloid and
blood administration and encouraged low­volume albumin use. Group A was defined as pre­protocol while group B
was defined as post­protocol. Basic demographics, preoperative data, and outcomes were analyzed. Primary
outcomes included Intraoperative crystalloid (IC), colloid and blood use. Additionally, data was collected on
intraoperative hypotension (IH), defined as systolic blood pressure below 80 mmHg, postoperative acute kidney
injury (AKI) as defined by the Kidney Disease: Improving Global Outcomes project (KDIGO) and hospital length of
stay.
Results
A total of 228 patient who underwent PD were analyzed. 102 underwent PD before protocol initiation (Group A) and
126 after initiation (Group B). Group characteristics were similar in regards to age (A 65.3 vs B 64.1 [p=0.4]), sex (A
M:F 54/46% vs B M:F 51/49% [p0.78]), BMI (A 27.4 vs B 27.1 [p=0.68]) and starting Hct (A 36.9 vs B 36.5
[p=0.59]). Pre­operative albumin (A 3.7 vs B 3.2 [p<0.01]) was lower in group B, and more patients in group B
underwent neoadjuvant therapies although this difference was non­significant (A 12% vs B 18% [p=0.28]). There
was a trend towards more vascular resections in group B (A 20% v B 30% [p=0.08]). IC use (A 7150ml vs B 4814ml
[p<0.01]), need for blood transfusion (A 47% vs B 23% [p<0.01]) and volume of blood used (A 1.2 units vs B 0.61
units [p<0.01]) were all significantly decreased in group B with a corresponding increase in the use of colloid fluids
(A 225ml vs B 612ml [p<0.01]). Despite this significant shift in volume administration, there were no differences
between groups in either frequency of intraoperative hypotension (A 63% vs B 68% [p=0.64]) or incidence of post­
operative AKI (A 0.06% vs B 0.13% [p=0.15]). Additionally, group B patients had a lower incidence of post­operative
pressor use (A 16% vs B 5% [p<0.05]), shorter ICU stay (A 2.23 days vs B 1.5 days [p<0.01]), and shorter overall
hospital stay (A 15.6 days vs B 11.9 days [p<0.01]).
Conclusions
A team, protocol­driven approach can minimize fluid and blood administration in PD patients. After initiation of our
protocol, crystalloid and blood use were significantly reduced, without an increase in IH or post­operative AKI. Overall outcomes appear to be improved as well, with a reduction in LOS noted after protocol initiation. Fluid and
blood management is likely an important component of an ERAS protocol for PD patients.
P005 MICRORNA­21 EXPRESSION AND OUTCOME IN RESECTABLE PANCREATIC DUCTAL
ADENOCARCINOMA ­ MULTICENTRE ANALYSIS Nigel B Jamieson, MRCS, PhD3, Asif Ali, MBChB2, Elisa
Giovannetti, MD, PhD1, Karin A Oien, FRCPath, PhD2, Fraser Duthie, FRCPath2, Euan J Dickson, FRCS, MD3,
Ross Carter, FRCS, MD3, Colin J McKay, MD, FRCS3; 3West of Scotland Pancreatic Centre, 2Wohlson Wohl
Cancer Research Centre, Institute of Cancer Sciences, MVLS, University of Glasgow, 1VU University Medical
Center, Amsterdam, The Netherlands, Glasgow, GB
Introduction
MicroRNAs (miRNA) have potential as diagnostic and prognostic biomarkers and as therapeutic targets in cancer.
We sought to establish the relationship between miR­21 expression and clinicopathologic parameters, including
prognosis, in resected pancreatic ductal adenocarcinoma (PDAC).
Methods
We assessed the potential clinical utility of miR­21 expression measured by In­situ hybridization (ISH) in 507
patients from three independent cohorts who underwent surgical resection for PDAC. ISH staining patterns were
scored semiquantively using a weighted histoscore method. A median cutoff of ≥ 45 was generated in the test
cohort and applied to the validation cohorts.
Results
Greater than the median miR­21 tumoural expression (histoscore ≥ 45, High) was associated with shorter survival as
compared to the low expression group (Histoscore <45) (14.7 (95%CI:12.4­17.0) Vs 26.5 (95%CI:20.4­32.6) months;
P < 0.0001). High epithelial miR­21 expression (histoscore ≥ 45) was found to be independently associated with a
poor prognosis in a multivariate analysis (Hazard Ratio (HR) 2.37, P = 0.001) along with the presence of nodal
metastases, high tumor grade, large tumour size , R1 margin status and no adjuvant chemotherapy. In two separate
validation cohorts including 69 and 249 patients respectively, miR­21 expression again independently correlated with
reduced overall survival (HR 2.03, P =0.006 and HR 2.58, P = 0.007 respectively).
We subsequently analyzed miR­21 expression and adjuvant chemotherapy allocation in the Test and Validation
cohort 2. High miR­21 expression was associated with improved outcome in patients receiving adjuvant
chemotherapy. After adjusting for the prognostic effect of mir­21 expression and chemotherapy, the interaction
variable (mir­21 X chemotherapy [≥3 cycles]) remained independently significant in both cohorts (HR = 0.516,
95%CI: 0.33–0.807, P = 0.004) and (HR = 0.329, 95%CI: 0.163–0.662, P = 0.002) respectively.
Finally we stratified the training cohort according to the presence of lymph node metastases and miR­21 expression.
The first group including those with low miR­21 expression and no lymph node involvement had an excellent
prognosis, (5­yr survival of 59%, median survival 90.3mths). Second, those patients with high miR­21 expression
and lymph node metastases had a very poor prognosis, (5­yr survival 5%, median survival 13.1mths). Third, the
remaining patients (with lymph node metastases or high miR­21 expression and no lymph node involvement) had an
intermediate prognosis, (5­yr survival 13% and median survival 24.0 months). This finding was replicated in both of
the validation cohorts.
Conclusion
We report the independent prognostic utility of miR­21 expression in three independent cohorts of resected PDAC
patients. Such molecular stratification may better delineate prognostic groups, aid in refinement of chemotherapeutic
strategies, better interpret past clinical trials, and facilitate future trial design. Furthermore assessment in pre­
operative EUS FNA and Procore samples has the potential to enhance the staging algorithm for patients with
borderline resectable disease.
P006 THE EFFICACY OF NEOADJUVANT THERAPY FOLLOWED BY SURGICAL RESECTION FOR
PATIENTS WITH BORDERLINE RESECTABLE PANCREATIC CANCER WITH ARTERY INVOLVEMENT
Hiroki Yamaue, Seiko Hirono, Manabu Kawai, Ken­ichi Okada, Motoki Miyazawa, Atsushi Shimizu, Yuji Kitahata;
Second Department of Surgery, Wakayama Medical University, Wakayama, JP
Background: It has been still controversial to perform surgical resection with borderline resectable pancreatic cancer
with artery involvement (BR­A), because an aggressive surgery leads to high morbidity and mortality with low R0
rate for the BR­A patients. In this study, we evaluated whether or not neoadjuvant therapy followed by surgical
resection improves survival benefits for the patients with BR­A pancreatic cancer.
Methods: There were 138 patients with BR­A among 330 pancreatic cancer patients underwent surgical resection at
Wakayama Medical University Hospital. Our definition of BR­A included the tumor with abutment of celiac axis
added to NCCN guideline definition. We compared clinicopathological factors between 38 BR­A patients with
neoadjuvant therapy followed by surgical resection and 100 BR­A patients with upfront surgery to evaluate the
clinical impacts of neoadjuvant therapy.
Results: The overall survival (OS) of BR­A patients was significantly shorter than that of the patients with borderline
resectbale pancreatic cancer with portal vein/ superior mesenteric vein (PV/SMV) involvement (n=76) and resectable
pancreatic cancer (n=105) who underwent surgical resection (median OS: 13.6 vs. 20.6 months, P<0.001). The OS
of BR­A patient with neoadjuvant therapy followed by surgical resection was significantly longer than those with
upfront surgery (median OS: 20.2 vs. 12.9 months, P=0.047). Multivariate analysis showed that older age (P=0.027,
odds ratio: 1.60), pathological PV/SMV invasion (P=0.031, odds ratio: 1.60), moderated or poor differentiated tumor
(P=0.008, odds ratio: 1.86), positive lymph node ratio ?0.1 (P=0.018, odds ratio: 1.60), and no postoperative
adjuvant chemotherapy (P<0.001, odds ratio: 2.80) were independent poor prognostic factors for BR­A patients who
underwent surgical resection.
Conclusions: Neoadjuvant treatment might bring the clinical benefits for BR­A patients, and it has been strongly
needed to develop the appropriate regimen of neoadjuvant therapy and postoperative adjuvant therapy for longer
survival in the patients with BR­A pancreatic cancer.
P007 AFTER NEOADJUVANT RADIATION THERAPY AN R1 RESECTION DOES NOT DECREASE
SURVIVAL IN PANCREATIC DUCTAL ADENOCARCINOMA Shadi Razmdjou, MD, Bl Collins, C Fernandez­del
Castillo, Ts Hong, Jy Wo, F Sabbatino, V Villani, D Dias Santos, Al Warshaw, Kd Lillemoe, Cr Ferrone;
Massachusetts General Hospital, Cambridge, US
OBJECTIVE(S): Neoadjuvant radiation therapy (XRT) with protons or photons is often utilized in patients with locally
advanced, borderline or resectable pancreatic ductal adenocarcinoma (PDAC). The aim of this study was to
evaluate the impact of an R1 resection on overall survival (OS).
METHODS: Clinicopathologic data was collected for PDAC patients who underwent XRT at our institution between
1/2001­4/20014. All patients underwent either photon (50.4 Gy) or short course proton beam therapy (25 Gy) and
concurrent chemotherapy with either 5­ Fluorouracil or Capecitabine with or without Hydroxychloroquine. None of
the patients had additional neoadjuvant chemotherapy.
RESULTS: Of the 105 patients 51% were female with a median age of 66 yrs . Protons were administered to 67
patients (63.8%) and photons to 38 patients (36.2%). An R0 resection was achieved in 93 patients. Twelve patients
(11%) underwent an R1 resection, of these 75% received proton and 25% photon therapy. There was no significant
difference between the R0 vs. R1 resection groups regarding gender, age, stage, grade, tumor size, and time to
recurrence. The median survival time was 28.9 months (range: 1.8­116.3 months). The log­rank test comparing the
Kaplan­Meier Survival Curves demonstrates no significant difference between an R0 and R1 resection (p=0.358),
see figure 1.
CONCLUSIONS: In patients who have received neoadjuvant radiation therapy there is no difference in overall
survival if patients undergo an R0 vs. R1 resection.
P008 LONG­TERM PATIENT­REPORTED SYMPTOMS AND QUALITY OF LIFE OUTCOMES ARE
FAVORABLE FOLLOWING RESECTION OF PANCREATIC NEOPLASMS Hop S Tran Cao, MD, Maria Q
Petzel, RD, Nathan H Parker, BS, Joe S Liles, MD, Michael Kim, MD, Jeffrey E Lee, MD, Thomas A Aloia, MD,
Claudius Conrad, MD, Jean N Vauthey, MD, Jason B Fleming, MD, Matthew H Katz, MD; U.T. MD Anderson
Cancer Center, Houston, US
Background: Patient­reported symptoms and quality of life (QOL) are critically important outcome metrics following
cancer operations but are poorly described following pancreatic resection for neoplasms. We sought to evaluate the
long­term QOL and surgery­related symptoms associated with pancreatectomy and to identify factors that may
influence them.
Methods: As part of a broader survivorship project, we conducted a cross­sectional survey of QOL (Functional
Assessment of Cancer Therapy­Hepatobiliary Questionnaire) and psychosocial distress (Hospital Anxiety and
Depression Scale) among patients with ductal (PDAC) or periampullary adenocarcinoma (NPAC) or pancreatic
neuroendocrine tumors (PNET) who were free of disease at least 6 months following pancreatectomy.
Results: Of 348 eligible patients, 232 (66.7%) participated at a median of 50 months (range, 8 ­ 238 months)
following pancreaticoduodenectomy or total pancreatectomy (PD/TP) (n=169), or distal pancreatectomy, central
pancreatectomy, or others (DP/Other) (n=63). Overall QOL was influenced by race and pancreatectomy type but not
histology; PD/TP survivors reported better QOL, including improved functional, emotional, and social well­being, and
lower symptom severity scores than DP/Other survivors (p<0.05 for all). Compared to DP/Other survivors, PD/TP
patients experienced more frequent problems with abdominal cramping and diarrhea, but less frequent problems with
poor appetite, constipation, fatigue, anxiety and depression (p<0.05 for all). Conclusion: In this, the largest study quantifying self­reported, long­term surgery­related symptoms and QOL
following pancreatectomy, patients generally reported favorable QOL but clinically significant gastrointestinal and
psychosocial symptoms were reported in nearly 20% of patients long after surgery. These critical data are needed
to optimize preoperative decision­making, design surveillance strategies, and identify therapeutic targets in the
survivorship period.
P009 THE INCIDENCE AND MANAGEMENT OF DELAYED GASTRIC EMPTYING FOLLOWING
PANCREATICODUODENECTOMY: A LARGE SINGLE­INSTITUTION ANALYSIS Joshua D Eisenberg, Janae
A Romeo, Ernest L Rosato, MD, Harish Lavu, MD, Charles J Yeo, MD, Jordan M Winter, MD; Department of
Surgery, Thomas Jefferson University Hospital, Philadelphia, US
ISGPS
DGE
Risk Factor
Risk Factors for DGE (Multivariate)
Odds
Primary
P­value
Risk Factor
Ratio
DGE
OddsRatio P­value
Abdominal infection
5.50
<0.001*
Abdominal infection
3.15
<0.001*
Gender (M)
1.92
0.007*
Smoking history
2.04
0.008*
Smoking history
1.75
0.033*
Malignant tumor
1.54
0.347
Periampullary
adenocarcinoma
1.66
0.041*
Periampullary
adenocarcinoma
1.35
0.421
Soft gland texture
1.32
0.263
Gender (M)
1.29
0.314
Estimated blood loss
1.00
0.649
Background: Delayed gastric emptying (DGE) is a common complication after pancreaticoduodenectomy (PD), yet
the evaluation, management, and impact remain incompletely understood. The International Study Group of
Pancreatic Surgery (ISGPS) in 2007 defined a three­tiered grading system to standardize studies of DGE.
Methods: Data were collected on 721 consecutive patients undergoing PD between October 2006 and May 2012 at
a high­volume academic medical center. Patients with DGE were retrospectively categorized according to the
ISGPS criteria (grades A, B, or C), as well as a modified grading system (referred to herein as primary DGE), in
which DGE was diagnosed if gastric function was documented to be abnormal by fluoroscopy, or if gastric
symptoms were not attributable to another complication (e.g., abdominal infection). Predictors of DGE and the
impact of DGE on outcomes were determined.
Results: Using ISGPS criteria, DGE was diagnosed in 140 (19.4%) patients, including 78 (56%) grade A, 36 (26%)
grade B, and 26 (19%) grade C. DGE was associated with an increased rate of abdominal infections (55% vs. 19%,
p<0.001), including pancreatic fistula (34% vs. 10%, p<0.001) and abscess (24% vs. 8%, p<0.001), as well as
higher rates of re­hospitalization (29% vs. 13%, p<0.001) and length­of­stay (median 12.5 days vs. 7 days, p<0.001).
Primary DGE occurred in 12.2% of the cohort (66% grade A; 19% grade B and 15% grade C). In a multivariate
logistic regression model (Table), abdominal infection (OR 5.5, p<0.001), male gender (OR 1.92, p=0.007), smoking
history (OR 1.75 p=0.033), and periampullary adenocarcinoma (OR 1.66, p=0.041) were statistically significant risk
factors for DGE. Statistically significant predictors of primary DGE included abdominal infection (OR 3.15, p<0.001)
and smoking history (OR 2.04, p=0.008). Tests and interventions performed on patients with grade B or C ISGPS
DGE included total parenteral nutrition (87%), pro­kinetic therapy (80%), upper­GI fluoroscopy (80%), endoscopy
(23%), and gastrostomy tube placement (17%). Median total hospital charges increased by more than $10,000 with
each severity grade of ISGPS DGE (p<0.001).
Conclusions: DGE is a morbid complication after PD and it is associated with a substantial increase in hospital
cost. In many instances, DGE is likely secondary to abdominal infection or other non­gastric complications, and
interventions aimed at preventing these complications may be the most effective strategy towards preventing DGE.
P010 THE OLDEST­OLD AND HOSPITAL­LEVEL RESOURCE USE AFTER
PANCREATICODUODENECTOMY AT HIGH VOLUME HOSPITALS Russell C Langan, MD, Chaoyi Zheng, MS,
Katherine Harris, PhD, Richard Verstraete, RN, Waddah B Al­Refaie, MD, Lynt B Johnson, MD, MBA; Georgetown
University Hospital, Washington, US
Introduction: Studies examining post­pancreaticoduodenectomy (PD) operative outcomes in patients older than 80­
years have found higher complications, longer length of stay (LOS) and higher mortality. However, there is limited
data reporting on the hospital resources consumed in caring for the oldest­old. We examined the use of PD­relevant
hospital resources in patients treated in high­volume­hospitals (HVH) participating in the University HealthSystem
Consortium (UHC).
Methods: Using the UHC database, we identified 210 U.S. hospitals performing ≥ 12 PD/year between 2010 and
2014. We compared mortality, complications, ICU­use, TPN­use, blood transfusions, LOS, readmissions and direct
costs by increasing age groups. Index hospitals performed a total of 12,766 PDs (< 70 years n=8,564, 70­79 years
n=3,302, ≥ 80 years n=900). We used linear regression models with and without adjusting for covariates to assess
the impact of older age. Hospital means were weighted based on age­specific procedure volume.
Results: Compared to the youngest cohort, those ≥ 80­years experienced more cardiopulmonary, genitourinary and
infectious complications, more blood transfusions, greater TPN use, longer LOS and higher direct costs (Table 1).
However, no statistical differences were found between patient’s ≥ 80­years and those 70 – 79 years with respect to
the administration of blood products or TPN and the direct cost of PD. Additionally, the oldest old experienced
fewer readmissions and had equivalent ICU­use and mortality rates to both younger cohorts.
Conclusion: Our findings suggest the
ability to deliver quality pancreatic surgical care to an aging population with minimal increases in resource utilization.
With growing pressure to control and reduce hospital costs, it is imperative to identify, understand, and modify
factors that contribute to elevated resource use both within the hospital and post­discharge. Additionally, as the
number of octogenarians undergoing PD continues to grow, the impact of this technically complex procedure on
other important cancer care metrics including patient reported outcomes and quality of life requires further
assessment.
P011 MICRORNA­145 TARGETS MUC13 AND SUPPRESSES GROWTH AND INVASION OF PANCREATIC
CANCER Sheema Khan, PhD1, Mara C Ebeling, BS2, Mohd S Zaman, PhD1, Mohammed Sikander, PhD1, Murali
M Yallapu, PhD3, Ashley Yacoubian4, Stephen W Behrman, MD5, Nadeem Zafar, MD4, Deepak Kumar, PhD6, Paul
A Thompson, PhD2, Meena Jaggi, PhD3, Subhash C Chauhan, PhD3; 1Univ. of Tennessee, Dept. of
Pharmaceutical Sciences, 2Sanford Research, Cancer Biology Research Center, 3University of Tennessee Health
Science Center, 4Univ. of Tennessee, Dept. of Pathology, 5Univ. of Tennessee, Dept. of Surgery, 6Univ. of the
District of Columbia, Dept. of Biological and Environmental Sciences, Memphis, US
Background: Mucins are a group of glycoproteins that under normal conditions protect mucosal surfaces, but have
tumor promoting properties when over­expressed in neoplastic conditions. Recent studies report that mucin 13
(MUC13) is a transmembrane mucin that is highly involved in PanCa progression suggesting its potential use as a
diagnostic and therapeutic target in PanCa. MicroRNAs (miRNAs) are small, noncoding RNAs that are highly
associated with cancer initiation and progression via their ability to affect expression of proteins regulating cell
proliferation. We interrogated whether miRNAs might regulate MUC13 expression in PanCa. The interplay, if
present, between miRNA and MUC13 was assessed relative to the progression from normal pancreas tissue to
PanIN and invasive PanCa as well as the efficacy of MUC13 inhibition on PanCa progression and gemcitabine
(GEM) sensitization.
Methods: In silico analysis predicted MUC13 binding sites for the miR­145 that was confirmed using reporter gene
assay. MUC13 expressing PanCa cell lines (HPAF­II and Capan­1) were used for the study. Western blotting and
immunofluorescence techniques were used to investigate effects of miR­145 on MUC13 expression and on
additional proteins affected by MUC13 expression. Functional studies demonstrating the effects of miR­145 on cell
proliferation, colony formation, cell migration, and cell invasion assays were performed. The in vivo therapeutic
efficacy of miRNA restitution and MUC13 expression was investigated using HPAF­II xenograft mice models. miR­
145 and MUC13 expression in human pancreatic tissues and xenograft mouse tissues were analyzed by in situ
hybridization and immunohistochemistry, respectively. The expression of MUC13 and miR­145 in PanCa cell lines
was investigated by q­PCR analysis.
Results: miR­145 directly targets the 3′ untranslated region of MUC13 and thus downregulates MUC13 protein
expression in cells as observed by reporter gene assay. In situ hybridization of miR­145 suggests its expression
predominantly in normal human pancreatic tissues and the early PanCa precursor lesion PanIN I but then is
progressively suppressed over the course of development from PanIN II/ III to late stage poorly differentiated PanCa
(see figure). miR­145 expression inversely correlated with MUC13 expression in PanCa cells as well as in human
tumor tissue. Transfection of miR­145 inhibited cell proliferation and invasion of PanCa cells as observed through
MTS and matrigel invasion assay. miR­145 reduced MUC13 and its effector oncoproteins, HER2, P­AKT, PAK1 and
increased tumor suppressor, p53. Similar results were found when MUC13 was specifically inhibited by shRNA
directed at MUC13. Additionally, miR­145 enhanced GEM sensitivity in GEM resistant AsPC­1 cells accompanied
by reduced cellular invasion and downregulation of MUC13 and HER2. Intratumoral injections of miR­145 in
xenograft mice inhibited tumor growth via suppression of MUC13 and its downstream target, HER2 as depicted by
immunohistochemical staining.
Conclusions: 1) miR­145 downregulates MUC13 expression and acts as a tumor suppressor miRNA in PanCa. 2)
miRNA­145 is progressively suppressed in the progression from PanIN to invasive carcinoma. 3) miR­145 restitution
inhibits pancreatic xenograft tumor growth and enhances GEM sensitivity in GEM resistant PanCa cells 4) It may be
a successful therapeutic strategy for PanCa progression and growth by its inhibitory effects on MUC13 expression.
P012 NICOTINE REDUCES SURVIVAL VIA AUGMENTATION OF PARACRINE HGF­MET SIGNALING IN
THE PANCREATIC CANCER MICROENVIRONMENT Daniel Delitto, MD, Dongyu Zhang, PhD, Song Han, PhD,
Brian S Black, BS, Andrea E Knowlton, PhD, Adrian C Vlada, MD, George A Sarosi, MD, Kevin E Behrns, MD,
Ryan M Thomas, MD, Xiaomin Lu, PhD, Chen Liu, MD, PhD, Thomas J George, MD, Steven J Hughes, MD,
Shannon M Wallet, PhD, Jose G Trevino, MD; University of Florida, Gainesville, US
Smoking is an established risk factor for the development of pancreatic adenocarcinoma (PC). However, the
relationship between smoking and PC tumor biology is incompletely defined. We report reduced overall survival (OS)
in PC patients who continued smoking after surgical resection with curative intention (HR 1.93; P = .040). We
further demonstrate augmented paracrine signaling via the hepatocyte growth factor (HGF)/c­Met pathway as a
result of nicotine exposure in the PC microenvironment. Specifically, HGF, secreted by patient­derived PC tumor
associated stroma (TAS), activates the c­Met receptor in PC cells. This paracrine activation subsequently leads to
downstream induction of inhibitor of differentiation­1 (Id1) in PC cells, previously established as a mediator of
chemoresistance. Further delineation of the signaling pathway demonstrates HGF­induced Id1 expression is
abrogated by silencing of c­Met or pharmacologic inhibition. In patient­derived PC xenografts, nicotine treatment
augmented tumor growth and metastasis; tumor lysates from nicotine­treated mice demonstrated elevated HGF
expression and phospho­Met levels. Additionally, patients with high intratumoral phospho­Met levels exhibited
reduced overall survival compared to those with low phospho­Met levels (Median OS 6.1 vs. 15.2 months,
respectively; P = .028). Taken together, our data reveal that nicotine promotes the progression of PC via a
microenvironment­dependent, paracrine signaling mechanism.
P013 TGFSS/EGFR CROSS­TALK MODULATES EMT PROCESS AND MIGRATION IN 3D TISSUE­
ENGENEERED MODEL OF PANCREATIC DUCTAL ADENOCARCINOMA Niccola Funel, PhD1, Claudio Ricci,
PhD1, Edwige Pugliesi, Dr2, Luca E Pollina, MD3, Fabio Caniglia, MD4, Serena Danti, Ing2, Ugo Boggi, Prof4,
Daniela Campani, Prof2; 1Department of Translational Research and New Technologies in Medicine and Surgery,
2Department of Surgical, Medical, Molecular Pathology and Emergency Medicine, University of Pisa, 3Division of
Surgical Pathology, Hospital of Pisa, Italy, 4Division of General and Transplants Surgery, University of Pisa, Italy,
Pisa, IT
Introduction: Pancreatic ductal adenocarcinoma (PDAC) is at umor with poor prognosis and few treatment options
are available. Indeed, preclinical models close to PDAC are needed. At present, growth factors play a pivotal role on
the aggressiveness and migration. In particular, TGF­β protein expression represents the master factor involved in
epithelial mesenchymal transition (EMT) phenomenon. Aim of this study was to investigate, EMT and cell migration
in 3D tissue­engeneered model of PDAC through TGF­β protein expression.
Methods: To create 3D cellular models, were fabricated a cylindrical polymers using polyvinyl alcohol/gelatin
(PVA/G). Primary PDAC cells were seeded on the sterile scaffolds at a density of 1×105 cells/mm3 and cultured for
4 different times (2, 5, 8 and 15 days). Six scaffolds for each time were performed. In 3 of them (group A) the
medium was replaced as needs, while in other 3 scaffold (group B), the medium was not replaced. A total of 24
formalin­fixed paraffin embedded (FFPE) 3D model were obtained. FFPE 3D model analyses included: viability
(AlamarBlue assay); morphology and histology (H&E staining); marker expressions by immunohistochemistry (IHC:
PanCk, EGFR, TGF­β, MMP9, Actin and Desmin); migration by computerized analyses of cellular distributions. Ten
measures for each FFPE block were done (software AnalySISb, Olympus, Italy). ELISA test (EGF concentration in
the medium). Finally, we used an inhibitor (NSC­631570 a mixture of alkaloids; [5μM]) of TGF­β and MMP9). Data
were analyzed using ANOVA and Student’s tests. A p value < 0.05 was considered significant.
Results: Significant differences comparing Group A vs Group B models were observed after 8 and 15 days in terms
of metabolic activity (50.9% vs 38.9% and 46.6% vs 57.2%, respectively; p <0.001). In group A cell migration
correlated with culture time (r2=0.9726; p=0.0138) and after 15 days the cells were more distributed in the border of
scaffold with respect to 2 days time, which were arranged throughout the volume (mean difference 74.6%;
p=0.0020). An increased number of mesenchymal­shape cells was observed (80% vs 10%, p <0.01). Cells in group
A showed high expression of PanCk, EGFR, TGF­β and MMP9 throughout the experiment duration, While the group
B cells lost TGF­β and MMP9 expression from day 8 onwards. Constitutive expression of TGF­β and was confirmed
by IF. NSC­631570 drug reduced TGF­β staining, mesenchymal phenotype and cellular migration. Finally, we found
a statistically significant difference in protein expression of EGF in the culture medium of 3D models at the time of
15 days (group A vs group B; 11:17 vs 6.97 pg / ml).
Discussion: 3D model could be representative of EMT process for PDAC. Recently is reported that cross­talk
between TGF­βRs and EGFR in pancreatic cancer where the autocrine secretions of their growth factors induce a
series of processes including cell proliferation, tumor growth and EMT. Autocrine secretion of EGF could control the
balance between proliferation and EMT process. NSC­631570 has been shown to regulate EMT and hold the
epithelial phenotype . These experiments may be use to suggest treatment anti TGF­β, and EGFR in patients with
PDAC.
P014 THE BIOLOGICAL BASIS OF HISTOPATHOLOGICALLY CONFIRMED PORTAL VENOUS INVASION
IN PANCREATIC HEAD CANCER H Lapshyn, MD1, P Bronsert, MD2, D Bausch, MD1, F Makowiec, MD3, U A
Wittel, MD3, M Werner, MD2, T Keck, MD1, U F Wellner, MD1; 1Clinic of Surgery, UKSH Campus Lübeck, Lübeck,
Germany, 2Institute of Pathology, University Medical Center Freiburg, Freiburg, Germany, 3Clinic for General and
Visceral Surgery, University Medical Center Freiburg, Freiburg, Germany, Lübeck, DE
Background:
Pancreaticoduodenectomy (PD) with portal venous resection (PVR) for pancreatic ductal adenocarcinoma (PDAC) is
performed routinely in case of adhesion to the portal vein. Several studies showed that survival in patients with PVR
is not limited compared to patients without PVR. We have previously demonstrated that true histopathologically
confirmed tumor invasion of the venous vessel wall (PVI) is associated with poor survival in patients with PVR. The
aim of the present study was to assess the tumor biology underlying PVI.
Material and Methods:
Retrospective analysis was performed on the basis of a prospectively maintained database and archived paraffin­
embedded formalin fixed tissue. Statistical analysis was performed with MedCalc software. The following biologic
factors were assessed for correlation with PVI: standard histopathologic factors (TNM stage, lymph node ratio,
microscopic lymphangiosis and hemangiosis, perineural invasion and tumor grading), standard tissue markers
(Cytokeratin 7, Cytokeratin 20 and Cdx2 staining pattern), morphologic activation of cancer associated fibroblasts,
tumor budding and epithelial­mesenchymal transition (EMT) markers (immunohistochemical E­Cadherin, ß­Catenin,
Vimentin and ZEB1 staining).
Results:
N=86 cases of PD with PVR for PDAC and sufficient tissue for re­assessment were identified. Histopathological re­
review disclosed true cancerous portal vein wall invasion (PVI) in 39 resection specimen. PVI correlated positively
with nuclear Cdx2 expression, shift from membranous to cytoplasmic E­Cadherin expression in cancer cells of
tumor buds, microvessel hemangiosis as well as increased cancer associated fibroblast activation.
Conclusion:
Histopathological portal venous tumor invasion in pancreatic ductal adenocarcinoma is associated with an
aggressive cancer biology characterized by features of epithelial­mesenchymal transition, nuclear Cdx2 expression,
stromal fibroblast activation and microvessel hemangiosis.
P015 TUMOR VOLUME RATIO (VTR) CORRELATES WITH METASTATIC LIMPH NODE RATIO (LNR) IN
PANCREATIC DUCTAL ADENOCARCINOMA Niccola Funel, PhD1, Linda Barbarello, MD4, Luca E Pollina,
MD2, Vittorio Perrone, MD4, Daniela Campani, Prof3, Ugo Boggi, Prof4; 1Department of Translational Medicine and
Surgery, University of Pisa, 4Division of General and Transplants Surgery, University of Pisa, Italy, 2Division of
Surgical Pathology, Hospital of Pisa, Italy, 3Division of Surgical Pathology, University of Pisa, Italy, Pisa, IT
Introduction: Tumor diameter and lymph node ratio (LNR) are among the main prognostic factors in resected
pancreatic ductal adenocarcinoma (PDAC). LNR, in particular, is emerging as an important prognostic factor.
Unfortunately, we have no preoperative factors that reliably identify patients with high LNR. We herein describe the
prognostic implications of tumor volume ratio (TVR) in resected PDAC and its relationships with LNR.
Methods: TVR was defined as the ratio between tumor volume and specimen volume. To define TVR both the
volume of the tumor and the volume of the resected specimen were assessed in cm3. LNR was defined as the ratio
between examined lymph nodes and metastatic lymph nodes. All other conventional staging parameters, including
residual tumor, were also assessed. Data were analyzed using ANOVA and Student’s tests, as appropriate. A p
value < 0.05 was considered significant.
Results: Specimens from 173 histologically proven PDAC were examined. According to the recommendations of the
last UICC staging manual, all primary tumors were staged as pT3. No metastasis were found in 29 patients
(pT3N0M0; stage IIA), metastasis in regional lymph nodes were identified in 125 patients (pT3N1M0; stage IIB), and
metastasis in extraregional lymph nodes were detected in 19 patients (pT3N1M1; stage IV). The mean number of
examined lymph node was 35.8 (7­108), with no difference across staging groups (stage IIA: 31.2) (stage IIB: 38.2)
(stage IV: 46.7). In stage IIB patients, mean LNR, was 17.3% (1.7%­77.7%), and mean TVR was 19.3%
(0.8%­72.9%). A good linear regression was observed between TVR and LNR (r2= 0.9942; p < 0.0001). Mean TVR
was lower in stage IIA patients as compared with stage II B patients (8.6% vs 19.3%; p= 0.0211). Further, mean
TVR showed a correlation with both N and M status (p=0.0125).
Conclusions: LNR is a well­known prognostic factor in resected PDAC. Patients with high LNR do poorly even after
seemingly radical resection. These patients could be best managed by neoadjuvant chemotherapy or chemo­
radiation therapy. Unfortunately, we are not able to reliably identify patients with high LNR until specimen
analysis. We have shown a strong correlation between TVR and LNR in resected PDAC. Should radiologic
determination of TVR lead to the same results, we would be able to identify preoperatively patients with anticipated
high LNR among those with otherwise immediately resectable PDAC.
P016 ACCURACY OF PREOPERATIVE IMAGING FOR VASCULAR INVOLVEMENT IN LOCALLY
ADVANCED, BORDERLINE RESECTABLE PANCREATIC ADENOCARCINOMA FOLLOWING
NEOADJUVANT CHEMOTHERAPY Jesse Clanton, J B Rose, Adnan Alseidi, Thomas Biehl, Scott Helton, Flavio
Rocha; Virginia Mason Medical Center, Seattle, US
Background: Radiographic imaging with arterial/venous phase, thin­slice computed tomography (MDCT) and
endoscopic ultrasound (EUS) are utilized for assessment and staging of locally advanced, borderline resectable
pancreatic adenocarcinoma (BR­LAPD). Neoadjuvant therapy followed by pancreaticoduodenctomy with vascular
resection is typically required for BR­LAPD. However, little is known about the correlation of MDCT and EUS
findings with need for vascular resection and true pathologic invasion.
Methods: A retrospective review of a prospectively­maintained database of consecutive patients with BR­LAPD
treated with neoadjuvant chemotherapy at our institution between 2011 and 2014 was performed. MDCT and EUS
results were compared to operative and pathology reports regarding vascular involvement. Results: A cohort of 33 patients underwent staging by MDCT and EUS followed by neoadjuvant chemotherapy and
successful resection. All patients were reported to have venous involvement on CT, EUS, or both. Based on
preoperative imaging and operative findings, a venous resection was performed in 15/33 (45.5%) patients, but
histological evidence of vascular invasion was only noted in 6/33 (18.2%) of pathologic specimens. These 6 patients
were noted to have vascular invasion on both CT and EUS. CT demonstrated vascular involvement without EUS
confirmation in 4/33 (12%) patients, while EUS demonstrated vascular involvement without CT confirmation in 3/33
(9%) patients. None of these patients had true pathologic vascular invasion. Conclusion: CT and EUS can be complementary modalities to detect BR­LAPD, but may overestimate the actual
incidence of venous involvement. This may be due to inherent limitations of these techniques versus a true
downstaging effect of neoadjuvant chemotherapy.
P017 AGE BIAS AND UNDER­TREATMENT IN OCTOGENARIANS WITH PANCREATIC CANCER Jonathan
C King, MD, Jennifer Steve, BS, Mazen S Zenati, MD, MPH, PhD, Sharon B Winters, MS, CTR, David L Bartlett,
MD, Amer Zureikat, MD, Herbert J Zeh III, MD, Melissa E Hogg, MD; UPMC Division of Surgical Oncology,
Pittsburgh, US
Introduction: Morbidity and mortality following pancreatic resection is at an all­time low and chemotherapeutic
options for pancreatic cancer (PC) are growing, yet there is still reluctance to treat elderly patients. We aimed to
examine the reason for failure to treat and analyze outcome in octogenarians with PC.
Methods: We performed retrospective chart review for patients ≥80 years old from 2005­2013. Demographics, tumor
characteristics, treatment, reason for lack of treatment, Charleston Comorbidity Index (CCI) and survival were
analyzed.
Results: 446 octogenarians were analyzed comprising 18% of all patients. Mean age was 83.9±3.3, 58.8% female.
Overall 44% received no treatment. Octogenarians with operable tumors (stage 1=35 [7.8%], 2a=100 [22.4%],
2b=120 [26.9%]) had surgery 39% of the time (compare to 58% of all­comers) with the smallest proportion
undergoing surgery for stage 1 (17.1% vs stage 2b 54.2%; p<0.001). Higher stage patients were more likely to
undergo surgery (OR 2.02 95%CI 1.34­3.03; p=0.001). Increasing age was a predictor of not receiving surgery (OR
0.82 95%CI 0.74­0.91; p<0.001) whereas CCI was not. The most common reason for no surgery was
‘contraindicated by comorbidity’ (29.8%) despite similar CCI for stage and treatment. Only 19.6% of patients with
resectable disease refused surgery of which 66% were female (p<0.01), in 11.4% the reason for not undergoing
surgery was unknown. Median overall survival was better in the surgical group 15.9 vs 5.6 mo in the nonsurgical
group (p<0.001). Advanced stage patients (stage 3=55 [12.3%], 4=136 [30.5%]) had similarly low treatment rates:
chemo stage 3=36.4%, stage 4=34.6% with better survival seen in treated patients (7.0±5.3 vs 2.3±2.7 mo; p<0.01).
Younger patients were more likely to undergo chemotherapy (OR 0.81 95%CI 0.72­0.92) but CCI was not related
(OR 0.99 95%CI 0.67­1.47).
Conclusion: There is significant deviation from expected treatment for octogenarians with PC. While no correlation
existed between CCI and treatment, age correlated with therapy for nearly all stages and few patients refused
therapy. Chronological age, not comorbidity, may drive recommendations for treatment in elderly patients.
P018 ANALYSIS OF GLYCEMIA IN PATIENTS UNDERGOING BYPASS SURGERY AND
PANCREATODUODENECTOMY DUE TO ADENOCARCINOMA OF THE PANCREATIC HEAD Mariusz
Seweryn, Katarzyna Kusnierz, MD, PhD, Aleksandra Kolarczyk­Haczyk, Weronika Bulska, Pawel Lampe,
Professor, MD; Department of Gastrointestinal Surgery , Medical University of Silesia, Katowice Poland.,
Sosnowiec, PL
Introduction:
The presence of adenocarcinoma of the pancreas may affect glucose tolerance and the development of diabetes is
related to the amount of pancreatic parenchyma. Pancreatoduodenectomy decreases amount of pancreatic
parenchyma and can trigger diabetes.
The aim of study was assessment of changes of glycemia by patients undergoing pancreatoduodenectomy (PD) and
bypass surgery (BS) propter adenocarcinoma of the pancreas.
Methods:
We perform prospective analysis of patients with pancreatic head tumor qualified to PD among I.2011­V.2012.
Within 262 patients with pancreatic tumor, 103 had adenocarcinoma. Patients were divided into two groups
according to surgery. First group of patients underwent PD­46 patients, second BS­57. We analyzed pre­ and
postoperative glycaemia, presence of pre­ and postoperative diabetes, diabetes treatment, preoperative BMI and
loss of weight.
Results:
There was significant difference (p<0.05) in terms of preoperative diabetes. In PD­group 43,5%(20) patients had
preoperative diabetes, in BS­group 38,6%(22) had diabetes and 17,5%(10) impaired fasting glucose. Preoperative
glycaemia in patients with diabetes was 114mg/dl in PD­group,138mg/dl in BS­group, and after operation 174mg/dl,
152mg/dl respectively(p<0.05). There was significant difference (p<0,05) in preoperative diabetes treatment: in PD­
group 65%(13) used metformin, 35%(7)only diet, in BS­group 50%(11)used insulin, 50%(11)diet. There was no
significant difference in preoperative BMI (p>0,05). Postoperative diabetes occurred in 45,7%(21)patients in PD­
group and 52,6%(30)in BS­group (p<0.05). In postoperative diabetes treatment 76,2%(16)patients in PD­group used
insulin, in BS­group 82,8%(24)used insulin(p>0.05).
Conclusion:
There was not relationship between the resection of the head of pancreas and diabetes. Patients undergoing PD had
often diabetes before surgery, but the average glycaemia were lower and not required using insulin in comparison to
BS­group. Postoperative glycaemia was higher in PD­group with preoperative diabetes and required a change in
treatment with insulin. There was significant increase of amount of patients with diabetes in bypass surgery.
P019 CANCER OF THE DISTAL BILE DUCT ­ A MULTICENTER RETROSPECTIVE ANALYSIS G Seifert,
MD1, S Zach, MD2, H Lapshyn, MD3, D Bausch, MD3, F Makowiec, MD1, U A Wittel, MD1, U T Hopt, MD1, T
Keck, MD3, F Rückert, MD2, U F Wellner, MD3; 1Clinic for General and Visceral Surgery, University Medical
Center Freiburg, Freiburg, Germany, 2Clinic of Surgery, University Medicine Mannheim, Mannheim, Germany,
3Clinic of Surgery, UKSH Campus Lübeck, Lübeck, Germany, Lübeck, DE
Background
Distal dile duct adenocarcinoma (DBDAC) is rare and usually not diagnosed before resection. Data on perioperative
outcome and survival is scarce. The aim of this study was retrospective analysis in a large patient cohort.
Methods
Retrospective exploratory data analysis was performed on the basis of prospectively maintained databases from
three high­volume academic centers for hepatopancreatic surgery. Data collection and analysis was performed with
MedCalc 14.8.1. software.
Results
From 1993 to 2013, n=111 patients (73 male, 38 female, median age 68 years) resected for DBDAC with
perioperative and long­term follow­up could be identified. Operations performed were pylorus preserving
pancreatoduodenectomy (PPPD, n=97), Whipple procedure (n=11) and total pancreatectomy (n=3), including 16
portal venous resections (PVR). Median operation time was 402 min, with 35% intraoperative blood transfusion
requirement. Perioperative grade B/C (ISGPS definition) pancreatic fistula, delayed gastric emptying,
postpancreatectomy hemorrhage and reoperation rates were 27%, 14%, 22% and 19%, respectively, with a
perioperative mortality of 7%. Overall median and 5­year survival was 28 months and 29%. BMI, operation time,
intraoperative transfusion, PVR, reoperation and surgical margin status were significant predictors of survival in
univariate analysis. In a multivariate Cox proportional hazards model, preoperative CA19.9, surgical margin status,
tumor grade and reoperation were independent predictors of survival.
Conclusion
DBDAC is a rare entity and associated with a high operative risk in terms of pancreatic fistula, bleeding and
perioperative mortality. Survival figures after resection are relatively favorable and influenced by the surgical
parameters resection margin status and reoperation rate, as well as the biologic factors tumor grading and CA19.9
level.
P020 CHARACTERISTIC OF THE OPERATIONS AND COMPLICATIONS IN PATIENTS WITH
NEUROENDOCRINE TUMOR OF THE PANCREAS Mariusz Seweryn, Katarzyna Kusnierz, MD, PhD,
Aleksandra Kolarczyk­Haczyk, Natalia Lampe, Pawel Lampe, Professor, MD; Department of Gastrointestinal
Surgery , Medical University of Silesia, Katowice Poland., Sosnowiec, PL
Introduction
Neuroendocrine tumors are rare neoplasms, characterized by beyond symptoms connected with tumor mass effect
and infiltration of the organs and vessels, also symptoms associated with secretion of hormones. Due to the new
possibilities of therapy in patients with advanced neuroendocrine tumor, complete resection of the primary tumor is
really important.
Aim of this study was to characterize operations and early complications in patients with neuroendocrine tumor of
the pancreatoduodenal field.
Methods
Retrospective analysis of patients operated due to tumor pancreatic­duodenal field in 2011­2014 in the Department
of Gastrointestinal Surgery, Medical University of Silesia in Katowice was performed. The analysis involved type of
surgery, age, tumor size, presence of liver metastases, postoperative hospitalization time and the complications
rate.
Results
Among 843 patients with a tumor of the pancreatic­duodenal field, in 53 patients histopathological neuroendocrine
tumor was confirmed. The average age of patients: 51,38 +/­13,08 years. Location of the tumor of the pancreatic­
duodenal field: 52.8%(28) patients ­ head of the pancreas, 47.2%(25) patients ­ body and cauda of the pancreas. In
96.2%(51) cases were performed radical tumor resections. The average postoperative hospitalization time was 13,23
+/­9,04 The average diameter of the tumor: 5,01+/­4,3cm. The percentage of liver metastases was 54,7%(29). The
postoperative course in 30,2%(16) cases was with complication. (in group of patients with tumor in head of
pancreas: 32,1%(9); in group of patients with tumor in body and cauda of the pancreas ­ 28,0%(7) respectively. In
4 (7,5%) cases, the death occurred during the postoperative time.
Conclusion
Early results of the treatment and the percentage of complications is satisfactory and should therefore strive to
achieve radical surgery in patients with neuroendocrine tumors of the pancreas, which is associated with early and
precise diagnostic procedure.
Complications are significantly more common in patients with tumor of head of panceras, which is why they should
be carefully monitored after surgery.
P021 COMPARISON OF THE CLINICAL CASES OF PATIENTS UNDERGOING BYPASS SURGERY AND
PANCREATODUODENECTOMY DUE TO PANCREATIC HEAD TUMOR Mariusz Seweryn, Katarzyna
Kusnierz, MD, PhD, Aleksandra Kolarczyk­Haczyk, Weronika Bulska, Pawel Lampe, Professor, MD; Department of
Gastrointestinal Surgery , Medical University of Silesia, Katowice Poland., Sosnowiec, PL
Introduction:
Pancreatoduodenectomy is the method of radical treatment of tumors of head of pancreas. Often the local
advancement of the tumor makes it impossible to perform radical surgery. In cases, where the radical resection is
not possible, the alternative option is the palliative bypass surgery. They do not eliminate the problem of the
presence of the tumor, but provides the patency of the gastrointestinal and biliary tract.The aim of study was to
compare the clinical cases of patients who underwent pancreatoduodenctomy(PD) and bypass surgery(BS).
Methods:
We perform data analysis of prospectively obtained, standardized medical interview which included 157 patients
qualified for surgery propter tumor of head of pancreas among I.2012 – V.2013. Histopathological diagnosis included
malignant neoplastic tumors.The patients were classified into PD­group and BS­group. There were 76 patients after
PD and 81 after BS. The following parameters were analyzed: gender, age, BMI, time of diagnostic procedures, the
number of diagnostic medical centers, the first symptoms of the tumor, major symptoms and postoperative course.
Results:
In BS­group dominated men, in PD­group women(p<0,05). There was found significant difference in terms of the age
of patients: average age in BS­group was 63+/­11 years, in PD­group 52+/­12. Time of diagnostic procedures: BS­
group: 10,5+/­3 months, PD­group: 7,5+/­3. First symptoms of the tumor: BS­group: jaundice, PD­group: pain.
Patients from both groups were being diagnosed average in 4 medical centers. In terms of the frequency of major
symptoms, both groups had pain and weight loss in 82%; lack of appetite in 41%; diarrhea, constipation and
vomiting in 34%. A higher number of previous treatments was in BS­group. The postoperative course was similar in
both groups.
Conclusions:
The analysis showed differences in terms of gender, age, time of diagnosis and the first symptom of tumor of head
of pancreas. Patients who underwent pancreatoduodenctomy can be characterized as young women, with shorter
time of diagnostic procedures and with pain as the first symptom.
P022 COMPLIANCE WITH SENDAI CRITERIA: A SINGLE INSTITUTION EXPERIENCE James C Padussis,
MD, Jennifer Steve, BS, Stephanie Novak, BS, Melissa E Hogg, MD, Amer H Zureikat, MD, Herbert J Zeh III, MD;
University of Pittsburgh Medical Center, Pittsburgh, US
Introduction
With the increase in frequency of cross­sectional imaging, non­inflammatory cystic lesions of the pancreas are
becoming increasingly diagnosed. Intraductal papillary mucinous neoplasm (IPMN) and mucinous cystic neoplasm
(MCN) are two distinct entities, which carry malignant potential. In 2004, during the Eleventh Congress of the
International Association of Pancreatology, held in Sendai, Japan, a consensus meeting was held to propose
guidelines for the diagnosis and treatment of IPMN and MCN. The “International Consensus Guidelines for
Management of IPMN and MCN of the Pancreas” was published in 2006 and have been variably adopted by medical
professionals. No publication however, has looked at compliance with Sendai criteria at an institutional level. We
report on how well the consensus guidelines are being followed at the University of Pittsburgh Medical Center three
years after publication of the consensus article. Methods
A retrospective chart review was performed to examine all patients at the University of Pittsburgh Medical
Center who had a pancreatic cyst identified on imaging during the 2009 calendar year. Patients with imaging
qualities suggestive of a pseudocyst were excluded and patients who had follow­up imaging of a cyst identified prior
to 2009 were excluded. We then examined the total number of cysts suspicious for IPMN or MCN and how they
were managed after diagnosis. We examined the percentage of patients with suspicious cysts >1cm who were
referred to gastroenterology. We also examined the percentage of patients with imaging or endoscopic evaluation
suspicious for MCN, main duct IPMN or side branch IPMN >3cm who were referred to a surgeon. Results
We identified 257 patients who had a pancreatic cyst identified by computed tomography (CT), magnetic
resonance imaging (MRI), or ultrasound (US) at the University of Pittsburgh Medical Center in 2009. Of those
cysts, 42 had imaging characteristics suggestive of a pancreatic pseudocyst. Patients that had imaging
characteristics suspicious for IPMN or MCN totaled 215, with an average cyst size of 1.61 cm. Of those patients,
78 (38%) were excluded, as their cyst was identified on imaging prior to 2009. The remaining 137 patients (64%)
had a newly diagnosed pancreatic cyst in 2009 with imaging characteristics suggestive of IPMN or MCN. The
number of suspicious cysts with size >1cm totaled 109 (80%). Of these patients, 83 (76%) were referred to
gastroenterology for further diagnostic evaluation with endoscopic ultrasound (EUS). Twenty­four patients had a
cyst suspicious for MCN, main duct IPMN or side­branch IPMN with size greater then 3 cm. Of these patients, 19
(79%) were referred to a surgeon and 17 (89%) proceeded to resection.
Conclusion
The University of Pittsburgh Medical Center has a high degree of compliance with Sendai criteria three years
after the landmark publication. Seventy­six percent of newly identified suspicious pancreatic cysts >1 cm in size
were appropriately referred to a gastroenterologist. Furthermore, gastroenterologists referred appropriate lesions to
surgery 79% of the time. Surgeons proceeded to pancreatic resection according to Sendai criteria 89% of the time. P023 ENGLISH AND SPANISH LANGUAGE READABILITY OF ONLINE PATIENT RESOURCES FOR
PANCREATIC CANCER Manuel Castillo­Angeles, MD, Alessandra Storino, MD, Ammara A Watkins, MD,
Christina R Vargas, MD, Jennifer F Tseng, MD, Mark P Callery, MD, A. James Moser, MD, Tara S Kent, MD; Beth
Israel Deaconess Medical Center, Boston, US
Background:
Nearly 50% of cancer patients use the internet as a source of health information. The National Institutes of Health
and American Medical Association have recommended patient health information be written at a sixth­grade reading
level. However, prior publications have reported that online patient resources are often written at a higher grade
level. We aimed to evaluate English and Spanish language websites for readability of pancreatic cancer information
in order to determine whether available information is at a reading level likely to be understood by readers.
Methods:
The top ten websites for “pancreas cancer” in English and Spanish found through the most visited search engine in
the United States were identified via a patient­simulated manner. Each initial website “hit”, as well as each subtopic
and its first generation link was designated as an article and was evaluated for reading level on its own and grouped
with the overarching website. A total of 122 articles in English and 76 in Spanish were available to be assessed
using ten readability analyses. Spanish­language websites native to 3 representative Spanish­language countries
were identified. Average readability scores were determined for all articles as well as by website.
Results: Overall mean reading level across US websites was 12.4±1.42 (English) and 10.5±0.95 (Spanish). Mean readability
by website ranged from 9.5 to 15.2 in English, and from 8.6 to 12.3 in US Spanish. Argentine, Chilean, and Mexican
websites had mean reading levels 9.7, 10.9, and 11 respectively. Spanish websites in each country were
significantly easier to read than those in English (Figure 1).
Conclusion:
Commonly searched online pancreatic cancer resources in English and Spanish exceed the recommended sixth
grade reading level. Spanish­language websites, US­based or not, were significantly easier to read than English­
language sites, though still above the recommended 6th grade level. With the growing focus on patient­centered
care, attention should be paid to ensuring availability of information at a level understandable by the general public.
More easily understandable materials may facilitate patient participation in shared decision­making.
P024 FEASIBILITY OF PANCREATECTOMY AFTER HIGH DOSE PROTON THERAPY FOR INITIALLY
UNRESECTABLE PANCREATIC CANCER. Romaine C Nichols, MD1, Christopher G Morris1, Debashish Bose,
MD2, Steven J Hughes, MD3, John A Stauffer, MD4, Scott A Celinski5, Robert C Martin6, Elizabeth A Johnson4,
Robert A Zaiden7, Michael S Rutenberg1; 1UF Health Proton Therapy Institute, 2UF Health Cancer Center ­ Orlando
Health, 3UF Health Cancer Center ­ Gainesville, 4Mayo Clinic ­ Jacksonville, 5Baylor University, 6University of
Louisville, 7Baptist Hospital ­ Jacksonville, Jacksonville, US
Purpose: Review surgical outcomes for patients undergoing pancreatectomy after 59.40 Cobalt Gray Equivalent
(CGE) proton radiotherapy with concomitant capecitabine (1000mg PO BID) for initially unresectable pancreatic
adenocarcinoma.
Methods and Materials: From 4/20/10 to 9/30/13 15 patients with initially unresectable pancreatic cancer were
treated with full dose proton therapy with concomitant capecitabine. All patients received 59.40CGE to gross
disease, One patient also received 50.40CGE to the high risk nodal targets. There were no treatment interruptions
and no chemotherapy dose reductions. 6 of these patients achieved a radiographic response sufficient to justify
surgical exploration. 1 was identified as having intraperitoneal dissemination at the time of surgery and the planned
pancreatectomy was aborted. 5 patients underwent resection. Procedures included: laparoscopic standard
pancreaticoduodenectomy (3); open pyloris sparing pancreaticoduodenectomy (1); and open distal pancreatectomy
with irreversible electroporation (IRE) of a pancreatic head mass (1).
Results: Median patient age is 60 years (range 51 to 67). Median duration of surgery was 419 minutes (range 290 to
484); Median estimated blood loss was 850cc (range 300 to 2000); Median ICU stay was 1 day (range 0 to 2);
Median hospital stay was 10 days (range 5 to14); 3 patients were readmitted to hospital within 30 days after
discharge for: wound infection (1); delayed gastric emptying (1); ischemic gastritis (1). 2 patients underwent R0
resections and demonstrated minimal residual disease in the final pathology specimen. 1 patient – after negative
pancreatic head biopsies – underwent IRE followed by distal pancreatectomy with no tumor seen in the specimen. 2
patients underwent R2 resections. Only one patient demonstrated ultimate local progression at the primary site.
Median survival for the 5 resected patients is 24 (range 10 to 30) months.
Conclusions: Pancreatic resection for patients with initially unresectable cancers is feasible after high dose proton
radiotherapy with a high rate of local control and median survival of 24 months.
P025 IMPACT OF PREOPERATIVE DIABETES AND DEGREE OF HYPERGLYCEMIA ON PROGNOSIS OF
PATIENTS WITH RESECTED PANCREATIC DUCTAL ADENOCARCINOMA Yoo­Seok Yoon, Woohyung Lee,
Ho­Seong Han, Jai Young Cho; Seoul National University Bundang Hospital, Seongnam­si, KR
Background: The impact of preoperative DM on prognosis of resected PDAC has been controversial. In most of
reports on this subject, the effect of presence of DM on prognosis of resected PDAC was only analyzed, but the
prognostic effect of degree of hyperglycemia related to degree of glycemic contol was not evaluated. The purpose of
this study was to evaluate the oncologic outcomes of patients with resected PDAC according to the presence of
preoperative DM and the degree of hyperglycemia using glycosylated hemoglobin (HbA1c).
Method: Of 167 patients with pathologically proven PDAC who underwent pancreatectomy in Seoul National
University Bundang Hospital between September 2003 and June 2012, 142 patients with R0 or R1 resection were
selected for this study. The patients were divided into three groups according to the presence of DM and level of
HbA1c: non­DM (n = 69), DM with HbA1c < 8.8% (n = 48), and DM with HbA1c ≥ 8.8 % (n = 25). We compared the
survival outcomes of the three groups with retrospective analysis using a prospectively collected database. Results: There was no significant difference in age, presence of jaundice, preoperative biliary drainage, tumor site,
tumor size, cancer stage, postoperative complication and adjuvant chemotherapy among three groups. After a mean
follow­up of 20 months, 3­year overall survival (OS) and disease­free survival (DFS) rates were similar between non­
DM and DM groups. However, DM with HbA1c ≥ 9.0 % group showed a significantly lower 3­year OS (22.3%) and
DFS (0%) compared with non­DM (34.3%/30.2%) and DM with HbA1c <9.0 % (40.2%/34.2%) groups
(p=0.028/0.036). The multivariate analysis revealed that DM with HbA1c ≥ 9.0 % (P = 0.007; RR = 2.531; 95% CI
1.287 – 4.978) and presence of angiolymphatic invasion (P = 0.039; RR = 1.842; 95 % CI 1.032–3.289) were
independent prognostic factors for OS. Conclusion: This study reveals that severely uncontrolled hyperglycemia rather than the presence of preoperative
DM negatively affects the survival outcome in patients with resected PDAC. Further large­scale studies are required
to draw the concrete conclusion about the impact of preoperative severe hyperglycemia on the prognosis of resected
PDA.
P026 INACCURACY OF PRE­OPERATIVE SIZE DETERMINATION IN PANCREATIC NEUROENDOCRINE
TUMORS: A RETROSPECTIVE STUDY ON 199 PATIENTS G Butturini1, A Malpaga1, H Impellizzeri1, G
Marchegiani1, M Miotto1, R Manfredi2, G Zamboni2, P Capelli3, S Cingarlini4, L Landoni1, R Salvia1, C Bassi1; 1The
Pancreas Institute Surgical Unit, 2The Pancreas Institute Radiology Unit, 3The Pancreas Institute Pathology Unit,
4The Pancreas Institute Oncology Unit, Verona, IT
Background: Preoperative radiologic size of pancreatic neuroendocrine tumors (PNET) is a crucial parameter in
selection of appropriate treatment especially for incidentally discovered small lesions (< 20 mm). The current
guideline bases the therapeutic recommendations on the pathologic size.
Aim: Evaluate the discrepancy between radiologic size (RS) and pathologic size (PS) of PNET Materials and
Methods: Patients resected for PNET between January 2004 and November 2014 at our Institution were retrieved
from a prospectively maintained electronic database and analyzed retrospectively. We excluded functioning and
cystic PNET and R2 resections. RS is defined as the mean of the largest diameters measured on CT, EUS and
MRI, and PS as the largest diameter of the tumor on the histological report of the surgical specimen.
Results: Study populations consisted of 199 patients (97 females/102 males), with a median age of 56 years (23­
83). Overall the RS (32.7 ± 22.8 mm) and PS (31.1 ± 23.4 mm) did not significantly differ (p = 0,477). In the
subgroup analysis, the size difference was statistically significant for tumor size of less then 30 mm (p=0.0006). In
this last group (n=133), RS overestimated PS in 66% (n=88) and RS underestimated PS in 21% (n=28).
Furthermore, we selected patients with PS<20 mm but with a RS>20 mm (n=23), and 13 of these were incidentally
found, sporadic PNET.
Conclusions: Radiologic preoperative imaging overestimates the size of tumor especially in small PNET (< 30
mm). The overestimation could be of paramount importance in the management of sporadic incidentally diagnosed
small (<20 mm) PNET.
P027 ISOLATED POSITIVE PERITONEAL CYTOLOGY IS ASSOCIATED WITH BETTER SURVIVAL THAN
GROSS METASTATIC DISEASE IN ADVANCED PANCREATIC CANCER Stephen Y Oh, MBBS, BSc,
FRACP, Alicia M Edwards, MBA, Margaret T Mandelson, PhD, Thomas Biehl, MD, FACS, Scott Helton, MD,
FACS, Flavio G Rocha, MD, FACS, Vincent Picozzi, MD, Adnan Alseidi, MD, EdM, FACS; Digestive Disease
Institute at Virginia Mason Medical Center, Seattle, US
Introduction/Background
Positive peritoneal cytology (PPC) in patients with pancreatic cancer indicates metastatic disease. However, it is
not known whether patients with PPC as their only site of distant metastasis have similar disease progression and
survival to other stage IV patients. We evaluated the natural history of patients with PPC and compared their
outcome to patients with gross metastasis. Methods
From 2003 to 2013, all patients with stage IV disease receiving oncologic therapy at our institution were identified
using our pancreatic cancer database. Inclusion criteria for PPC cohort were as follows; 1. No radiographic evidence
of metastasis 2. No surgical evidence of metastasis by laparoscopy and 3. PPC via peritoneal washings. The gross
metastasis cohort were patients with biopsy­proven or radiographic evidence of metastasis. Clinicopathologic
information was obtained via medical records. Disease progression was defined as radiographic or pathological
evidence of primary tumor growth (local) or new metastasis (systemic) necessitating switch of oncologic therapy.
Results
There were 44 patients with PPC and 265 with gross metastasis. Baseline characteristics between two groups were
similar except for significantly higher median Ca 19­9 (440 vs. 1904 IU/ml, p<0.0001) and worse functional status
(ECOG score ≤1 98 vs. 87%, p=0.02) in gross metastasis. Tumor resectability status in the PPC cohort prior to
peritoneal washings are as follows: 7% (3) resectable; 18% (8) borderline resectable; 75% (33) unresectable.
Treatment details in the PPC cohort are as follows: 66% (29) chemotherapy only following diagnosis; 32% (14)
additional consolidative chemoradiation; 2% (1) resection. Patients received consolidative chemoradiation if there
was no systemic progression with initial chemotherapy. Median time from diagnosis to initiation of chemoradiation
was 9.6 months (6.5 – 19.5). The one patient who underwent resection died after 6.5 months following diagnosis
with systemic progression. Three patients were excluded from subsequent analysis as progression status could not
be determined (death within 2 months of diagnosis in 2, no repeat imaging in 1). In the remaining 41 patients, 80%
(33) experienced disease progression. 20% (8) without disease progression had a follow­up ranging from 6.4 to 18.1
months. Initial sites of progression were: 32% (13) local; 27% (11) systemic; 22% (9) combined; median time to
primary progression 7.6 months (0.6 – 27.3). Systemic progression ultimately occurred in 73% (30): median time to
progression 10.1 months (0.6 – 34.9); 13 liver metastasis; 16 peritoneal disease; 9 lung, bone and/or adrenal.
Survival was significantly better in patients with PPC compared with gross metastasis: median survival 13.9 vs. 9.5
months, p<0.0001; 1­year survival 67 vs. 31%, p<0.001; 2­year survival 27 vs. 12%, p=0.01.
Discussion/Conclusion
Patients with advanced pancreatic cancer with PPC had significantly better survival when compared to those with
gross metastasis. However, the majority of patients with PPC ultimately experienced systemic progression,
predominantly involving the liver and peritoneum. Further study is warranted to determine whether PPC is a
consistent marker of systemic disease and such selected patients would benefit from additional alternative
therapies.
P028 MALIGNANT PROGRESSION IN INTRADUCTAL PAPILLARY MUCINOUS NEOPLASMS OF THE
PANCREAS: OUTCOME OF INITIALLY SELECTED FOR RESECTION OR PRIMARY SURVEILLANCE
Takuya Sakoda, MD, Yoshiaki Murakami, Kenichiro Uemura, Yasushi Hashimoto, Naru Kondo, Naoya Nakagawa,
Kazuhide Urabe, Hayato Sasaki, Hiroki Ohge, Taijiro Sueda; Dep.of Surgery, Appli. Life Sciences Institute of
Biomedical & Health Sciences, Hiroshima University, Hiroshima, JP
BACKGROUND: There is international consensus that most branch duct IPMNs (BD­IPMN) harbor a low risk of
progression to malignancy and can be safely managed with observation alone. However, the natural history of
patients undergoing primary surveillance is poorly understood and there is still ongoing controversy about surgical
indication of BD­IPMN. This study describes the natural history of IPMN in patients initially selected for resection or
primary surveillance with respect to risk factors of malignant transformation.
METHODS: All consecutive patients who underwent surgery for IPMN between April 1990 and August 2014 were
included. Low versus high oncologic risk was stratified prospectively and compared with histopathological features.
RESULTS: A total of 162 patients underwent pancreatic resection for IPMN. Based on the preoperative imaging, 38
(23%) patients had main duct involved IPMN, 56 (35%) had mixed­type, and 68 (42%) had suspected BD­IPMNs.
Initial resection was selected for 131 patients (81%) and invasive disease was identified in 38% (50/131). Primary
surveillance was selected for 31 patients (19%) and ultimately underwent resection: 6 mixed­type and 25 BD­
IPMNs. Median follow­up for the primary surveillance was 5.5 years. Of the 31 patients for primary surveillance, 24
(77%) initially stratified as low­risk IPMN (guideline negative) developed a new indication for pancreatic resection:
development of intramural nodule (n=22) and extension to main pancreatic duct (n=16), new onset of symptom
(n=3), and invasive lesions developed in a region remote from the index IPMN (n=2). On histopathology, 18 patients
had low­grade, 2 had intermediate­grade, 8 had high­grade dysplasia, and 3 had invasive disease associated with
IPMN on histopathology. Overall, invasive disease was identified in 23% (7/31) for primary surveillance, of which 4
(13%) patients developed distinct invasive ductal adenocarcinomas in a region separate from the index cyst,
representing 57% (4 of 7) with invasive disease during surveillance. Among the 24 patients initially determined to be
low­risk (guideline negative), 5 patients developed high­grade dysplasia (4.0­10.4 years after initiating surveillance)
and only one patient developed invasive cancer associated with IPMN over 10 years after initially diagnosed as BD­
IPMN (10.5 yeas).
CONCLUSIONS: Progression to invasive cancer during primary surveillance for low­risk BD­IPMN was rare.
Diagnostic, operative, and surveillance strategies for IPMN should consider increased risk not only to the index cyst
but also the entire gland for developing distinct pancreatic cancer.
P029 PANCREATIC CANCER PATIENTS WITH LYMPH NODE INVOLVEMENT BY DIRECT TUMOR
EXTENSION HAVE SIMILAR SURVIVAL TO THOSE WITH NODE­NEGATIVE DISEASE Jennifer L Williams,
MD1, Andrew H Nguyen, MD2, Matthew Rochefort, MD2, James S Tomlinson, MD2, Oscar J Hines, MD2, Howard
A Reber, MD2, Timothy R Donahue, MD2; 1Department of Surgery, Harbor­UCLA Medical Center, 2Department of
Surgery, David Geffen School of Medicine at UCLA, Los Angeles, US
Introduction: Lymph node (LN) involvement is a well­known poor prognostic factor in patients with pancreatic ductal
carcinoma (PDAC). However, PDAC often invades into peri­pancreatic tissues, and there have been conflicting
results on the prognostic significance of the mechanism of LN involvement: “direct” tumor invasion versus
“metastatic” disease.
Methods: Clinicopathologic and survival records from all patients who underwent resection for PDAC from 1990 to
2014 at a single­institution were reviewed from a prospectively maintained database. Results: Of the 381 total patients, most (n= 335, 87.9%) underwent pancreaticoduodenectomy. One­hundred and
ninety­one patients (50.1%) were female, and median age was 66 years. There was tumor growth outside of the
pancreas (AJCC T stage 3 or 4) in 290 (76.2%) patients, and margins were microscopically positive in 79 (20.7%).
Overall, 237 (62.2%) had node­positive disease: (i) 218 (92.0%) by “metastatic” involvement, (ii) 14 (5.9%) by “direct”
tumor extension, and (iii) 5 (2.1%) by a mix of “metastatic” and “direct”. The median number of LNs pathologically
examined was significantly lower in the “direct” versus “metastatic” group (11 vs. 18, P=0.015); however, there was
no difference in the median number of involved LNs (1 vs. 2, P=0.64). “Direct” involvement ranged from 1 – 7 (mean:
2.0) LNs per patient. Similar to LN involvement, there were no significant differences in other clinicopathologic
factors associated with PDAC survival between “metastatic” and “direct” LN patients, including tumor size, tumor
grade, margin status, lymphovascular invasion, perineural invasion, and neoadjuvant or adjuvant therapy. The
median and 5­year overall survivals for the whole cohort were 30.1 months and 27.3%. As compared to overall
survival in patients with LN negative disease (median: 40.4 months, 5­year 37.1%), those with: (i) any number or
mechanism of LNs involved was significantly shorter (median: 26.1 months, 5­year: 19.3%; P < 0.001), yet (ii)
“direct” LN extension was similar (median 48.1 months, 5­year survival 29.2%; P=0.719). Furthermore, there was no
survival benefit to having only 1 “metastatic” LN involved (median: 22.8 months, 5­year: 18.3%) as compared to 2 or
more (median: 26.7 months, 5­year: 18.3%) (P =0.821). Conclusions: These results indicate that the mechanism of LN involvement matters in PDAC. Patients with LNs
involved by direct extension have a similar survival to those with node negative disease.
P030 PANCREATICODUODENECTOMY IN THE SETTING OF INTESTINAL MALROTATION Canaan Baer,
MD1, Randall Zuckerman, MD2, Thomas Biehl, MD1, Scott Helton, MD1, Flavio G Rocha, MD1; 1Virginia Mason
Medical Center, 2St. Vincent's Medical Center, Seattle, US
INTRODUCTION: Intestinal malrotation (IM) is a developmental anomaly resulting from a failure of the embryonic
midgut to complete its rotation around the superior mesenteric artery (SMA). Although typically recognized in the
pediatric population, IM in adults with periampullary malignancies present particular challenges for the pancreatic
surgeon. Herein, we present our experience with patients with IM requiring pancreaticoduodenectomy.
METHODS: Retrospective review of 5 patients with IM undergoing pancreaticoduodenectomy for periampullary
lesions at two institutions. Clinical, radiographic, and pathologic information was obtained from the medical record.
RESULTS: Five patients presented with resectable periampullary neoplasms (2 pancreatic adenocarcinoma, 1
ampullary, 1 duodenal, and 1 intraductal papillary mucinous neoplasm). On preoperative imaging they were found to
have intestinal malrotation with inversion of the superior mesenteric artery (SMA) and vein (SMV) and failure of the
duodenum to cross the midline. Successful pancreaticoduodenectomy was performed in all cases with a modified
approach. Aberrant vascular and anatomic locations mandated careful parenchymal resection. Although the SMA
was more accessible at the head of the pancreas as it passed to the right of the SMV, the location of intestinal and
pancreatic branches required precise identification of their actual compared to expected course. Particular attention
was paid to the dissection of the uncinate process in order to avoid inadvertent injury to arteries supplying the small
bowel. The reconstruction was performed in standard fashion in all cases. All patients tolerated their procedures
well. Postoperative complications included 1 peripancreatic abscess requiring drainage.
CONCLUSION: Pancreaticoduodenectomy can be performed in patients with intestinal malrotation although a
thorough understanding of the vascular anatomy is required to complete the operation safely.
P031 PARTIAL COVERED BILIARY METALLIC STENT WITH/WITHOUT DUODENUM METAL STENT AND
NEOADJUVANT CHEMORADIATION THERAPY PROVIDE SYMPTOMATIC BORDERLINE RESECTABLE
PANCREATIC HEAD CANCER WITH A SAFE R0 SURGERY Kensuke Kubota, MD, Sho Hasegawa, MD, Ken
Ishii, MD, Yuji Fujita, MD, Yusuke Sekino, MD, Kunihiro Hosono, MD, Atsushi Nakaima, MD; Gastroenterology and
Hepatology,YokohamaCity University, Yokohama, JP
Background: Neoadjuvant chemradiationtherapy (NACRT) may lead to successful margin­negative resection (R0) in
pts with borderline resectable pancreatic head cancer (BRPHC). NACRT using a covered metallic biliary stent has
been attempted in pts with BRPHC, however, the efficacy of this therapy with/without metallic duodenal stent
(MDS) and the influence of using partially covered metallic stent (PCMS) for its delivery in the treatment of BRPHC
has not been evaluated. Aims: To evaluate the efficacy of and complications associated with the use of PCMS with/without MDS during
NACRT and the surgical period. Methods: We reviewed the outcomes of consecutive pts with BRPHC had histopathologically proven pancreatic
adenocarcinoma, who presented with symptomatic biliary obstruction, and divided the pts chronologically, in terms
of the period of stent placement into two groups: group A; plastic stent (PS) deployment plus NACRT between
August 2009 and October 2010; group B; prospectively PCMS deployment with/without MDS plus NACRT between
November 2010 and September 2014. The pts were categorized as having borderline resectable cancer based on
the NCCN clinical practice guideline established in 2013. Data on the pts demographics, complications, non re­
intervention rate (NRR), surgical time, operative blood loss, length of hospital stay, complications after resection,
the rate of R0 and prognosis were studied. Safe R0 surgery was defined as R0 surgery without the need for
endoscopic re­intervention or postoperative complications.
Results: There were a total of 85 pts with BRPHC (group A and B: 44 and 41 pts, respectively). The median time
from stent placement to surgery in the overall subject population were 126.1 days in group A and 127.6 days in
group B. MDS was deployed in two pts with group A and four pts with group B. NPR for the 1st 30 days in group A
(PS) and B (PCMS) were 60.5% and 97.6%, respectively. NPR for the 2nd 30 days in group A and B were 40.6%
and 85.1%, respectively. NPR for the 3rd 30 days in group A and B were 40.6% and 79.9%, respectively.
Regarding NPR, PCMS is superior to group using PS (logrank, p<0.0001). There were no significant differences
between groups regarding surgical time, operative blood loss, length of hospital stay and prognosis. The rates of
achievement of R0 surgery in groups A and B were 66.7% (30/45) and 66.9% (25/41), respectively. The PCMS and
MDS did not interfere with the conduct of the NACRT and pancreaticoduodenectomy in any pts. The rates of
achievement of safe R0 surgery in groups A and B were 9.1% (4/44) and 39% (19/41), respectively (p<0.0001).
Multivariate analysis showed that odds ratio for safe R0 surgery was 8.636 (p=0.0004) for PCMS placement. Conclusions: Insertion of PCMS should be considered for the relief of biliary and/or duodenum obstruction in pts with
BRPHC scheduled to receive NACRT, in view of the minimize need for re­intervention for recurrent biliary
obstruction, and a potentially higher rate of achievement of safe R0 surgery, as compared to the results obtained
with PS deployment.
P032 PERSONALIZED MEDICINE: A NEW MODEL FOR PRIMARY AND SECONDARY PANCREATIC
NEOPLASIA PREVENTION Milena Di Leo, MD1, Raffaella A Zuppardo1, Roberta Maselli1, Elisa Radice1, Andrea
M Tamburini2, Paola Zanelli3, Maurizio Ferrari4, Luca Albarello5, Michele Reni6, Monica Ronzoni6, Pier Alberto
Testoni1, Giulia Martina Cavestro1; 1Gastroenterology Unit, IRCCS San Raffaele Scientific Institute, Vita­Salute
San Raffaele University, 2Gastrointestinal Surgical Unit, Department of Surgery, IRCCS San Raffaele Scientific
Institute, 3Immunogenetic Unit, Parma University Hospital Parma, Parma, Italy, 4Clinical Molecular Biology,IRCCS
San Raffaele Scientific Institute,Vita­Salute San Raffaele Univers, 5Department of Pathology, IRCCS San Raffaele
Scientific Institute, Vita San Raffaele University, 6Department of Oncology,Division of Experimental
Oncology,IRCCS San Raffaele Scientific Institute, Milano, IT
BACKGROUND: Pancreatic ductal adenocarcinoma (PCA) is the fourth most common cause of death from cancer
in the USA and one of the deadliest cancers in the world with a 5­year survival of less than 7%. Therefore, follow up
of high­risk individuals is fundamental to improve the overall benefit of prevention.
AIM: The aim of this proposal is to apply the personalized based approach in high­risk patients for pancreatic
cancer.
METHODS: The project will include all individuals who have at least one first­degree relative who developed
pancreatic cancer. We propose an initial encounter contact phase with a geneticist in order to better define if the
disease has a syndromic inheritance or not (endogenous risk factor) and propose, if necessary, specific genetic
screening. The second summary phase is performed by a gastroenterologist in order to define the exogenous
associated risk factors. In this second phase, we will provide all the options and decisions for the next clinical
diagnostic step. A continued personalized follow­up will be performed, creating a specific diagnostic flow chart
based on the current literature. If necessary, a multidisciplinary team composed by a geneticist, a surgeon, an
oncologist and a pathologist is involved in the decision process.
RESULTS: On the basis of the current literature we created a multidisciplinary diagnostic flow chart (Figure 1).
The individuation of high risk subpopulations based on genetic features has the potential: (i) to define the appropriate
follow up in order to reach early identification of neoplastic/pre­neoplastic lesions, (ii) to optimize diagnostic
benefit/risk ratio.
CONCLUSIONS: Gastrointestinal personalized medicine is a multidisciplinary new model for primary and secondary
pancreatic neoplasia prevention in high risk subpopulations based on genetic features and exogenous risk factors
analysis.
P033 PRACTICE PATTERNS AND IMPACT OF IMAGING SURVEILLANCE AFTER RESECTION FOR
PANCREATIC DUCTAL ADENOCARCINOMA June S Peng, MD, Colin O'Rourke, Gareth Morris­Stiff, MD, PhD,
R. Matthew Walsh, MD, Sricharan Chalikonda, MD; Cleveland Clinic, Cleveland, US
Background
Current strategy for surveillance after surgical resection for pancreatic ductal adenocarcinoma (PDAC) varies widely
with no evidence basis to guide practice. The current NCCN recommendation is to perform a history and physical,
computed tomography (CT), and obtain CA 19­9 levels every 3­6 months for two years and annually thereafter. We
examined our practice in order to elucidate the impact of imaging studies in the care of patients after surgical
resection.
Methods
Patients who underwent resection for PDAC at our academic, tertiary care institution between 1/1/2010 and
12/31/2012 were identified from institutional databases. A total of 149 patients underwent partial or total
pancreatectomy for pancreatic cancer during the time period. Patients were screened and excluded if they did not
undergo adjuvant oncologic treatment (13 patients), had treatment or surveillance for a concurrent malignancy (7
patients) or had incomplete records due to follow­up at outside institutions (70 patients). Due to the nature of our
practice as a tertiary referral center, the majority of patients received oncologic treatment locally after surgical
resection. A total of 59 patients were included in the final analysis. Patient demographics, symptoms, imaging
results, and treatment regimens were extracted from a prospectively collected electronic medical record. All CTs,
PET­CTs, and MRIs of the chest and/or abdomen obtained after surgical resection were correlated with a clinical
encounter at the time of imaging to determine associated symptoms and decision making. Further intervention was
defined as initiation of therapeutic anticoagulation, additional imaging, and any procedure or surgery. A change in
treatment plan was defined as a change in the chemoradiation regimen or enrollment in hospice or palliative care.
Results
Of the 59 patients included in the study, 33 were male (56%) and the median age was 68 years (range 35­83 years).
The median length of follow­up was 726 days (range 83­1660 days). In total, 430 cross­sectional imaging studies
were acquired, with a median number of 6 studies per patient (range 1­27) and median time between studies of 81
days (range 0­490 days). Of the 430 studies, 77 studies prompted further intervention (17.9%) and 57 studies altered
the treatment plan (13.3%). Patients were asymptomatic for 253 of the 430 studies (58.8%), with 32 of the 253
studies (12.6%) prompting further intervention and 25 studies (9.9%) altering the treatment plan. Symptoms were
reported for 177 of the 430 studies (41.2%), with 45 of the 177 studies (25.4%) prompting further intervention and 32
studies (18.1%) altering the treatment plan. Of the 107 studies (24.9%) acquired when new symptoms were
reported, 34 of the 107 studies (31.8%) prompted further intervention and 18 studies (16.8%) changed the treatment
plan.
Conclusion
Our institutional experience demonstrates that the majority of patients are asymptomatic at the time of surveillance
imaging for follow­up after resection for PDAC. Although the presence of symptoms, and especially new symptoms,
increased the likelihood that an imaging result would prompt further intervention or a change in the treatment plan,
imaging studies impacted decision making in a minority of cases both for symptomatic and asymptomatic patients.
P034 RADIOGRAPHIC RESPONSE AND RESECTABILITY OF LOCALLY ADVANCED, BORDERLINE
RESECTABLE PANCREATIC ADENOCARCINOMA AFTER EXTENDED NEOADJUVANT
CHEMOTHERAPY. Jesse Clanton, J B Rose, Adnan Alseidi, Thomas Biehl, Scott Helton, Flavio Rocha; Virginia
Mason Medical Center, Seattle, US
Background: Neoadjuvant therapy is often utilized for locally advanced, borderline resectable pancreatic cancer (BR­
LAPD) prior to resection, despite previous reports suggesting minimal downstaging. However little is known about
the effect of extended preoperative chemotherapy on disease response.
Methods: Retrospective review of a prospective database of consecutive patients with BR­LAPD treated with a 24­
week course of gemcitabine and docetaxel between 2011 and 2014 was performed. Patients with high­quality
imaging before and after treatment, and who completed full course of therapy with intention to resect were included.
RECIST 1.1 criteria were used to assess radiographic response.
Results: Forty­five patients who completed extended neoadjuvant chemotherapy were included for analysis. Thirty­
two (71.1%) patients underwent pancreaticoduodenectomy and 14 (31.1%) received a vein resection. Thirteen
(28.9%) were not resected due to local or distant progression or reduced performance status. The majority of
patients demonstrated stable tumor size (24/45, 53.3%) or partial response (20/45, 44.4%). Suspicious lymph nodes
remained stable (27/45, 60%) or improved (11/45, 24.4%) while vascular involvement was stable (21/45, 46.7%) or
improved (16/45, 35.6%) after treatment. None of the ten patients not resected after chemotherapy had an increased
tumor size, three had enlarging lymph nodes, and one had more extensive vascular involvement by CT. None of
these factors were associated with ability to undergo resection (p > 0.05)
Conclusions: Extended neoadjuvant chemotherapy for BR­LAPD commonly results in either stable disease or a
partial response. One­third of patients have demonstrable downstaging of vascular involvement, however
radiographic response or downstaging is not associated with successful resection.
P035 RESULTS OF THE RAMPS PROCEDURE IN 78 PATIENTS WITH ADENOCARCINOMA OF THE
DISTAL PANCREAS: DOES THE PROCEDURE ATTAIN THE ONCOLOGIC GOALS FOR RESECTION OF
LEFT SIDED PANCREATIC ADENOCARCINOMAS? Julie G Grossman, MD1, Feng Gao, MD, PhD, MPH1,
Ryan Fields, MD1, William Hawkins, MD1, David Linehan, MD2, Steven Strasberg, MD1; 1Washington University
School of Medicine in St.Louis, 2University of Rochester, Saint Louis, US
Introduction: Tangential margin negativity and adequate lymph node counts are the principal oncologic goals for
pancreatic adenocarcinoma. RAMPS (Radical Antegrade Modular Pancreato­Splenectomy) was designed to
improve results for these outcomes in cancers of the distal pancreas by standardizing the tangential planes of
resection and the lymph node groups to be resected. Herein, we report the outcomes of patients undergoing
RAMPS at a single institution.
Methods: Between 1999 and 2012, 78 consecutive patients underwent RAMPS for adenocarcinoma of the distal
pancreas. There were 47 females (60.2%) and the mean age was 67 years. Anterior RAMPS was performed in 56
patients and posterior RAMPS (left adrenal resected en bloc) in 22 patients. Tangential margins were inked in the
resected specimen and studied microscopically. R0 was declared when tumor was >1mm from ink. R1(t) indicates
a positive tangential margin and R1(p) a positive parenchymal margin. High lymph node ratio was declared when
>20% of resected nodes contained cancer. Results: Mean tumor diameter was 4.5 cm (2.4 SD), 84.6% of tumors were T3 (extra pancreatic invasion) and 95%
were histologic grades G2­4. Microscopic perineural, lymphatic and venous invasion were present in 68%, 52.6%
and 48.6% of specimens respectively. Tangential margins were negative in 73/78 patients (93.6%), that is to say
only 5 cases were R1(t) and 7 were R1(p). Mean lymph node counts were 20.2 (12.1 SD). 82.1% of patients had at
least 10 nodes in the resected specimen and 65.4% had 15 nodes. 39/78 specimens (50%) were N1. With a mean
follow­up of 78.4 months (49.1 SD), overall 5 year survival was 25.2%, and median survival was 24.2 months. In
multivariate analysis, factors associated with improved survival included: low lymph node ratio, histologic grade G1
or G2, and Caucasian race. N1 patients had poorer survival, but this did not reach statistical significance as only
5/78 patients had microscopically positive tangential margins the effect of this factor on survival could not be
determined. Conclusion: RAMPS fulfills the oncologic goals of negative tangential margins and adequate lymph node resection.
Not unexpectedly, factors related to tumor aggressiveness (i.e. histologic grade and lymph node status) are
important determinants of survival. Adequate lymph node counts are needed to make this determination. The cause
of poor survival in non­Caucasians needs further investigation. Table 1 Factors influencing surivival­ multivariate analysis
Variable
Hazard ratio
95% CI
P value
Race (black, white)
0.24
0.12­0.48
<.001
Grade (G1/G2, G3/G4)
2.01
1.15­3.51
0.014
Lymph node ratio (LNR<0.2, LNR>0.2)
3.51
1.71­7.20
<.001
Tobacco use (no, yes)
2.08
1.04­4.17
0.039
P036 RISK OF MISDIAGNOSIS AND OVERTREATMENT IN PATIENTS WITH MAIN PANCREATIC DUCT
DILATATION AND SUSPECTED COMBINED/MAIN­DUCT IPMNS Stefano Crippa1, Ilaria Pergolini1, Corrado
Rubini1, Giorgia Marchesini1, Paola Castelli2, Alessandro Pucci1, Giuseppe Zamboni2, Massimo Falconi1;
1Universita' Politecnica delle Marche, 2Ospedale Sacro Cuore Negrar, Ancona, IT
Introduction/Background: Diagnostic errors are quite common in the evaluation of cystic neoplasms of the pancreas.
Segmental/diffuse dilatation of main pancreatic duct (MPD) more than 5 mm is the typical feature of combined/main­
duct intraductal papillary mucinous neoplasms (CMD­IPMNs). Although MPD dilatation in IPMNs is commonly
caused by tumor involvement, it may be also the expression of mucus hypersecretion/obstructive chronic
pancreatitis (OCP). Moreover, pathologic conditions other than CMD­IPMNs can cause MPD dilation. Aim of this
study is to evaluate the real presence and extension of MPD involvement by tumor in patients who underwent
pancreatic resection for suspected CMD­IPMNs
Methods Retrospective analysis of patients with a preoperative diagnosis of CMD­IPMNs undergoing
pancreatectomy between January 2009 and October 2014. Pathological correlation among MPD dilatation, presence
of IPMN, OCP or other diseases was searched.
Results 93 patients (60 males, median age 67 years) were identified. Surgical procedures included 37
pancreaticoduodenectomies (40%), 31 total (33%) and 25 distal pancreatectomies (27%). At pathology, CMD­IPMNs
were found in 69 patients (74%). In the remaining 24 patients (26%), branch­duct IPMNs (BD­IPMNs) were found in
8 cases (9%), PDAC in the absence of IPMN in 9 (10%), cystic neuroendocrine tumor (NET G2) in 1 (1%), serous
cystadenoma (SCA) in 2 (2%), and OCP alone/mucinous metaplasia in 4 patients (4%). Six out of eight BD­IPMNs
were symptomatic and/or with high­risk stigmata. Considering 69 CMD­IPMNs, 27 (39%) underwent total
pancreatectomy because of positive resection margin after initial partial pancreatectomy or because diffuse and
extensive involvement of MPD at preoperative imaging. However, in 9/27 (33%) patients the IPMN was found only
in a pancreatic segment at final pathology, while OCP was present in the remaining portion. In these patients partial
pancreatectomies could be performed. Median size of MPD in IPMN­involved area was 9.5 mm compared with 6
mm when only OCP was found (P<0.05). Considering the entire cohort of 93 patients, 18 (19%) underwent an
overtreatment because unnecessary surgery (2 BD­IPMNs, 2 serous cystadenomas and 4 OCPs only) or too
extensive resections (9 CMD­IPMNs and one PDAC with associated OCP). Total pancreatectomy was the most
common procedure (67%) performed in these patients.
Conclusion There is a considerable risk of overtreatment in patients with preoperative diagnosis of CMD­IPMNs.
New strategies, including advanced endoscopic techniques such as pancreatoscopy, are necessary to improve the
diagnostic yield in this setting. Partial pancreatectomy with intraoperative margin examination and pancreatoscopy
should be performed instead of upfront total pancreatectomy when surgery is mandatory and diffuse dilation of MPD
is evident. Strict radiological observation instead of immediate resection can be considered in asymptomatic patients
with mild MPD dilatation (< 9mm) lacking high­risk stigmata.
P037 ROLE OF COMBINED 68GA­DOTATOC AND 18F­FDG PET­CT IN THE DIAGNOSTIC WORKUP OF
WELL AND MODERATELY DIFFERENTIATED NEUROENDOCRINE TUMORS OF THE PANCREAS
(PNETS): A SURGICAL SERIES. G Butturini1, S Ortolani2, A Malpaga1, S Cingarlini2, V Malfatti3, P Capelli4, A
Ruzzenente5, M D'Onofrio6, Mv Davì7, P Vallerio7, E Grego2, C Trentin2, Gp Tortora2, A Scarpa4, M Salgarello3, C
Bassi1; 1The Pancreas Institute Surgical Unit, 2The Pancreas Institute Oncology Unit, 3Department of Nuclear
Medicine Ospedale Sacro Cuore Negrar­Verona, 4The Pancreas Institute Pathology Unit, 5Department of Surgery
and Oncology, Hepatobiliary Unit, 6The Pancreas Institute Radiology Unit, 7The Pancreas Institute Endocrinology
Unit, Verona, IT
Background­ pNETs are characterized by a broad spectrum of aggressiveness. Somatostatin­receptor based
functional imaging (i.e.[68Ga]DOTATOC­PET/CT, GP) is recommended in the diagnostic workup of pNETs;
preliminary evidences show that [18F]FDG­PET/CT (FP) can provide additional informations about prognosis.
Purpose­ To assess the role of combined GP and FP in the evaluation of G1­2 pNETs and to test the correlation
between FP positivity and tumoral grade.
Methods­ Preoperative GP and FP of 35 patients with surgically resected G1­2 pNETs were evaluated. Correlation
between FP positivity (SUV max > 3.5) and tumor grade was calculated.
Results­ Among 35 surgically resected pNETs, 10/35 (28.6%) were G1 and 25/35 (71.4%) were G2. GP showed
high sensitivity in detecting G1­2 pNETs (94.3%). FP resulted positive in 2/10 (20%) and 19/26 (76%) G1 and G2,
respectively. Therefore, FP was able to identify G2 pNETs with high positive predictive value (PPV, 90.5%).
Conclusions­ The high sensitivity of GP in NETs detection is well recognized. The high PPV (90.5%) of FP in the
identification of G2 forms suggests its potential role in prognostication and risk stratification for pNETs. The
complementary informations provided by FP may have relevance in the decision­making for “small” (< 2 cm)
potentially resectable pNETs. Given the reliability of FP correlation with tumor grade determined on surgical
specimens, FP could be proposed in the diagnostic workup of pNETs to target the bioptic sample on FP positive
areas and to guide the “interpretation” of bioptic diagnosis.
P038 SIGNIFICANCE OF HISTOLOGICAL RESPONSE FOR PREDICTING THE OUTCOME IN PATIENTS
WITH PANCREATIC DUCTAL ADENOCARCINOMA RESECTED AFTER GEMCITABINE­BASED
CHEMORADIOTHERAPY Hiroyuki Kato, MD, PhD, Ryosuke Desaki, MD, PhD, Yasuhiro Murata, MD, PhD,
Akihiro Tanemura, MD, PhD, Naohisa Kuriyama, MD, PhD, Yoshinori Azumi, MD, PhD, Masashi Kishiwada, MD,
PhD, Shugo Mizuno, MD, PhD, Masanobu Usui, MD, PhD, Hiroyuki Sakurai, MD, PhD, Shuji Isaji, MD, PhD;
Department of Hepatobiliary pancreatic and transplant surgery, Mie university hospital, Tsu, Mie, JP
Background
Prognosis of patients with pancreatic ductal adenocarcinoma (PDAC) has been extremely poor because of the high
rates of distant metastasis even after R0 resection. To overcome this problem, we conducted the gemcitabine
(Gem)­based chemoradiotherapy (CRT) followed by surgery for the patients with locally advanced PDAC from 2005,
and sometimes experienced the cases whose resected specimens showed an excellent histological response
showing that the necrosis area of tumor was more than 90%. However, there have been no studies in which this
effect of CRT really contributes the improvement of patient prognosis. The aim of this study is to elucidate the
characteristics of PDAC whose histological response was categorized as Evans III (necrosis >90%) or IV (100%)
and compare the patient prognosis according to the Evans criteria.
Methods
Among 129 patients who underwent surgical resection after Gem­based CRT (50.4Gy+Gem 800mg/m2 alone: n=77
from 2005 to 2011 or Gem 600mg/m2+S­1 60mg/m2: n=52, from 2011­2013), we evaluated the reduction rate of
CA19­9 after CRT, R0 resection rate and patient survivals according to the Evans criteria III and IV or I, IIa and IIb.
Results
The 10% (13/129) of patients achieved significant tumor reduction categorized as Evans III (n=11) or IV (n=2). The
reduction rate of CA19­9 in patients whose tumor response was regarded as Evans III or IV was significantly higher
than that in the other patients (81.5% vs. 43.0%, p<0.001). Moreover, R0 resection rate of the patients with Evans
III or IV was significantly better than that of other patients (100% vs. 73.9% p<0.023). As for the patient survival, the
3­ and 5­ year survival rates and MST were dramatically improved in the patients with Evans III or IV, as compared
to the other patients (3­ and 5­year survival: 75.0% and 37.5% vs.28.8% and 24.0%,respectively MST: 37.9 vs. 22.4
months, p=0.012)
Conclusion
Our result suggested that the patients with Evans III or IV achieved the excellent prognosis because of the high
successful rate of R0 resection. Further study is required to reveal the precise mechanism by which these PDAC
have favorable tumor responses to CRT.
P039 SURGICAL STRATEGY FOR PATIENTS WITH RIGHT HEPATIC ARTERY VARIATIONS IN
PANCREATICODUODENECTOMY Ken­ichi Okada, MD, PhD, Manabu Kawai, Seiko Hirono, Motoki Miyazawa,
Atsushi Shimizu, Yuji Kitahata, Hiroki Yamaue; Wakayama Medical University, Wakayama, JP
Background/Purpose. Resectable/borderline resectable category at increased risk of higher likelihood of an R1
resection in patients with right hepatic artery (RHA) variation has not been well discussed. The aim of the present
study was to clarify the strategy in the patients with RHA variation undergoing pancreaticoduodenectomy (PD)
based on the tumor position and R1 resection rate.
Methods. The medical records of 180 consecutive patients who underwent PD for pancreatic ductal
adenocarcinoma between January 2000 and May 2013 were evaluated for the R1 resection rate, surgical outcome,
and survival. In this study, three types were defined as (i) resectable type, tumors situated more than 10 mm away
from the root of aberrant right hepatic artery (aRHA)/aberrant common hepatic artery (aCHA), (ii) adjacent type,
tumors situated within 10mm from the root of aRHA/aCHA without tumor abutment of the SMA, and, (iii) borderline
resectable type, the patients with tumor abutment of the SMA not to exceed 180° of the circumference of the vessel
wall, were classified.
Results. In the present study, 25 patients were identified to have RHA variation in preoperative imaging study. There were sixteen patients with resectable type, 5 with adjacent type, and 4 with borderline resectable type; the R1
resection rates revealed 6% (n=1), 80% (n=4), and 75% (n=3) respectively. aRHA/aCHA were preserved in 14
(88%) patients with resectable type, all of adjacent type, and none of borderline type pancreatic carcinoma. In
resectable/adjacent type cases, all positive margins were identified in the retropancreatic tissue adjacent to the
periarterial nerve plexuses around the aRHA/aCHA.
Conclusion. The strategy to resect aRHA of the adjacent type pancreatic carcinoma would improve the R0
resection rate.
P040 SYSTEMATIC REVIEW OF INNOVATIVE ABLATIVE THERAPIES FOR THE TREATMENT OF
LOCALLY ADVANCED PANCREATIC CANCER Sje Rombouts, MD1, Ja Vogel, MD2, Hc V Santvoort, MD,
PhD2, Kp V Lienden, MD, PhD2, R V Hillegersberg, MD, PhD, Prof1, Orc Busch, MD, PhD, Prof2, Mgh Besselink,
MD, PhD2, Iq Molenaar, MD, PhD1; 1University Medical Center Utrecht, 2Academic Medical Center Amsterdam,
Amsterdam, NL
Background: Locally advanced pancreatic cancer (LAPC) is associated with a very poor prognosis. Current palliative
(radio)chemotherapy provides only a marginal survival benefit of 2 – 3 months. Several innovative local ablative
therapies have been explored as new treatment options. This systematic review aims to provide an overview of the
clinical outcomes of these ablative therapies.
Methods: A systematic search in PubMed, Embase and the Cochrane Library was performed to identify clinical
studies, published before 1 June 2014, involving ablative therapies in LAPC. Outcomes of interest were safety,
survival, quality of life and pain.
Results: After screening 1037 articles, 38 clinical studies involving 1164 patients with LAPC, treated with ablative
therapies, were included. These studies concerned radiofrequency ablation (RFA) (7 studies), irreversible
electroporation (IRE) (4), stereotactic body radiation therapy (SBRT) (16), high­intensity focused ultrasound (HIFU)
(5), iodine­125 (2), iodine­125 – cryosurgery (2), photodynamic therapy (1) and microwave ablation (1). All strategies
appeared to be feasible and safe. Outcomes for postoperative, procedure­related morbidity and mortality were
reported only for RFA (4 – 22 and 0 – 11 per cent respectively), IRE (9 – 15 and 0 – 4 per cent) and SBRT (0 – 25
and 0 per cent). Median survival of up to 25.6, 20.2, 24.0 and 12.6 months was reported for RFA, IRE, SBRT and
HIFU respectively. Pain relief was demonstrated for RFA, IRE, SBRT and HIFU. Quality­of­life outcomes were
reported only for SBRT, and showed promising results.
Conclusion: Ablative therapies in patients with LAPC appear to be feasible and safe and promising median survival
rates are shown.
P041 VALIDATION OF IMMEDIATE PERITONEAL WASHING CYTOLOGY RESULTS IN PANCREATIC AND
GASTRIC CANCER Andrea Porpiglia, MD, Hormoz Ehya, MD, John P Hoffman, MD; Fox Chase Cancer Center,
Philadelphia, US
Introduction:
The mainstay of treatment for pancreatic and gastric cancers is surgical resection. Unfortunately many of these
patients present with locally advanced, unresectable or distant disease and therefore medical management may be
of more benefit. Accurate staging of patients with pancreatic and gastric cancer is essential in determining the best
treatment strategy. Despite preoperative imaging there remains a group of patients that has clinically occult
metastatic disease. Positive peritoneal cytology is a poor prognostic indicator for survival in both gastric and
pancreatic cancer. Surgical resection may not be of benefit in those with positive peritoneal cytology. At our
institution, a diagnostic laparoscopy with peritoneal washings is performed prior to surgical resection. We performed
a retrospective review to evaluate the accuracy of immediate peritoneal washing interpretation in both gastric cancer
and pancreatic cancer.
Results:
There were 51 patients that underwent immediate peritoneal washing interpretations. There were 5 patients with
gastric adenocarcinoma, 2 patients with cholangiocarcinoma, and 44 patients with pancreatic adenocarcinoma. Four
of the patients had positive cytology for tumor cells with immediate interpretation, and 47 patients had cytology
negative for tumor cells with immediate interpretation.
There was only one patient with negative cytology on immediate interpretation with cytology positive for malignant
cells on final cytopathologic results. There were many RBCs noted within the peritoneal fluid specimen, which is
probably why the cancer cells were missed. Immediate peritoneal cytology results had a 100% positive predictive
value, and a 97% negative predictive value. There were no false positive results.
Discussion:
Positive peritoneal cytology is considered a poor prognostic factor for survival in both gastric and pancreatic cancer.
In pancreatic and gastric cancer, previous studies have shown resection in the presence of metastatic disease does
not improve survival. Diagnostic laparoscopy has been used as an adjunct to help stage patients before proceeding
with a radical resection that has an associated significant morbidity and mortality rate is futile and potentially
harmful, delaying or obviating systemic therapy that may be of more benefit to the patient. Our results show that
utilizing immediate interpretation of peritoneal cytology is reliable and accurate. There was only one patient with a
false negative result and there were no false positive results. Therefore, prior to surgical resection immediate
interpretation can be utilized.
Conclusion:
Diagnostic laparoscopy with peritoneal washings and immediate interpretation of cytology can be used prior to
proceeding with surgical resection. A futile operation can potentially be avoided with the use of immediate
interpretation of the peritoneal fluid.
P042 A PROSPECTIVE STUDY OF SURGICAL OUTCOME AND DIFFERENCES ON HISTOPATHOLOGY IN
PATIENTS WITH ALCOHOLIC AND NON ALCOHOLIC CHRONIC PANCREATITIS (CP) Srinath S R, MS,
Rajesh Gupta, Professorsurgical, gastroenterology, Sunil Shenvi, MSMchsurgical, gastroenterology, Deepak Bhasin,
ProfessorGastroenterology, Ritambhra Nada, Associate, Professor, Dept, of, Histopath, Mandeep Kang, Associate,
ProfessorDeptRadiodiagnosis, Naresh Sachdeva, Associate, ProfessorDeptEndocrinology; PGIMER,Chandigarh,
Chandigarh, IN
Background:Alcoholic & non­alcoholic chronic pancreatitis appears to be two different diseases with common final
outcome of pancreatic exocrine & endocrine failure.Present prospective study was planned to find the differences in
both these in relation to postoperative pain relief ,postoperative changes in exocrine & endocrine function,Imaging
and histopathology.
Materials and methods:All patients of chronic pancreatitis admitted in Division of Surgical gastroenterology and
Dept. Medical Gastroenterology, PGIMERfrom January 2012 to June 2014were included in the study.Informed
written consent was taken from all patients .Surgery was offered in patients who continue to suffer from symptoms
despite best medical, endoscopic & radiological management
Results: Total of 24 patients were included in the study with 13 patients being alcoholic pancreatitis and 11 patients
being non alcoholic pancreatitis.Mean followup was 15 months.All patients with non alcoholic pancreatitis underwent
Frey’s procedure. Four patients with alcoholic pancreatitis underwent Pylorus preserving pancreaticoduodenectomy
due to head mass and rest all underwent Frey’s procedure. When we compared impact of surgery on pain relief,
patients with nonalcoholic pancreatitis had statistically significant better results compared to alcoholic
pancreatitis.Three patients in non alcoholic pancreatitis and 2 patients with alcoholic pancreatitis had complete
resolution of diabetes during follow­up.Patients with alcoholic pancreatitis continued to have steatorrhea in follow up
period but none of non alcoholic pancreatitis patients had deterioration in exocrine function .When we compared
histopathology in both the groups, patients with alcoholic pancreatitis has statistically significant increase in
intralobular inflammatory infiltrate & fibrosis,neuritis , nerve hypertrophy and predominant plasma cell
infiltration.Patients with non alcoholic pancreatitis had duct centric infiltration with predominant lymphoplasmacytes.
Conclusion:In this prospective study, non­alcoholic CP patients had better outcome after surgery than alcoholic CP
in terms of postoperative pain relief, endocrine and exocrine insufficiency.Histopathologically, pancreatic changes in
patients with non­alcoholic chronic pancreatitis markedly differ from those with alcoholic chronic pancreatitis.
P043 A VERIFICATION STUDY OF THE FISTULA RISK SCORE NEWLY LAUNCHED ON PANCREAS
CLUB WEBSITE Hisashi Kosaka, Y Asano, K Suzumura, A Kurimoto, T Okamoto, K Ohashi, S Hai, Y Kondo, I
Nakamura, N Uyama, T Okada, T Hirano, Y Iimuro, J Fujimoto; Hyogo College of Medicine, Nishinomiya, JP
Introduction: Pancreatic fistula (PF) remains troublesome complication after pancreas head resection. The Pancreas
Club newly launched the fistula risk score (FRS) calculator on their web in 2014. In this study, we verified the
scoring system and tried to elucidate the relationship between the score and pathophysiological characteristics of
the pancreatic tissue.
Methods: Clinical data of 88 patients who underwent pancreas head resection between 2009 and 2012 were
retrospectively reviewed. FRS was calculated on the web site based on the data of gland texture, pathology,
pancreatic duct size and intra­operative blood loss. The FRS was categorized as follow: Negligible (0), Low (1­2
pts), Mod (3­6 pts) and High (7­10 pts). Degree of PF were classified according to ISGPF definition, and grade B/C
were considered as clinically relevant PF (cPF). Percentage of the pancreatic fibrous area (%PFA) and the
pancreatic secretary gland area (%PSGA) were evaluated by Image­J following Azan­Mallory staining of pancreatic
tissues. Pancreatic juice was collected via external trans­anastomosis stent of pancreaticojejunostomy.
Results: cPF was noted in 30.7% of patients. cPF rate was increased according to FRS: Negligible (0%), Low (0%),
Mod (21.7%) and High (56.7%). %PFA was decreased according to increase of FRS: Negligible (37.8%), Low (25.9
± 4.1%), Mod (16.3 ± 1.5%) and High (8.0 ± 1.3%), while %PSGA was increased: Negligible (1.6%), Low (18.1 ±
4.2%), Mod (25.6 ± 2.1%) and High (34.6 ± 1.8%). Quantity of pancreatic juice per day was also increased according
to increase of FRS: Negligible (5.6ml), Low (52.5 ± 16.1ml), Mod (101.3 ± 15.8ml) and High (202.9 ± 18.0ml). These
findings indicated FRS could reflect the characteristics of each pancreatic tissue regarding to pancreatic juice
production. To evaluate the diagnostic ability of FRS, receiver operating curve analysis was performed. Area under
the curve of FRS to cPF diagnosis was 0.81 and cut­off value was 5.5. These findings indicated FRS could fairly
detect the high risk case.
Conclusion: FRS could represent the characteristics of pancreatic tissue and had an acceptable potential to detect a
high risk case of cPF.
P044 DISTAL PANCREATECTOMY WITH ISLET AUTOTRANSPLANT FOR SELECT PATIENTS WITH
FOCAL CHRONIC PANCREATITIS Sydne Muratore, MD, Melena Bellin, MD, Ty Dunn, MD, FACS, Timothy
Pruett, MD, Alfred Clavel, MD, Josh Wilhelm, MS, Srinath Chinnakotla, MD, David Sutherland, MD, PhD, Greg
Beilman, MD, FACS; University of Minnesota, Minneapolis, US
Introduction: In select patients with chronic pancreatitis involving the tail only, distal pancreatectomy (DP) may be
preferred over total pancreatectomy to avoid the morbidity of total resection. The aim of this report is to examine
metabolic outcomes of distal pancreatectomy patients undergoing simultaneous islet autotransplant for chronic
pancreatitis compared to matched total pancreatectomy islet autotransplant (TPIAT) counterparts.
Methods: All patients undergoing distal pancreatectomy with islet autotransplant (DPIAT) at UMN for chronic
pancreatitis between 2004­ 2014 were identified for inclusion in this single­center observational study. These
patients were compared to age­, gender­, and BMI­matched controls (n=24) with a history of TPIAT. Outcomes
included insulin use, HgbA1C, C­peptide positivity, maximum stimulated C­peptide, narcotic use, reoperation for
pain, and islet equivalent (IEQ) transfused. Nonparametric statistics were used to compare groups.
Results: Twelve patients (50% female) with a mean age of 40 years underwent DPIAT. Disrupted duct was the
preoperative etiology in 7 of 12 (58%) DPIATs. Metabolic data was available for 8 of 12 DPIATs and 24 of 24
TPIATs at 1 year (+/­ 8 months) follow­up. Five of eight (63%) DPIATs were insulin free compared to 9 of 24 (38%)
TPIATs. All 8 (100%) DPIATs had HgbA1c < 7.0 compared to 19 of 24 (79%). C­peptide positivity (> 0.6ng/mL)
was 5 of 5 (100%) for DPIATs compared to 21 of 23 (91%) for TPIATs. The maximum stimulated C­peptide was
also higher for DPIATs than TPIATs (5.95 vs 2.8ng/mL, p = 0.003). IEQ/kg was significantly lower for DPIATs vs
TPIATs (1,418 vs 5,267, p = 0.001). Narcotic data was available for 7 of 12 DPIATs at 1 year (+/­ 8 months) follow­
up. Five of seven (71%) of DPIATs compared to 13 of 24 (54%) TPIATs were no longer taking chronic narcotics. Three of twelve DPIAT patients (25%) underwent completion pancreatectomy for recurrence of pain at a mean of 44
months (7­92 months). All 3 patients had evidence of diffuse disease at time of distal pancreatectomy.
Conclusion: This is the first report of a series of DPIAT for chronic pancreatitis and demonstrates 100% C­peptide
positivity and improved metabolic outcomes compared to TPIAT. DPIAT was effective in relieving pain in over two­
thirds of patients, as measured by narcotic­free status at 1 year. IAT preserves islet cell mass in patients
undergoing pancreatectomy and may decrease the complications of diabetes in patients with progressive disease. DPIAT may be appropriate for patients with isolated distal pancreatic disease due to disconnected duct or other
similar pathology.
P045 DOES POST­OPERATIVE COMPLICATIONS REALLY AFFECT THE ONCOLOGICAL RESULTS AFTER
PANCREATICO­DUODENECTOMY FOR CANCER? Gennaro Nappo, MD1, Michel El Bechwaty, MD1, Julie
Perinel, MD1, Roberto Coppola, MD, Ph, FACS2, Mustapha Adham, MD, Ph1; 1HPB Surgery, Edouard Herriot
Hospital, Lyon, France, 2General Surgery, Campus Bio­Medico University of Rome, Lyon, FR
Introduction
Pancreatico­Duodenectomy (PD) is the curative treatment for patients affected by periampullary tumours. With the
improvement of surgical techniques and perioperative care, morbidity and mortality rates after PD have decreased in
the last decades (ranging around 30­50% and 3­5%, respectively, in high volume pancreatic centers). However,
oncological outcomes after curative resection remain unsatisfactory (5­year survival: 20%). In other surgical
specialities, recent studies demonstrated that post­operative complications could affect long­term outcomes. The
relationship between post­operative complications and oncological outcomes after pancreatic surgery has not been
well evaluated. The aim of the present study is to evaluate the influence of post­operative complications on
prognosis after PD for cancer.
Material and methods
From 2008 to 2014, all consecutive PD performed for periampullary tumours at Edouard Herriot Hospital of Lyon
were retrospectively evaluated. Post­operative complications were classified according to the Clavien­Dindo’s
classification. All cases were stratified in two group: a) absence of complications or presence of mild complications
(grade I­II) (group 1); b) presence of severe complications (grade > 3) (group 2).
Oncological data were collected and differences in terms of overall survival (OS) and Disease Free Survival (DFS)
between the two groups were evaluated.
Results
We retrospectively analysed 208 consecutives PD performed in the study period. PD was performed for pancreatic
ductal adenocarcinoma in 169 cases (81%), for ampullary cancer in 15 cases (7.2%), for distal cholangiocarcinoma
in 24 cases (11.5%). One­hundred­twenty­six (60.6%) and 82 (39.4%) cases were stratified in group 1 and 2,
respectively. Overall mortality was 3.8%. There were no differences in TNM stage and R status between the two
groups, while difference in terms of access to adjuvant treatment was found (64.3% and 47.3% in group 1 and 2,
respectively) (p=0.014)
Mean OS and DFS for the entire cohort was 37.6 and 32.6 months, respectively. No statistical difference in terms of
OS and DFS were found between the two groups (39.5 and 34,4 months, in group 1 versus 34.4 and 29.4 months in
group 2, respectively) (p > 0.05). The multivariate analysis showed that positive nodal metastasis (N1) (p=0.006)
and R1 resection (p=0.001) were independent prognostic of survival.
Discussion
The results of this retrospective study don’t seem to show a relationship between the occurrence of post­operative
complications and oncological results after PD for cancer. Our study confirms that lymphnode metastases (N1) and
R1 resections are the most useful prognostic factors.
P046 EVALUATION OF CENTRAL PANCREATECTOMY AND PANCREATIC ENUCLEATION AS
PANCREATIC RESECTIONS ­ A COMPARISON Marius Distler, MD1, Steffen Wolk, MD1, Stephan Kersting,
MD2, Weitz Jürgen, Prof1, Grützmann Robert, Prof1; 11Department for General, Thoracic and Vascular Surgery,
Universityhospital Carl Gustav Carus, TU­Dr, 2Department for General and Vascular Surgery, RKK Hospital­
St.Josefs, Freiburg, Germany, Dresden, DE
Introduction: For minor pancreatic resection such as enucleation and central pancreatectomy comparative data are
rare. These techniques provide parenchyma­sparing alternatives to major resections (e.g. pancreaticoduodenectomy)
for neuroendocrine tumours, cystic tumours or metastases. This study retrospectively compares the morbidity and
mortality of both techniques, with special regard to the formation of postoperative pancreatic fistulas.
Methods: Between December 1996 and November 2013 the postoperative events and clinical outcomes of 17
patients after pancreatic enucleation and 26 patients after central pancreatectomy were retrospectively analyzed
from a prospectively collected database.
Results: Perioperative mortality was 0% in both groups. There was no significant difference in the overall peri­
operative morbidity (CP 80.8% vs. PE 82.4%). The major cause of the high morbidity was the formation of a
postoperative pancreatic fistula with 26.9% of the patients after central pancreatectomy and 35.3% after pancreatic
enucleation.. Univariate analysis showed a BMI over 30 kg/m2 in the central pancreatectomy group to be an
independent risk factor. Additional minor complications, e.g. urinary tract infection, pleural effusion, etc. furthermore
contributed to the perioperative morbidity.
Conclusion: Pancreatic enucleation and central pancreatectomy are feasible techniques for selected patients, but
the indications are limited. Morbidity after these resections is high with the major cause being the development of a
postoperative pancreatic fistula.
P047 FAST­TRACK PATHWAY AFTER PANCREATICODUODENECTOMY. SPECIFIC DIET THERAPY
PROTOCOLS REDUCES THE RATE OF DELAYED GASTRIC EMPTYING Sergio Valeri, Paolo Luffarelli, Sara
Emerenziani, Domenico Borzomati, Giovanbattista Giorgio, Rossana Alloni, Roberto Coppola; Campus Bio­Medico
University, Rome, IT
Background
Pancreaticoduodenectomy (PD) is the treatment of choice for periampullary tumors but it is still affected by high
morbidity (40­50%) and mortality (2­3%). Delayed Gastric Emptying (DGE) is the most common post­operative
complications after PD. According to literature data, DGE is associated with other post­operative major
complications such as pancreatic fistula and abdominal abscess. The present study reports the impact on DGE
incidence of a Fast­Track protocol and specific diet protocols. The aim was to value the impact of early oral feeding,
with specific diet protocols, on postoperative course focusing on the occurrence of DGE.
Methods
From January 2013 to October 2013, 23 consecutive patients underwent PD at the Department of Surgery of the
Campus Bio­Medico University. All patients started early oral feeding on post­operative day 2 (Table 1). Compliance
to diet protocol was monitored by using a food diary recording system. Incidence of DGE, defined according to
International Study Group Pancreatic Fistula (ISGPF), time of first stool canalization and length of hospital stay were
recorded. Data were compared with those obtained from 26 patients underwent PD from January 2012 to December
2012 in which specific diet therapy protocols were not applied.
Results
DGE incidence was significantly lower in the study groups compared to study control group (38% vs 8%
respectively p<0.05). The mean length of stay decreased from 14.6 day of the study control group to 13.4 day in the
study group. Time to first stool canalization was significantly shorter in the study group.
Conclusions
In our experience, the implementation of the Fast­Track program and the introduction of a specific diet therapy
protocol reduced the incidence of DGE, the length of hospital stay and the time to first stool canalization. A
randomized trial will provide further evidence on the impact of early feeding protocol on post­operative course after
PD.
Table 1
POD II III IV V VI
Kcal 405 640 1200 1580 1770
Proteins (g) 9 36 61 72 84
Lipids (g) 2 6 29 42 45
Carbohydrates (g) 90 103 185 241 274
Fibers (g) 2 3 6 14 18
P048 HAS SURVIVAL IMPROVED FOLLOWING RESECTION FOR PANCREATIC ADENOCARCINOMA?
Alexander Rosemurgy, MD, Robert Klein, BS, Carrie Ryan, MS, Prashant Sukharamwala, MD, Benjamin Sadowitz,
MD, Kenneth Luberice, MS, Sharona B Ross, MD; Florida Hospital Tampa, Tampa, US
Introduction: Billions of dollars have been spent on the research and treatment of pancreatic cancer. This study was
undertaken to determine if survival after resection of pancreatic adenocarcinoma has been extended over the past
two decades.
Methods: The SEER database was queried for patients who underwent pancreatectomy for pancreatic
adenocarcinoma from 1992 through 2010. AJCC Stage and survival were determined for each patient. Data were
analyzed using Mantel­Cox test and linear regression. Significance was accepted at p < 0.05.
Results: 15,604 patients underwent pancreatectomy from 1992 through 2010. Survival improved from 1992 through
2010 (p < 0.0001), as denoted in Figure 1 with the patients divided into three cohorts for illustrative purposes (1992­
97, N=1,846; 1998­2003, N=4,528; 2004­10, N=9,230). Similarly, median survival increased 1992 through 2010 (14
vs. 15 vs. 18 months for the cohorts, p< 0.0001). However, 5­year survival rates did not change 1992 through 2010
(14.4% vs. 15.2% vs. 17.0% for the cohorts; p = 0.07). More patients (p= 0.007) and relatively more patients (p=
0.004) underwent resections of Stage I and Stage II cancers 2004 through 2010 with commensurately smaller
tumors (p= 0.01).
Conclusions: From 1992 through 2010, progressively more patients underwent pancreatectomy for pancreatic
adenocarcinoma with progressively smaller tumors and earlier stages. These patients lived more years (e.g.,
improved survival curves and median survival) but without improved 5­year survival, denoting better early and
intermediate survival. Early detection, better perioperative care, more efficacious noncurative chemotherapy
undoubtedly play a role, but better solutions for long­term survival must be sought. P049 HOW MUCH SHOULD WE PAY TO MINIMIZE PANCREATIC LEAK: THE COST­EFFECTIVENESS OF
PASIREOTIDE IN PANCREATIC RESECTION De Abbott, Jm Sutton, Pl Jernigan, A Chang, P Frye, Mj
Edwards, Sa Shah, Dp Schauer, Mh Eckman, Sa Ahmad, Jj Sussman; University of Cincinnati, Cincinnati, US
Introduction/Background: Pasireotide has recently been shown to decrease leak after pancreatic resection. However, the significant cost of this drug may be prohibitive for universal use. We conducted a cost­effectiveness
analysis to determine whether prophylactic Pasireotide administration after pancreatic resection possesses a
reasonable cost profile while improving patient outcomes.
Methods: A cost­effectiveness model was constructed to compare Pasireotide administration after pancreatic
resection versus usual care. Probabilities of completion of Pasireotide therapy, grade 3 complication, and
readmission were populated from a recent randomized trial1. Hospital costs (direct and indirect) for pancreatic
resection at a university pancreatic disease center, any subsequent readmission, and Pasireotide were weighted and
matched to outcomes from the index study, and also adjusted for inflation. Sensitivity analyses were performed
around financially and clinically important components of the model.
Results: Without considering the cost of Pasireotide, prophylactic use of the drug saved an average of $8,109 per
patient. However, when the cost of Pasireotide was included in the analysis, the costs of perioperative care
increased from $42,159 to $77,202; this was associated with a 56% reduction in pancreatic leak rate (21.9%­9.2%). The resultant cost per leak avoided was $301,628. Threshold analysis demonstrated that for this intervention to be
cost neutral, either the purchase price of Pasireotide (base case $43,172) must be reduced by 92.3% or
reimbursement for Pasireotide must be $39,848. If, hypothetically, societal willingness to pay was $50,000 to avoid
the morbidity of a pancreatic leak, the cost of Pasireotide would need to be reduced by 77%, or reimbursed at
$9,930 per course by payers. Sensitivity analyses around variable perioperative mortality, pancreatic leak, and
readmission rates did not significantly alter model outcomes.
Discussion/Conclusions: Pasireotide, though clinically useful in reducing pancreatic leak rates and readmission
following pancreatic resection, is expensive. Our analyses demonstrate the cost per leak avoided is approximately
$300,000 based on retail Pasireotide costs. Aggressive pricing negotiation, payer reimbursement for the drug, high
volume use (to help recoup research and development expenses) and an agreement in the surgical community about
what the morbidity of a pancreatic leak is worth will ultimately determine the utility of Pasireotide in pancreatic
resection.
1Allen, PJ, et al. Pasireotide for Postoperative Pancreatic Fistula. N Engl J Med;370:2014­22
P050 LONG­TERM OUTCOMES FOLLOWING SELECTIVE APPLICATION OF LAPAROSCOPIC
PANCREATICODUODENECTOMY FOR PERIAMPULLARY MALIGNANCIES Daniel Delitto, MD, Casey
Luckhurst, BS, Brian S Black, BS, Thomas J George, MD, George A Sarosi, MD, Ryan M Thomas, MD, Jose G
Trevino, MD, Kevin E Behrns, MD, Steven J Hughes, MD; University of Florida, Gainesville, US
Background: Recent literature supports laparoscopic pancreaticoduodenectomy (Lap PD) as feasible, safe and
effective. However, data regarding patient selection criteria or evaluating long­term outcomes following selective
application of Lap PD are lacking.
Methods: Consecutive patients (11/2010 – 02/2014) who underwent PD by a high­volume pancreatic surgeon
experienced in Lap PD were reviewed. Exclusion criteria for Lap PD included: borderline resectable disease,
extensive prior surgery, aberrant anatomy, and BMI > 40.
Results: Of 121 identified patients, 77 patients underwent PD for adenocarcinoma of which Lap PD was offered to
57 (74%); 7 (9%) required intraoperative conversion to Open PD. Mortality (90­day) was similar between Open and
Lap PD groups, but increased in patients requiring conversion (P < .012). The Lap PD group had smaller tumors (2.5
+ 0.1 vs. 3.4 + 0.2 cm; P = .004) and a higher R0 resection rate (92% vs. 70%; P = .017). Lap PD also had
significantly reduced blood loss (240 + 40 vs. 700 + 110 mL; P < .001) compared to open PD. Median overall
survival (OS) was significantly higher in Lap PD (18.6 vs. 10.9 mo; P = .010). Significant predictors of OS on
multivariate analysis included poor tumor differentiation, nodal metastasis, margin­positive resection, blood loss, and
operative approach.
Conclusion: The selective application of Lap PD results in a high percentage of eligibility, an excellent R0 resection
rate, and a low conversion rate. Long­term oncologic outcomes are superior in the Lap PD cohort as compared to the
Open PD cohort. P051 MEDICAID BENEFICIARIES UNDERGOING COMPLEX SURGERY AT QUALITY CARE CENTERS:
INSIGHTS INTO THE AFFORDABLE CARE ACT E C Hall, MD, MPH1, C Zheng2, R C Langan, MD1, L B
Johnson, MD, MBA1, N Shara, PhD3, W B Al­Refaie1; 1MedStar Georgetown University Hospital, 2MedStar
Georgetown Surgical Outcomes Research Center, 3MedStar Health Research Institute, Washington, US
Background: Medicaid beneficiaries do not have equal access to high quality care including access to high volume
centers for complex surgical procedures. We hypothesize that access to surgical services for the Medicaid
population varies within the same hospital, with a large Medicaid gap between those receiving emergency general
vs. complex surgery.
Methods: Using the National Inpatient Sample 1998­2010, we identified hospitals that performed at least 10
pancreatic resections/year (as a proxy for complex surgery). We compared the percentage of Medicaid patients
receiving appendectomies vs. pancreatic resections within each hospital. Multivariable logistic modeling was used to
identify characteristics associated with increased Medicaid gap.
Results: 603 hospitals were included. Most were urban (98%), and teaching (88.1%). Median percentages of those
with Medicaid receiving appendectomy and pancreatic resection were 11.9% (IQR: 5.6%­19.8%) and 6.5% (IQR: 0­
15.3%) respectively. Teaching hospitals (OR=7.9, 95% CI: 1.1­58.5) and hospitals that performed ≥ 40 pancreatic
resections in the year (OR=2.1, 95% CI: 1.2­3.6) were more likely to have a higher Medicaid gap.
Conclusions: Gaps exist between the percentage of Medicaid patients receiving emergency general surgery vs.
more complex surgical care at the same hospital, and are particularly exaggerated in teaching hospitals and
hospitals with very high volume. While targeted increase in intra­hospital access for complex cases at these
institutions may serve as a potential inlet strategy, it is unclear if the current Medicaid expansion program will truly
increase access to specialized high quality surgical care.
P052 NO SUPERIORITY OF PANCREATICOGASTROSTOMY OVER PANCREATICOJEJUNOSTOMY IN THE
PREVENTION OF PANCREATIC FISTULA AFTER PANCREATICODUODENECTOMY: AN UPDATED
META­ANALYSIS OF RANDOMIZED CONTROLLED TRIALS Stefano Crippa1, Roberto Cirocchi2, Justus
Randolph3, Stefano Partelli1, Amilcare Parisi2, Alessandro Pucci1, Michele Pagnanelli1, Massimo Falconi1;
1Universita' Politecnica delle Marche, 2Universita' di Perugia, 3Mercer University, Atlanta, USA, Ancona, IT
Backgound: Recent meta­analyses of seven randomized controlled trials (RCTs) with 1121 patients showed that
pancreaticogastrostomy (PG) is superior to pancreaticojejunostomy (PJ) to decrease the rate of postoperative
pancreatic fistula (POPF) after pancreaticoduodenectomy. However these meta­analyses did not perform any
subgroup analysis of different anastomotic techniques (i.e. duct to mucosa PJ or PJ with invagination versus PG)
and few considered specifically the rate of clinically­significant POPF. Moreover, new RCTs have been published. A
new meta­analysis of RCTs comparing PJ and PG was carried out with these specific aims.
Methods: Meta­analysis of randomized controlled trials (RCTs) comparing PJ versus PG after
pancreaticoduiodenectomy. Primary outcomes were overall and clinically significant POPF. Comparisons included
PG versus all types of PJ, PG versus duct­to­mucosa PJ and PG versus PJ with invagination.
Results: Nine RCTs were identified including 1309 patients, 655 undergoing PG and 654 PJ. RCTs showed
heterogeneity regarding definitions of POPF, intraoperative and perioperative management (i.e. use of octreotide, use
of pancreatic stents), characteristics of pancreatic gland (i.e. rate of soft pancreas, main pancreatic duct size), and
surgical techniques. No statistically significant differences were found in the rate of overall and clinically significant
POPF, morbidity and mortality, rate of reoperation, rate of intra­abdominal sepsis when PG was compared with all
types PJ, with duct­to­mucosa PJ or with PJ with invagination.
Conclusions: Compared to previous meta­analyses, we found a loss of statistical significance in regard to the rate of
POPF between PG and PJ by adding only two RCTs with 188 patients to the 1121 patients of the seven RCTs
previously analyzed. Subgoup analysis did not show any difference as well. PG is not superior to PJ in the
prevention of POPF, postoperative overall morbidity and mortality. Although randomized and prospective, current
available RCTs have major methodological limitations with significant heterogeneity. Further RCTs with
standardized surgical techniques, perioperative management and definitions of POPF are warranted.
P053 PANCREATECTOMY SURGICAL SITE INFECTIONS: WHAT ARE THE RISK FACTORS? Gareth Morris­
Stiff1, Colin O'Rourke2, R Matthew Walsh1, Henry A Pitt3; 1HPB Surgery, Cleveland Clinic Foundation,
2Quantitative Health Sciences, Cleveland Clinic Foundation, 3Department of Surgery, Temple University, Cleveland,
US
BACKGROUND: The morbidity of pancreatectomy remains unacceptably high. Surgical site infections (SSIs)
account for approximately half of this morbidity, and the SSI rate following pancreatectomy is the highest among all
gastrointestinal operations. However, the risk factors for SSI following pancreatectomy have not been adequately
defined. Therefore, the aim of this analysis was to attempt to identify those factors associated with SSI occurring
after pancreatectomy.
METHODS: The American College of Surgeons­National Surgical Quality Improvement Program Participant Use
File was queried from 2011­13. Patients undergoing proximal, distal, and total pancreatectomy as well as
enucleation were included in the analysis. Multiple pre­ and intraoperative parameters were analyzed as potential risk
factors. SSIs were characterized as superficial, deep and organ space infections (OSIs). Multivariable logistic
regression models were performed to determine the risk factors for OSI and any SSI.
RESULTS: 16,664 patients were available for analysis. The superficial, deep and OSI rates were 7.6, 2.3 and 10.7
percent, respectively, for an overall SSI rate of 20.6 percent. Any SSI rates were significantly higher for proximal
resections when compared to distal ones or to other resection types (p<0.001). The SSI rates were similar for benign
and malignant pathology (malignant versus benign, p= 0.13). Risk factors for both OSI and any SSI included male
gender, Hispanic ethnicity, increased ASA class and BMI, Preoperative sepsis and operative time. Wound
classification was an additional risk factor for superficial SSI, and smoking was an independent predictor for deep
SSI.
CONCLUSIONS: This analysis documents that a) SSI rates following pancreatectomy are unacceptably high, b)
SSI rates are higher for proximal than for distal pancreatectomy c) and the only modifiable factors are preoperative
cholangitis and operative time. Future efforts to reduce SSIs should focus on antibiotic choice and redosing, wound
protection and operative technique.
P054 PANCREATICOGASTROSTOMY FOLLOWING PANCREATICODUODENECTOMY IS ASSOCIATED
WITH LOW RE­OPERATION AND PANCREATIC FISTULA RATES Jennifer K Plichta, MD, MS, Gerard Abood,
MD, MS, Eileen O'Halloran, MD, Sam Pappas, MD, Gerard Aranha, MD; Loyola University Medical Center,
Maywood, US
Introduction: Significant morbidity and mortality following pancreaticoduodenectomy (PD) has been attributed to the
potential development of a pancreatic fistula, which has been shown to be as high as 40% in some studies. Our aim
was to review the development of post­operative pancreatic fistulas from a consecutive series of PD with
pancreaticogastrostomy (PG) patients.
Methods: Retrospective review of a prospective database identified 435 patients who underwent PD with PG
between 1996 and 2013. Of these patients, pancreatic texture and duct size data were available for 239 patients.
Clinical and pathological data for this subset were reviewed, and statistical analyses using univariate models were
performed.
Results: The median age was 66±12 years, with 56% males. Median intra­operative blood loss was 750ml, intra­
operative transfusions was 0 units, operative time was 6h, and post­operative length of stay was 8 days. Pathology
revealed: 96 pancreatic cancers, 35 ampullary cancers, 14 duodenal cancers, 12 bile duct (CBD) cancers, and 82
other lesions (including IPMN, pancreatitis, neuroendocrine tumors, and others). The post­operative complication
rate was 36% (n=86), most commonly pancreatic fistula (11%, n=27), delayed gastric emptying (DGE, 10%, n=24),
and wound infection (4.6%, n=11). Three patients (1.3%) required re­operation (1 for a type C fistula and 2 for
bleeding), no associated deaths. There were 5 peri­operative mortalities (2%). The development of a post­operative
pancreatic fistula was associated with a soft pancreas and pancreatic duct size <3mm. Other factors, including age,
pre­operative weight loss, use of intra­operative octreotide, estimated blood loss, intra­operative transfusions, and
operative time, were not correlated with fistula formation. Specifically, the pancreatic fistula rate for patients with a
soft pancreas was 18%, compared to 6.5% in patients with a firm or hard pancreas. Patients with a pancreatic duct
<3mm had a 17% rate of fistula formation, compared to 2.3% in those with a duct >3mm. The rates of pancreatic
fistula formation were not significant for benign vs. malignant lesions in general. However, patients with duodenal
adenocarcinoma had a statistically significant higher fistula formation rate.
Conclusions: Our study demonstrates that PG appears to be associated with overall low fistula formation rates,
although the rates are increased with a soft pancreas, pancreatic duct size <3mm, and certain types of peri­
ampullary tumors. Overall, it is also associated with low post­operative mortality and re­operation rates, which are
higher when a pancreatic fistula develops.
P055 POSTOPERATIVE PAIN CONTROL IN ENHANCED RECOVERY AFTER SURGERY (ERAS)
PROTOCOLS FOR PANCREATIC SURGERY: THE ROLE OF CONTINUOUS LOCAL ANESTHETIC WOUND
INFILTRATION Fara Uccelli, MD, Maria Carla Tinti, Giovanni Capretti, Francesca Gavazzi, Barbara Fiore, Maria
Rachele Angiolini, Monica Caravaca Martinez, Marco Montorsi, Alessandro Zerbi; Humanitas Research Hospital,
Rozzano, Italy, Vimodrone, IT
Background
ERAS models have been implemented to reduce length of stay (LOS) and postoperative morbidity. Pain control is
essential in these settings and from the initial experiences epidural catheter was accepted as gold standard in
colorectal surgery. New pain control methods had been proposed. The aim of this study is to assess if an ERAS
protocol in pancreatic surgery could achieve its goals if pain control was provided by continuous local anesthetic
wound infiltration, a far less dangerous pain control methods.
Methods
From July 2013 to September 2014 162 consecutive patients underwent a standard pancreatic resection in our
center and were managed according to an ERAS program. The program includes specific pre,intra and postoperative
items such as fluid balance, early mobilization and feeding. Pain control was obtained through a multi­holed wound
catheter with naropin 0.2% at 10ml/h up to 72 postoperative hours in association with paracetamol and ketorolac if
not contraindicated. 30days after discharge mortality was collected, pancreatic fistula (PF) was defined following the
ISGPF definition, Clavien­Dindo classification was used to asses postoperative morbidity. Data are expressed as
percentage or median [interquartile range].
RESULTS
102 patients underwent pancreaticoduodenectomy (PD), 45 distal pancreatectomy (DP), 15 total pancreatectomy
(TP). Median age was 69 [15] years, BMI was 24.5 [3.9]. 15.1% were ASA 3­4. 79% patients had malignant
diseases. No preoperative fasting longer than 6 hours was indicated and only 1 patient had a preoperative bowel
preparation. Naso­gastric tube (NGT) was removed in 96.8% of cases before POD1. 15 (33%) DP were performed
laparoscopically. A wound catheter for continuous local anesthesia was placed in 150 (92.5%) cases and removed in
POD 3. 19 (11.7%) patients required postoperative use of opioids. Subcostal incision was performed in open
procedures.13.4% of patients had a mean pain numeric rating scale (NRS) up to 3 in POD1 and 8.3% in POD2.
Median NRS at rest was 2[3] in POD1, 0[2] POD2, 0[1] POD3. Wound infection rate was 7.4%. 71.8% of PD and
TP started oral feeding on POD4 and 71.1% of DP started oral feeding on POD3. 17.1% of PD and TP and 4.4% of
DP required repositioning of NGT. 92.4% stay more than 4 hours out of bed in POD3, 59% in POD2. Major
complications (Clavien score >II) occurred in 19 patients (11.7%). 9 patients underwent reoperation and only one
died (0.6%). Pancreatic fistula rate was 39%. Median length of stay was 12 [10] days for DP, 7 [3] for DP, 9 [3] for
TP. Patients were consider fit for discharge (median) POD 10 [9] after PD, POD 6 [2] after DP, 8 [2] after TP.
Median LOS for uncomplicated PD was 8 [3]. 30­day readmission rate was 4.3%.
CONCLUSIONS
Postoperative continuous local anesthetic wound infiltration as part of an ERAS protocol in pancreatic surgery
guarantee a good pain control. The use of this strategy allow the achievement of major postoperative goals, reducing
the LOS and maintaining a very low readmission rate, without using epidural analgesia, avoiding the danger related
to the placement procedure and management difficulties.
P056 PRE­DIAGNOSIS IMPAIRMENT IN ACTIVITIES OF DAILY LIVING PREDICTS WORSE OVERALL
SURVIVAL IN PANCREATIC ADENOCARCINOMA Clancy J Clark, MD, Pradeep Yarra, MD, Nora Fino, MS,
Rishi Pawa, MD; Wake Forest Baptist Health, Winston Salem, US
Background: Pancreatic ductal adenocarcinoma (PDAC) is a cancer of the elderly with a 5­year overall survival
(OS) of only 6.7 %. Poor functional status, such as the inability to perform activities of daily living (ADLs), typically
limits treatment options due to the high morbidity associated with current therapies. While impairment in ADLs is
frequently attributed to the burden of the cancer, we have limited understanding of how the impairment in ADLs prior
to diagnosis impacts cancer survivorship. The aim of the current study is to evaluate the effect of pre­diagnosis
ADLs on OS for patients with PDAC.
Methods: The 1998­2011 Surveillance, Epidemiology and End Results­Health Outcomes Survey linked database
(SEER­MHOS) was queried for patients with histologically confirmed PDAC who reported their ability to perform
ADLs within 24 months prior to diagnosis. Patients with any impairment in ADLs were compared with patients who
reported no ADL impairment. Overall survival was evaluated using Kaplan­Meier and Cox proportional hazard
regression.
Results: 377 patients (median age 77, 53.6% male) with PDAC were identified with 41% (n=155) reporting
impairment in ADLs. Median time of reporting ADLs was 11.4 (IQR 12.5) months prior to PDAC diagnosis. The
majority of patients presented with metastatic disease (50.7%, n=188) and only 15% (n=60) underwent pancreatic
resection. No long­term survivors were identified in the study cohort. Median OS was 4.5 months (IQR 7.5). OS
for patients with any impairment in ADLs prior to diagnosis was significantly worse than patients reporting no ADL
impairment, 3.4 months vs. 5.5 months, respectively (log­rank p=0.012). After adjusting for patient age, stage at
diagnosis, and resection, any impairment in ADLs within the 24 months prior to diagnosis was an independent
predictor of OS for pancreatic cancer patients (HR 1.3, 95% CI 1.1­1.6, p=0.013).
Conclusions: Impairment in ADLs is common among patients with PDAC and is associated with significantly
worsen overall survival. Adjunctive interventions focused on addressing ADL impairments may broaden treatment
options and thus improve clinical outcomes. P057 PRESENTATION, MANAGEMENT AND OUTCOMES OF PANCREATIC ADENOCARCINOMA AT A
VETERANS AFFAIR TERTIARY MEDICAL CENTER Ali Mokdad, MD2, David Kim, MD2, Sergio Huerta, MD2,
Mathew Augustine1, Alexandra Webb, MD2, Michael A Choti, MD1, Zeeshan Ramzan, MD2, Patricio M Polanco,
MD2; 2University of Texas Southwestern Medical Center/Veterans Affairs North Texas Health Care System,
1University of Texas Southwestern Medical Center, Dallas, US
Introduction: Pancreatic cancer is the fourth cause of cancer deaths in the US. Ductal pancreatic adenocarcinoma
(DPA) is by far the most common type of pancreatic cancer. It portends poor survival with only 15 to 20% of the
cases are resectable at the time of diagnosis. Previously published national data demonstrated low rates of
resection and use of multimodality therapy in DPA. The primary objective of this study was to report the
presentation, management and outcomes of DPA in a veteran patient population. Comparison with nationally
reported data of pancreatic cancer in the general population was a secondary objective.
Methods: All patients with diagnosis of pancreatic cancer from 2005­2010 were identified from the cancer registry of
a tertiary referral VA hospital. Chart review and retrospective data collection was performed. Descriptive data
regarding demographics, symptoms, tumor location, stage, and treatment were reported in rates and percentages.
Median overall survival (OS), 3­year survival and Kaplan­Meier curve analysis were used for assessment of
oncologic outcomes.
Results: During the study period, 116 patients were identified with a diagnosis of pancreatic cancer; 104 patients
were DPA and were used for our analysis. All patients were men with a median age of 65 years (range 46­86) and
predominantly white (65.4%). Most common symptoms at presentation were abdominal pain (58.3%), weight loss
(55.3%) and jaundice (45.6%). Tumor location was head (66%), tail (14.6%), body (9.7%) and body/tail (6.9%).
Initial staging distribution was: I­II/localized (n=20, 19.2.%), III/regional (n=15, 14.4%) and IV/metastatic (n=69,
66.3%). Of the 20 patients with localized disease, 12 (60%) underwent curative resection (9 Whipple procedures and
3 distal pancreatectomies), 3 were unfit for surgery, 1 was explored but not resected, 1 declined surgery, and in 3
the cause of no­resection was undetermined. Based on staging, the median OS was 8.6 (95%CI: 2.6 – 15.9), 6.8
(95% CI:3–9.3), and 3 (95% CI:2–5.1) months for localized (resected and not resected), regional and metastatic
disease respectively. Median OS of patients that underwent curative resection was 11.3 months (95% CI: 3.5 –
34.5) months with a 3­year survival of 16.7%. Only 39.3% (33/84) of patients with advanced disease (Stages III­IV)
received chemotherapy. Those had an improved median OS when compared to patients that did not receive
chemotherapy (6.8 months [95%CI: 5.3­9.7] vs. 2 months [95%CI: 1.2­2.5], p<0.001)
Conclusions:
Presentation and outcomes of DPA in our cohort of veteran patients are comparable with previously reported
national data for the general population. Pancreatic resection with curative intent is the only means to obtain long­
term survival but it was only feasible in a small subset of patients. Palliative chemotherapy improved overall survival
in patients with advanced stage DPA. Further investigation and strategies are needed to improve the rate of
resection in early DPA and increase the use of palliative chemotherapy in advanced stages of the disease.
P058 RELATIVE CONTRIBUTIONS OF COMPLICATIONS AND FAILURE TO RESCUE ON MORTALITY IN
OLDER PATIENTS UNDERGOING PANCREATECTOMY Nina Tamirisa, MD1, Abhishek Parmar1, Gabriela
Vargas1, Hemalkumar Mehta1, Elizabeth Kilbane2, Bruce Hall3, Henry Pitt4, Taylor Riall, MD, PhD1; 1UTMB
Galveston and UCSF East Bay, 2Indiana University Health, 3Washington University in St Louis; BJC Healthcare,
St Louis, MO, 4Department of Surgery, Temple University Health System, Philadelphia, PA, Houston, US
OBJECTIVE AND SUMMARY BACKGROUND DATA: For pancreatectomy patients, mortality increases with
increasing age. Our study evaluated the relative contribution of overall postoperative complications and failure to
rescue rates on the observed increased mortality in older patients undergoing pancreatic resection at specialized
centers.
METHODS: We identified 2,694 patients who underwent pancreatic resection from the ACS­NSQIP
Pancreatectomy Demonstration Project at 37 high volume centers. Overall morbidity and in­hospital mortality were
determined in patients <80 (N=2,496) and >80 (N=198) years old. Failure to rescue was the number of deaths in
patients with complications divided by the total number of patients with postoperative complications.
RESULTS: No significant differences were observed between patients <80 and >80 in the rates of overall
complications (41.4% vs. 39.4%, p=0.58). In­hospital mortality increased in patients >80 compared to patients <80
(3.0% vs. 1.1%, p=0.02). Failure to rescue rates were higher in patients >80 (7.7% vs. 2.7%, p=0.01). Across 37
high volume centers, unadjusted complication rates ranged from 25.0%­72.2% and failure to rescue rates ranged
from 0.0%­25.0%. Among patients with postoperative complications, comorbidities associated with failure to rescue
were ascites, COPD, and diabetes. Complications associated with failure to rescue included acute renal failure,
septic shock, and postoperative pulmonary complications.
CONCLUSION: In experienced hands, the rates of complications after pancreatectomy in patients >80 compared to
patients <80 were similar. However, when complications occurred, older patients were more likely to die.
Interventions to identify and aggressively treat complications are necessary to decrease mortality in vulnerable older
patients. P059 RISK FACTORS OF NEW­ONSET DIABETES MELLITUS AFTER PANCREATICODUODENECTOMY,
PAYING ATTENTION TO LONG­TERM MORPHOLOGICAL CHANGES IN THE REMNANT PANCREAS
Yusuke Iizawa, MD, Masashi Kishiwada, MD, PhD, Yoshinori Azumi, MD, PhD, Hiroyuki Kato, MD, PhD, Akihiro
Tanemura, MD, PhD, Yasuhiro Murata, MD, PhD, Naohisa Kuriyama, MD, PhD, Shugo Mizuno, MD, PhD,
Masanobu Usui, MD, PhD, Hiroyuki Sakurai, MD, PhD, Shuji Isaji, MD, PhD; Hepatobiliary Pancreatic and
Transplant Surgery, Mie University School of Medicine, Tsu, Mie, JP
Background/introduction: New­onset diabetes mellitus (DM) after pancreaticoduodenectomy (PD) is an important
problem. The aim of the present study is to examine the risk factors of new­onset DM after PD, paying attention to
remnant pancreatic volume (RPV) and postoperative pancreatic duct dilatation (PDD). Methods: Among 204 patients who underwent PD from 2007 to 2013, the subject was 83 patients who didn’t have
DM preoperatively, could have RPV measured by CT volumetry at 1 month and had been followed for more than 6
months. RPV had been serially measured from 1 to 36 months. RPV ratio was calculated by dividing RPV at 3
months or more by RPV at 1 month. PDD was defined as greater than two­fold the pancreatic duct diameter at 3
months postoperatively by CT. Postoperative new­onset DM was retrospectively determined when we found either
one of fasting blood sugar level of 126 mg/dl or more or casual blood sugar level of 200 mg/dl or more and HbA1c of
6.5% or more, or when DM treatment had been postoperatively introduced. According to new­onset DM, these
patients were classified into DM group (n=11) and non­DM group (n=72). Results: The median follow­up of 83 patients was 23.9 months. Preoperative diagnosis, portal vein / artery
resection, diameter of pancreatic duct (measured intraoperatively), pancreatic consistency and incidence of
pancreatic fistula were not significantly different between DM and non­DM groups. RPV at 1 month was not
significantly different between DM and non­DM groups: 16.5ml (3.1­42.3) vs. 21.5ml (3.1­68.1), p=0.301. However,
RPV at 36 months was significantly smaller in DM group than in non­DM group: 10.9ml (1.4­19.8) vs. 21.8ml (3.1­
47.6), p=0.04. In both groups, RPV was reduced at 6 months, followed by almost no changes. RPV ratio at 6
months and 36 months in DM and non­DM groups were 0.79+/­0.25 vs. 0.76+/­0.18 (p=0.738) and 0.65+/­0.32 vs.
0.82+/­0.18 (p=0.09). PDD was evaluated in 81 patients (97.6%), and 7 (8.6%) developed PDD. Incidence of PDD
was higher in DM group than in non­DM group: 45.5% vs. 2.9% (P<0.001). Multivariate analysis revealed that PDD
was a significant independent risk factor for new­onset DM (odds ratio 24.0, 95%CI 1.95­295.0, P<0.013). Risk
factors of PDD couldn’t be identified in univariate and multivariate analysis. Conclusion: New­onset DM after PD may be influenced by RPV at 36 months. PDD was a significant independent
risk factor for new­onset DM.
P060 SARCOPENIA ­ AN UNDERESTIMATED BUT IMPORTANT ADVERSE PROGNOSTIC FACTOR IN
PATIENTS UNDERGOING SURGERY FOR PANCREATIC DUCTAL ADENOCARCINOMA Klaus Sahora, MD,
Gregor Werba, MD, Dietmar Tamandl, MD, Irene Kuehrer, MD, Martin Schindl, MD, Michael Gnant, MD; Medical
University Vienna, Vienna, AT
Background: Pancreatic cancer (PC) is one of the most lethal malignancies, with a persistently poor prognosis over
the last decade. Preoperative predictors of poor outcomes are needed to facilitate counselling of patients diagnosed
with pancreatic cancer. Recently, it has been shown that body composition is prognostic of survival in several
malignancies, particularly the degenerative loss of skeletal muscle mass known as sarcopenia. The purpose of this
study is to investigate the impact of sarcopenia on survival in patients undergoing surgery for pancreatic ductal
adenocarcinoma.
Methods: Patient body­composition was analysed using routine preoperative staging CT scans. The relationship
between sarcopenia, nutritional status, clinic pathological factors and prognosis was examined.
Results: We retrospectively reviewed a series of 133 patients undergoing surgery for pancreatic cancer. The
presence of sarcopenia had a significant negative impact on short term (1­year survival 58% vs. 85%, p= .001) and
overall survival (15 vs. 20 months, p= .019). Moreover, overweight patients with sarcopenia had the poorest
prognosis, with a median survival of 11 months, whereas overweight patients without sarcopenia had a significant
favourable median survival of 25 months (p= .001). In addition, positive lymph nodes and poor tumour differentiation
were independent predictors in a multivariate analysis. Neither sarcopenia nor sarcopenic obesity was associated
with perioperative mortality or morbidity.
Conclusion: Sarcopenia is a strong independent predictor of survival after surgery for pancreatic ductal
adenocarcinoma, specifically the presence of sarcopenic obesity. Incorporating body composition parameters into
clinical decision­making may allow for a more adequate and personalized treatment of pancreatic cancer.
P061 SHOULD I STAY OR SHOULD I GO NOW: FACTORS INFLUENCING HIGH LENGTH OF STAY AFTER
PANCREATECTOMY Michal Radomski, MD, MS, Amer Zureikat, MD, J.Wallis Marsh, MD, Kenneth K Lee, MD,
Allan Tsung, MD, David Bartlett, MD, Herbert J Zeh, III, MD, Melissa E Hogg, MD; University of Pittsburgh,
Pittsburgh, US
Introduction:
In this healthcare climate, much scrutiny is being paid to cost, readmission, and length of stay (LOS). Complex
pancreatic surgeries have been associated with prolonged postoperative courses. Many studies have looked at
overall data and low outliers for pancreaticoduodenectomy. However, little exists on high outliers and factors that
contribute to prolonged stays. We sought to evaluate the contribution of social, preoperative, operative, and
postoperative patient related factors to extended LOS in major pancreatic resections.
Methods:
A retrospective review of a single institution’s pancreaticoduodenectomies (PD) and distal pancreatectomies (DP)
was performed from 6/2009 to 3/2014 for all pathologies and technical approaches. Interquartile ranges (IQR) were
calculated and the highest quartile was evaluated and compared to the lower three quartiles.
Results: A total of 350 PD (42% open) and 127 DP (21% open) patients were analyzed with a 3rd IQR of >14 and
>8 days respectively. Social factors including race, distance, insurance status, and marital status were not
significant for PD; however, divorce was associated with longer LOS in DP (p<0.0001). Pre­operative
characteristics of higher age (p=0.0003), age adjusted Charlson Comorbidity Index (p=0.002), body mass index
(p=0.01), and American Society of Anesthesia assessment (ASA, p=0.005) were associated with increased LOS in
PD; whereas, only higher ASA (p=0.0188) was associated in DP. Albumin, Ca19­9, previous abdominal surgery,
and neoadjuvant therapy were not significant for PD or DP. Increased operative time (p=0.009), blood loss (EBL,
p=0.03), and transfusion (p=0.03) all were associated with longer LOS in PD; however, only EBL (p=0.03) was
associated with longer LOS in DP. A trend toward more LOS outliers was seen in the open PD group (p=0.06)
compared with robotic, but not in the DP group (p=0.63). For the PD group, pancreatic fistula (p<0.0001), delayed
gastric emptying (p<0.0001), and pseudoaneurysm (p<0.0001) were associated with extended hospitalizations. No
specific post­operative complications led to increased LOS in the DP group but the high LOS group was more likely
to have had any complication (p=0.007) compared to the rest of the cohort. High LOS outliers were more likely to go
to skilled nursing facilities, acute care facilities, and rehab than home in both PD (p=<0.0001) and DP (p=0.005)
groups. The high LOS group had more readmission in the PD group (p=0.005) but not in the DP group (p=0.64).
Conclusion:
More patient pre­, intra­, and post­operative factors lead to high LOS outliers seen in the PD group than the DP
group; but the DP group also had a social factor associated with increased LOS. Pre­ and Intraoperative factors are
hard to change, but further subgroup analysis in the PD group looking at management of specific complications and
physician related factors may help identify better or sooner management to decrease LOS for these factors. P062 SMOKING NEGATIVELY AFFECTS OUTCOMES AFTER TOTAL PANCREATECTOMY WITH ISLET
AUTOTRANSPLANTATION William P Lancaster, MD, David B Adams, MD, Katherine A Morgan, MD; Medical
University of South Carolina, Charleston, US
Introduction
Selected patients with pain from chronic pancreatitis (CP) benefit from total pancreatectomy with islet
autotransplantation (TPIAT). Recent studies have implicated smoking as an etiologic factor in the development of
fibrosing chronic pancreatitis. The effects of smoking on outcomes after TPIAT have not been previously examined.
Methods
Evaluation of a prospectively­collected database of patients undergoing TPIAT from March 2009 to October 2014
was conducted, with particular attention to demographics, smoking status, and outcomes as measured by SF­12
quality of life (QOL). Two­tailed t­tests were used comparing continuous data and Fisher’s exact test comparing
categorical data. Results
141 patients (102 women,76%; mean age 40.4years) underwent TPIAT. Forty­four patients were smokers (31%).
Patients who smoked had lower preoperative physical quality of life (physQOL) than
nonsmokers(25.4vs29.2,p=0.03) as well as lower postoperative physQOL at 1 year(30.4vs37.3, p=0.003). The islets
harvested were significantly fewer in smokers (2781 vs. 3860 IE/kg, p=0.05). There was a trend toward lower
insulin independence rates in smokers (20,17,11% at 1,2,3 years postoperative vs 32,28,21%, p=0.10).
Conclusions
Smoking is associated with poorer outcomes after TPIAT, including lower physQOL and islet harvest. Smoking
cessation should be counseled and smoking might be considered a relative contraindication to TPIAT.
P063 STARTING A PANCREATIC SURGERY PROGRAM AT A COMMUNITY HOSPITAL: BUCKING THE
TREND Jeffrey M Hardacre, MD1, Siavash Raigani, BA2, John Dumot, DO3; 1University Hospitals Case Medical
Center, 2Case Western Reserve University School of Medicine, 3University Hospitals Ahuja Medical Center,
Cleveland, US
Background: Most literature suggests that pancreatic resections should be done by high­volume surgeons at high­
volume hospitals to optimize patient outcomes. However, patient preference and insurance requirements may
restrict hospital location. After careful planning, a high­volume pancreatic surgeon started performing
pancreatectomies at a community hospital.
Methods: During a 14­month period, 81 pancreatectomies were performed, 60 at an academic medical center
(AMC) and 21 at a 144­bed community, non­teaching hospital (CH). Patients were selected for surgery at the CH
based on insurance carrier, health status, anticipated difficulty of resection, and patient preference. The operations
performed at the CH were done with the help of a surgical resident; however all post­operative care was provided by
the attending surgeon alone. Sixty­day outcomes were recorded.
Results: There were no statistically significant differences between the AMC and CH with regard to the median age
of the patients (66 vs 59 years), the gender distribution (57% vs 62% female), or the median BMI (28 vs 25
kg/m2).There was a significant difference in the American Society of Anesthesiologists (ASA) class distribution
between the AMC and CH (1: 0% vs 5%, 2: 7% vs 24%, 3: 88% vs 71%, 4: 5% vs 0%, p=0.0039). For
pancreaticoduodenectomy (PD)/total pancreatectomy (TP) patients at the CH, length of stay (LOS) was significantly
less than for patients at the AMC. For distal pancreatectomy/splenectomy (DPS) patients at the CH, the shorter
LOS was clinically relevant, but not statistically significant. Major complications and readmissions tended to be
lower at the CH. Greater than 80% of patients with adenocarcinoma who were recommended to receive adjuvant
therapy started their treatment within 60 days of surgery.
Conclusions: With appropriate planning and careful patient selection, high­quality pancreatic surgery can be
performed at a community hospital by a high­volume pancreatic surgeon.
AMC (N=60)
CH (N=21)
p value
Operation
0.99
PD
33 (55%)
12 (57%)
TP
6 (10%)
2 (10%)
DPS
21 (35%)
7 33%)
Median LOS (days)
PD/TP
7
5
0.012
DPS
5
3
0.25
Mortality
PD/TP
0
0
n/a
DPS
1 (5%)
0
0.91
Major Complication
(Accordion ≥ 3)
PD/TP
10 (26%)
2 (14%)
0.48
DPS
6 (29%)
0
0.16
Readmission
PD/TP
7 (18%)
2 (14%)
0.76
DPS
6 (29%)
1 (14%)
0.45
P064 SURVIVAL AFTER DISTAL PANCREATECTOMY FOR PANCREATIC DUCTAL ADENOCARCINOMA:
A NATIONWIDE RETROSPECTIVE COHORT STUDY Thijs De Rooij, Bsc1, Johanna Tol1, Casper Van Eijck,
MD, PhD2, Djamila Boerma, MD, PhD3, Bert Bonsing, MD, PhD4, Koop Bosscha, MD, PhD5, Ronald Van Dam,
MD, PhD6, Marcel Dijkgraaf, PhD1, Michael Gerhards, MD, PhD7, Harry Van Goor, MD, PhD8, Erwin Van Der
Harst, MD, PhD9, Ignace De Hingh, MD, PhD10, Geert Kazemier, MD, PhD11, Joost Klaase, MD, PhD12, Quintus
Molenaar, MD, PhD13, Gijs Patijn, MD, PhD14, Hjalmar Van Santvoort, MD, PhD1, Joris Scheepers, MD, PhD15,
George Van Der Schelling, MD, PhD16, Egbert Sieders, MD, PhD17, Olivier Busch, MD, PhD1, Marc Besselink,
MD, PhD1; 1Academic Medical Center, 2Erasmus Medical Center, 3St Antonius Hospital, 4Leiden University
Medical Center, 5Jeroen Bosch Hospital, 6Maastricht University Medical Center, 7Onze Lieve Vrouwe Gasthuis,
8Radboud University Medical Center, 9Maasstad Hospital, 10Catharina Hospital, 11VU University Medical Center,
12Medisch Spectrum Twente, 13University Medical Center Utrecht, 14Isala Clinics, 15Reinier de Graaf Gasthuis,
16Amphia Hospital, 17University Medical Center Groningen, Amsterdam, NL
Introduction
Nationwide data on survival after distal pancreatectomy (DP) for pancreatic ductal adenocarcinoma (PDAC) with
predictors for survival are lacking.
Methods
Adult patients who underwent elective DP for PDAC in one of 17 Dutch pancreatic centers between January 1st
2005 and September 1st 2013 were analyzed retrospectively. Patients were excluded when DP was not the primary
procedure or the histopathological diagnosis was not PDAC. Primary outcome was postoperative survival.
Predictors for survival <1 year were identified.
Results
In total, 761 consecutive patients were identified, of whom 620 patients were excluded because DP was not the
primary procedure (n=124) or non­PDAC histopathology (n=496). Therefore, 141 patients (45% (n=63) male, mean
age 64 years) who had undergone DP for PDAC were included, with a median number of 7 procedures per center
[range 2 to 22]. Multivisceral resection was performed in 30% (n=43) and laparoscopic resection in 5% (n=7) of
patients. In­hospital mortality was 4% (n=6). Mean tumor size was 44 mm [SD 23]. R0 resection was performed in
50% (n=70), R1 resection in 45% (n=64) and R2 resection in 5% (n=7) of patients. After DP, 45% (n=63) of patients
received adjuvant chemotherapy without differences in age and ASA physical status between patients with and
without adjuvant chemotherapy. Median follow­up was 14 [IQR 8­27] months. Postoperative median survival was 17
[IQR 13­21] months. One­year, three­year and five­year cumulative survival were 63%, 28% and 17%, respectively.
Median survival of patients who had undergone multivisceral resection was significantly reduced compared with
patients who had undergone DP only (10 [95% confidence interval (CI) 6­14] months versus 22 [95% CI 17­27]
months, respectively; P < 0.001). Survival was 12 [95% CI 5­19] months after R0 and 8 [95% CI 3­14] months after
R1 multivisceral resection. After multivariable analysis, AJCC (6th ed.) T­category (odds ratio 2.83 [95% CI 1.41­
5.67]) and not receiving adjuvant chemotherapy (odds ratio 4.98 [95% CI 2.03­12.21]) were independent predictors
for postoperative survival <1 year. Sex, age, ASA physical status, multivisceral resection, tumor size, lymph node
ratio and resection margin did not predict postoperative survival <1 year.
Conclusion
In this nationwide series, DP for PDAC was associated with a similar in­hospital mortality, R1 resection rate and
postoperative survival as seen after pancreatoduodenectomy for PDAC. These findings highlight the need for focus
on adequate (neo­)adjuvant treatment strategies in these patients.
P065 SURVIVAL BENEFIT ASSOCIATED WITH ADJUVANT CHEMORADIOTHERAPY IN PANCREATIC
DUCTAL ADENOCARCINOMA. Patrick J Worth, MD, Erin W Gilbert, MD, Raphael El Youssef, MD, Charles R
Thomas, MD, Brett C Sheppard, MD; Oregon Health & Science University, Portland, US
Background: Given the poor long­term outcomes of resected pancreatic ductal adenocarcinoma (PDAC), the addition
of adjuvant therapy may potentially improve outcome. There is no clear consensus regarding the optimal adjuvant
therapy, which can consist of either chemotherapy­­typically with gemcitabine or fluorouracil­based regimens +/­
radiotherapy. Identifying which pts will benefit from adjuvant therapy and estimating the magnitude of this benefit
remains challenging. Our working hypothesis is that adjuvant therapy may be impacted by select pre­treatment and
treatment­related parameters. Hence, the specific aim was evaluate the impact of these parameters within our
dataset
Materials and Methods: This was a retrospective review of prospectively collected data from a multidisciplinary
pancreatic cancer tertiary referral center within a NCI­Designated Cancer Center. Pts that underwent
pancreaticoduodenectomy for pathologically confirmed PDAC between 2004­2012 were included. Complete 2­yr
follow up mortality data was available for all pts. Age, gender, body mass index, recent onset of diabetes, abdominal
pain, obstructive jaundice, smoking history, weight loss, height­adjusted pre­operative total psoas area (a surrogate
of sarcopenia), operative time, intraoperative transfusion, portal vein (PV) reconstruction, tumor size, positive lymph
nodes, margin status, pathologic grade, perineural invasion, lymphovascular invasion, stage, presence of PanIN,
hospital length of stay, and overall survival from surgery were analyzed.
Results: 155 pts met inclusion criteria for the study; at the time of analysis, 78.1%had expired. Median survival was
451 days for the cohort, 331 days for pts who received surgery alone (n = 50), 398 days for pts who received
adjuvant chemotherapy (n = 43), and 606 days for pts receiving adjuvant chemoradiotherapy (CRT) (n = 62).
Treatment groups differed in operative time, intraoperative transfusion, margin positivity, and length of stay. There
was no difference in age, gender, tumor size, grade, lymph node positivity, perineural invasion, lymphovascular
invasion, stage, smoking status, BMI, or presence of jaundice with respect to treatment group. Univariate analysis
of all covariates demonstrated a correlation between sarcopenia, PV reconstruction, # positive lymph nodes, margin
status, perineural invasion, pathologic stage, and adjuvant treatment (all p < 0.05). A multivariate logistic regression
model was generated from these covariates. Of these, only adjuvant treatment group predicted two year survival.
Conclusion: Pancreatic cancer remains a clinically significant challenge. While our data are limited by their
retrospective nature, they suggest that the successful completion of systemic chemotherapy with the addition of
radiotherapy for locoregional control may be of significant benefit for selected patients.
P066 THE EVALUATION OF PREOPERATIVE INFLAMMATORY MARKERS IN PREDICTION OF POST­
OPERATIVE COMPLICATIONS AND SURVIVAL AFTER PANCREATIC SURGERY FOR CANCER. Gennaro
Nappo, MD1, Julie Perinel, MD1, Tommasangelo Petitti, MD2, Michel El Bechwaty, MD1, Roberto Coppola, MD,
Ph, FACS3, Mustapha Adham, MD, Ph1; 1HPB Surgery, Edouard Herriot Hospital, Lyon, France, 2Public Health
and Statistics, Campus Bio­Medico University of Rome, 3General Surgery, Campus Bio­Medico University of
Rome, Lyon, FR
Background
Accumulating evidence in literature supports a positive relationship between inflammation and cancer development
and progression. Recently, some preoperative inflammatory markers, as Neutrophil­to­Lymphocyte Ratio (NLR),
Platelet­to­Lymphocyte ratio (PLR), C­Reactive Protein (CRP) and modified Glasgow Prognostic Score (mGPS),
have been proposed as prognostic factors for pancreatic cancer. Moreover, recent studies showed the role of the
inflammatory markers in the early phase of post­operative course in predicting Post­Operative Pancreatic Fistula
(POPF) and inflammatory complications. However, no study to date has analyzed the role of preoperative
inflammatory markers as predictors of complications after pancreatic surgery. The aims of this study were to
evaluate the role of pre­operatory inflammatory markers in prediction of post­operative complications and survival
after pancreatic surgery for cancer.
Materials and methods
All pancreatectomies (Pancreatico­Duodenectomy (PD), Distal Pancreatectomy (DP), Total pancreatectomy (TP)
and completion of pancreatectomy) performed for cancers between 2008 and 2014 at Eduard Herriot Hospital were
retrospectively evaluated.
Data regarding baseline inflammatory markers were retrospectively recorded. Continuous parameters were
categorized as follows: NLR < 5 or > 5, CRP < 5 or > 5 mg/L, mGPS 0 (CRP and albumin normal) or 1 (CRP > 10
mg/L and albumin normal) or 2 (CRP > 10 mg/L and albumin < 35 g/L). The overall morbidity, all post­operative
complications and overall survival were also collected.
The correlation between preoperative inflammatory markers with occurrence of post­operative complications and
overall survival was evaluated.
Results
168 pancreatectomies were included in the retrospective analysis (108 PD, 23 DP, 34 TP, and 3 completion
pancreatectomies). The overall morbidity was 76%, while the major morbidity (grade III­V according the Clavien­
Dindo classification) was 55.5%. Mortality rate was 7%.
The incidence of complications was not statistically related to the value of preoperative inflammatory markers
(76.8% and 73% in case of NLR < 5 and > 5, respectively, p= 0.84; 76% and 76% in case of CRP < 5 g/L and CRP
> 5 g/L, respectively, p = 0.99; 79% and 71.1% in case of mGPS = 0 and 1 or 2, respectively, p = 0.52; 75% versus
90.3% in case of PLR < 200 and > 200, respectively). Moreover, no statistical relationship was found between the
value of preoperative inflammatory markers with the occurrence of major complications and each single complication
(particularly, with POPF).
The median survival was significantly correlated to NLR and mGPS preoperative values (30 versus 16 months for
NLR <5 and >5, respectively, p < 0.001; 33 versus 17 months for mGPS = 0 and 1 or 2, respectively, p < 0.05), but
not to CRP and PLR values (33 versus 21 months for CRP <5 and >5, respectively, p = 0.22; 28 versus 29 months
for PLR < 200 and > 200, respectively, p = 0.59).
Conclusions
This study demonstrated that mGPS and NLR are useful prognostic markers after pancreatic surgery for cancer.
Moreover, the preoperative inflammatory markers do not seem to have a role in predicting the development of
postoperative complications. Further prospective studies with a large number of patients are needed.
P067 THE IMPACT OF MINIMALLY INVASIVE DISTAL PANCREATECTOMY ON 90­DAY READMISSIONS
AND COST: IS IT ANY BETTER THAN OPEN? Janak Parikh, MD, Sandeep Anantha Sathyanarayana, Scott
Bendix, MD, Michael J Jacobs, MD; Providence Hospital Medical Center, Southfield, US
Introduction: Introduction: Laparoscopic distal pancreatectomy (LDP) is commonly performed for lesions of
pancreatic body and tail. Surprisingly, recent literature suggests an increase in readmission rates after LDP, hence
potentially negating any gained length of stay (LOS) benefit compared to open distal pancreatectomy (ODP).
Therefore, we sought to examine readmission rates and total cost of LDP versus ODP at a high­volume community
hospital.
Methods: Between January 2003 to December 2013, 81 distal pancreatectomies were performed at a community
teaching hospital. A retrospective analysis on demographics, 90­day outcomes, readmission rates, length of stay
(LOS), and total cost were collected. Results: Eighty­one patients underwent distal pancreatectomy (41 open and 40
laparoscopic). Median age was 62 years. Two­thirds of patients were female. LDP had significantly shorter mean
operative time (150 vs. 183 minutes; p<0.01) and decreased blood loss compared to ODP (135 vs. 568 mL;
p<0.001). Table 1 compares tumor characteristics, LOS, readmission rates, and costs. Pancreatic fistula rates were
comparable with no Grade C fistulae in either group. Overall 90­day morbidity was lower in the LDP group with no
mortalities. The 30­day and 90­day readmission rate was lower in LDP; hence LDP has lower total hospital days.
The overall costs for both the index admission and the total hospital stay (including readmission) were lower for LDP
group.
Conclusion: LDP has significantly lower index LOS, fewer total hospital days and lower overall costs compared to
ODP. LDP should be the standard of care for amenable lesions in the body or tail of the pancreas.
P068 VASCULAR RESECTION IN THE SURGICAL TREATMENT OF PANCREATIC ADENOCARCINOMA;
EXPERIENCE OF A CENTER E. Vigia, MD, S. Corado, M Sobral, A Nobre, L Bicho, E Filipe, J Paulino Pereira, A
Martins, E Barroso; Hospital Curry Cabral ­ Centro hepatobiliopancreatico e Transplantação, Lisbon, PT
OBJECTIVE / INTRODUCTION ­ Pancreatic cancer is the fourth leading cause of death by tumor in developed
countries, mostly adenocarcinomas ductal (adcD). Surgical resection being the only potential cure, we present the
results of our center for the treatment of this pathology.
MATERIALS AND METHODS ­ We analyzed our series of patients undergoing surgery for pancreatic last 10 years
for the clinical and pathological variables relevant for the prognosis. We uses SPSS 20 for processing of the data.
RESULTS ­ Between January 2004 and December 2013, there were 593 surgeries ­ 208 per adcD the pancreas. Of
these, 180 (87%) were located in the head of the pancreas; 158 (75%) were resected, 123 (54%) of the resections
were Duodenopancreatectomias. It was necessary vascular resection in 39 patients (15%). The Overall Survival
(OS) was 27.5% and 9.7% at 3 and 5 years, and the disease­free survival 18.5% and 8.2% at 3 and 5 years,
respectively. In univariate analysis, influenced the SG the stage (p = 0.035), positive margin (HR 1.95; 95% CI 1.28
to 2.95), vascular resection (HR 1.87, 95% CI 1, 18 to 2.95), reintervention (HR 3.76, 95% CI 2.2 to 6.4), major
morbidity (HR 4.8; 95% CI 2.9 to 8.0) and adjuvant therapy (p = 0.002).
DISCUSSION ­ The pancreas adcD prognosis is very reserved, even with resectable disease. Factors such as the
profit margin need vascular resection, reoperation, major morbidity and the absence of adjuvant therapy may
influence survival.
P069 WHIPPLES IN OCTOGENARIANS: PATIENT SELECTION TRUMPS AGEISM Audrey E Ertel, MD,
Jeffrey M Sutton, MD, Koffi Wima, MS, Richard S Hoehn, MD, Syed A Ahmad, MD, Jeffrey J Sussman, MD,
Shimul A Shah, MD, MHCM, Daniel E Abbott, MD; Department of General Surgery, University of Cincinnati,
Cincinnati, US
Introduction: The average age at diagnosis of pancreatic cancer is 72 years, and resection is the only potentially
curative intervention. Debate continues regarding the safety of pancreaticoduodenectomy (PD) in aging patients,
particularly those > 80 years old. We hypothesized that PD in selected patients 80 years or older was associated
with an acceptable outcomes profile.
Methods: The University HealthSystems Consortium database was queried for all pancreaticoduodenectomies
between 2009­2013 (n=16,983). We utilized univariate, multivariate, and 1:1 propensity score matching (comparing
patients age > 80 to those age 70­79) analyses to determine the impact of increasing age on post­operative
outcomes.
Results: 1370 patients (8%) > 80 years of age underwent PD in this cohort. There was an even distribution between
genders, though elderly patients were more often white (83.6%, p=0.001). Center volume differed across age groups,
with patients > 80 years more likely to receive care at highest volume centers vs. lowest volume centers (28.8 vs.
21.3%, p=0.001). Univariate analysis showed small increases in mortality (4.0 vs. 2.3%, p<0.001), length of stay (10
vs. 9, p<0.001), and cost ($21,839 vs. $20,485, p<0.001) between patients > 80 and those aged 70­79. Multivariate
analysis confirmed increase in mortality (OR 1.90, 95% CI [1.36­2.67], p<0.002), length of stay (OR 1.09, [1.04­
1.13], p<0.001), and cost (OR 1.08, [1.03­1.14], p<0.001). After propensity scoring, the effects of age on mortality
(p= 0.089) and cost (p=0.089) were no longer significant while length of stay (p<0.001) and discharge to a
rehabilitation facility (p<0.001) remained different across the stated age groups.
Conclusion: For appropriately selected patients, these data show that PD in the elderly (> 80) is safe and associated
with a minimal increase in perioperative and post­discharge resource utilization. Age > 80 should not be a
contraindication to PD this is important to realize as both the elderly population and the incidence of pancreatic
cancer continue to grow. P070 LAPAROSCOPIC FREY PROCEDURE Igor Khatkov, Viktor Tsvirkun, Roman Izrailov, Ruslan Alikhanov,
Aleksey Andrianov, Pavel Tyutyunnik, Artur Khisamov; Moscow Clinical Scientific Centre, Moscow, RU
Aim: To demonstrate the first experience of laparoscopic Frey procedure.
Materials and Methods: From December 2012 to November 2014 laparoscopic Frey procedure were performed in
12 patients (9 male and 3 female) with chronic pancreatitis type C (classification of M.Buchler).
The age of the patients was 48.8±9.6 years. The average size of the pancreatic head was 34.1±8mm, the average
diameter of the main pancreatic duct was 7.6±2.3 mm. The procedures were performed through the 5 trocar
accesses. Intraoperativelly were used following instruments: harmonic scalpel, monopolar coagulation, 5­10 mm
trocars, Endo GIA Universal stapling system.
Operative technique: After the pancreas mobilization and visualization vena mesanterica superior the head of the
pancreas was stitched with the stay sutures on the border of resection. The main pancreatic duct was opened with
the unipolar coagulator or an active branch of the Harmonic scalpel. Ventral part of the head of pancreas was
resected. A side­to­side pancreaticojejunoanastomosis was formed with single­layer continuous sutures using
nonabsorbable materials. The pancreaticojejunoanastomosis was covered additionally with a strand of greater
omentum in nine cases.
Results: The average operating time was 445±67.6 minutes. Blood loss was less than 200 ml. Conversion was
required in two cases: in the first case due to the peritoneal comissures after laparotomy and in the second case by
virtue of the impossibility of main pancreatic duct. There were no deaths. Complications developed in the post­
operative period in 2 patients: grade II (gastrointestinal intraluminal hemorrhage from the
pancreaticojejunoanastomosis) and grade IIIA (intrabdominal fluid collection) (classification of Clavien­Dindo). The
average postoperative stay period was 9,3±3.3 days. The entire of the total amount of patients (12) had been pain
free after the surgical treatment during the period 1 ­ 20 month.
Conclusions: The first experience shows that laparoscopic Frey procedure for the chronic pancreatitis are safe,
feasible with an accessible short­time outcome.
P071 PANCREAS STUMP CLOSURE TECHNIQUE AFFECTS PANCREATIC FISTULA RATE AFTER
RADICAL DISTAL PANCREATECTOMY Roderich E Schwarz, MD; IUH Goshen Center for Cancer Care,
Goshen, US
Background: Pancreatic stump leak after distal pancreatectomy (DP) is a major determinant for impaired
postoperative recovery. Factors influencing pancreatic fistula (PF) occurrence after DP remain of interest.
Methods: Prospectively collected outcomes observed by a single surgeon in an academic surgical oncology
practice were examined, and clinicopathologic and operative predictors of PF were analyzed.
Results: Seventy­two of 285 pancreatectomy patients underwent DP (25%). There were 30 men and 42
women with a median age of 61 years (range: 18­85). Underlying conditions included cancer (58%) and benign
processes (42%). Resections were comprised of open DP (63%), laparoscopic DP (26%) and open multivisceral
resections (11%). Spleen preservation was accomplished in 74%, despite planned splenic vascular resection in 93%
of cases. Median operative blood loss was 250 ml, with 14% of patients receiving transfusions (50% for
multivisceral, 9% for pancreas only resections). The mean lymph node count for cancer patients was 17.1 (+/­ 10.4),
with a R1 resection rate of 16%. Of 20 cases with postoperative complications (28%, no death), 6 were PFs (8%, 2
grade A and 4 grade B). The median LOS was 6 days (4­24). PF rate was 2.1% after sutured fish­mouth closures,
but 21% after other techniques including stapling (p=0.007); no other variable was linked to PF occurrence. LOS was
linked to complications, resection extent, underlying malignancy and transfusions (all at p<0.02) but not to PF.
Conclusion: PF rate after DP in this experience is unaffected by splenic vessel resection but appears to be
minimized through a sutured fish­mouth closure technique.
P072 PANCREATIC LIPOMA: INNOCENT BYSTANDER OR PATHOLOGICAL PROCESS? Maxwell T
Fohtung, BS, Nicholas J Zyromski, MD, Kumar Sandrasegaran, MD; Indiana University School of Medicine,
Indianapolis, US
Introduction: Pancreatic lipomas are a rare and benign form of the mesenchymal neoplasms that make up 1%­2% of
all pancreatic neoplasms. They are often diagnosed incidentally during radiographic imaging and treatment in the 69
reported cases in the literature has mostly been conservative although there is no consensus on the histopathologic
significance of lipomas or whether surgical intervention is warranted. In this study, we describe 74 cases of
intrapancreatic lipomas from a single institution over a 12­year period to elucidate the natural history including
associated symptoms, the need for intervention and the potential for dedifferentiation to liposarcoma. Methods: In the period from January 2001 to December 2013, we selected patients over the age of 18 diagnosed
with pancreatic lipoma based on Ultrasound, CT scan and MR imaging. Clinical data was coupled with radiographic
images and reviewed. Age at the time of diagnosis, gender, presenting symptoms, location of lipoma in pancreas,
size on imaging, surgical intervention and pathological findings were evaluated. Results: Pancreatic lipomas were identified in 74 patients including 41 women and 33 men aged 31­88 (median age
64). Most of the neoplasms were located in the head of the pancreas (n=32) followed by the body (n=16), tail (n=16),
neck (n=6) and uncinate (n=6). The lipomas ranged in size from 0.1cm to 4.8cm. The majority of patients had no
symptoms at presentation (n=36) while others had abdominal pain (n=26), dysphagia (n=1), jaundice (n=1), pelvic
pain (n=1), chest pain (n=2), nausea/vomiting (n=2), dyspnea (n=1), constipation (n=1). One lipoma was
pathologically confirmed and no patients underwent surgical resection. Conclusion: Pancreatic lipomas are rare benign mesenchymal neoplasms with increasing incidence as the use of
radiographic imaging continues to expand. Most reported cases are managed conservatively with follow­up imaging
especially when the lipoma has well defined margins and causes no obstruction to the pancreatic duct or common
bile duct. Given the lack of consensus, questions remain about the histopathological significance concerning
dedifferentiation to liposarcoma and whether surgical resection should then be considered as an intervention. This
large single center study will help shed more light on these questions.
P073 TOTAL LAPAROSCOPIC PANCREATICODUODENECTOMY: A SINGLE – INSTITUTIONAL
EXPERIENCE Alessandro Paniccia, Richard D Schulick, MD, MBA, Barish H Edil, MD; Department of Surgery,
University of Colorado Anschutz Medical Campus, Aurora, CO, Aurora, US
Introduction:
Laparascopic pancreaticoduodenectomy represents one of the most advanced abdominal surgical procedures. Since
its first description in 1994 several techniques have been detailed in the literature, however a standard approach is
still lacking. Herein we present our initial experience with total laparoscopic pancreaticoduodenectomy (TLPD).
Methodology:
Retrospective review of all cases consecutively performed by two operators between January 2013 and May 2014 at
The University of Colorado. Results:
Twenty patients underwent TLPD and conversion to open procedure was required in 2 cases (10%). Median age at
diagnosis was 55.4 years (IQR 42.8­66.5). Operative characteristics and postoperative complications are
summarized in table 1. The operative time decreased from 353 minutes (IQR 320­421) in the first 10 cases to 323.5
minutes (IQR 272­379) in the second 10 cases (r2= ­6.7; p=0.012). The estimated blood loss decreased from 300 mL
(IQR 330–400) in the first 10 cases to 200 mL (IQR 100 – 500) in the second 10 cases (r2= ­8.5; p=0.544).
Conclusion:
Laparoscopic Pancreaticoduodenectomy is a challenging operation, which is not done in a high volume at most
centers. We present our initial experience as a new laparoscopic pancreas program. Our experience shows that
oncologic outcomes are acceptable in terms of margin and lymph node harvest. Minimally invasive pancreas surgery
can be done safely with comparable complications rates to the traditional open approach and with no mortality. The
perioperative outcomes are similar to the traditional approach and long­term benefits are likely comparable to those
seen with other laparoscopic abdominal operations.
Table 1. Operative characteristics and postoperative complications
VARIABLE
N 20 SURGICAL MARGIN
Negative R0
20 (100%)
NUMBER OF NODES HARVESTED Median (range)
16.5 (13­20.5)
OPERATIVE TIME (min)
Median (range)
340 (300.0 – 381.5)
EBL (ml)
Median (range)
300 (175 – 450)
PANCREATIC FISTULA GRADE
Grade A
7 (35%)
Grade B
2 (10%)
Grade C 1 (5%)
DELAYED GASTRIC EMPTYING (DGE)
Grade A
3 (15%)
Grade B
4 (20%)
Grade C 1 (5%)
BILE LEAK
3 (15%)
PSEUDOANEURYSM
Hepatic artery
2 (10%)
GDA
1 (5%)
CHYLE LEAK
1 (5%)
SSI TYPE
Superficial
2 (10%)
Deep
0
Organ Space
3 (15%)
P074 INTESTINAL BARRIER DYSFUNCTION IN AGEING ANIMALS WITH ACUTE PANCREATITIS:
INCREASED INTESTINAL INFLAMMATION? Marcel C C Machado, MD, PhD1, Fabiano Pinheira­Silva, MD,
PhD1, Debora G Cunha1, Denise F Barbeiro, PhD1, Ana Maria M Coelho, PhD2, Heraldo P Souza, MD, PhD1;
1Department Emergency Medicine, University of Sao Paulo, SP, Brazil, 2Department of Gastroenterology (LIM/37),
University of Sap Paulo, SP, Brazil, Sao Paulo, BR
Introduction/Background: Acute pancreatitis in elderly patients in spite of similar occurrence of local complications is
followed by a substantial increase in morbidity and mortality rates with a significant financial impact. The
mechanisms underlying this age related vulnerability remain unknown. Intestinal barriers dysfunction resulting in
bacterial translocation from the intestinal lumen to distant organs has been incriminating as the main cause of
infected necrotizing pancreatitis with increased mortality. The aim of the present study was to investigate if
intestinal barrier dysfunction could be related to an increased intestinal inflammation in aged animals.
Methods: AP was induced in male Wistar rats by an intraductal 2.5% taurocholate injection and divided in 2
experimental groups (20 rats each group) G­1 young ( 3 month old rats) and G­2 older (18 month old rats). Twelve
hours after AP fragments of distal ileum were collected for evaluation of the gene expression of Cycloxygenase­
2(COX­2), and tight junction proteins (JAM­A and Occludin) and determination of inflammatory mediators (TNF­alfa
and IL­10). Twenty four hours after AP induction pancreas tissue was collect in sterile conditions for bacterial
culture.
Results: It was observed an increased bacterial translocation in the group of aged animals with AP (p< 0.05). We
also observed an increase of intestinal Cox­2 gene expression in aged animals with AP when compared to young
animals (p<0.05). It was also demonstrated an increased intestinal levels of TNF­alfa (p<0.0001) and decreased
levels of IL­10 in aged animals when compared to young ones (p<0.0001). We also observed an increased gene
expression of tight junction proteins (JAM­A and Occludin) in young animal when compared to aged animals
(p<0.005).
Conclusion: These results suggest that intestinal dysfunction in ageing animal with AP is related to increased
intestinal inflammation and delay recovering of barrier breakdown. These results also suggest Cox 2 could be a
potential target for reduction or prevention of barrier dysfunction in AP.
P075 RECIPROCAL STIMULATION OF PANCREATIC ACINAR AND STELLATE CELLS IN A NOVEL LONG­
TERM IN VITRO CO­CULTURE MODEL Merja Blauer, PhD1, Matias Laaninen, MD2, Juhani Sand, MD, PhD2,
Johanna Laukkarinen, MD, PhD2; 1Tampere Pancreas Laboratory; Tampere, Finland, 2Department of
Gastroenterology and Alimentary Tract Surgery, Tampere University Hospital, Finland, Tampere, FI
BACKGOUND: We have previously developed culture techniques for long­term in vitro maintenance of mouse and
human pancreatic acinar cells and demonstrated with the former the possibility of acinar cell cryopreservation for on­
demand use. These methods have enabled us to set up a co­culture system in which interactions between acinar
cells and pancreatic stellate cells (PSCs) can be studied on a long­term basis in vitro.
METHODS: Acinar cells and PSCs were obtained from mouse pancreata by explant outgrowth in cell type­specific
media. Co­cultures were set up in acinar cell­specific medium in 24­well format. Acinar cells were seeded in the
wells and PSCs in a separate compartment in tissue culture inserts. After 4­day culture, acinar cells were analyzed
for basal and caerulein­stimulated amylase release and PSCs for collagen I and fibronectin expression.
RESULTS: The viability of both cell types was excellently maintained for a minimum of 4 days in co­culture. Co­
culturing caused stimulation of acinar cell basal amylase secretion 2­fold compared to acinar cell monoculture.
Further stimulation with 0.1nM caerulein was prevented in co­culture, while in monoculture the normal 2.4­fold
amylase release compared to basal secretion was seen. The low level of extracellular matrix protein expression in
PSC monocultures was markedly increased in co­cultures.
CONCLUSIONS: Humoral communication between acinar and PSCs in co­culture was shown to lead to their
reciprocal stimulation. With its two separable cell compartments our co­culture system provides a versatile in vitro
setting that allows independent analysis of both cell types.
P076 VITAMIN D INDUCES UP­REGULATION OF ITS COGNATE RECEPTOR AND INHIBITS
PROLIFERATION AND EXTRACELLULAR MATRIX PROTEIN EXPRESSION IN MOUSE PANCREATIC
STELLATE CELLS Merja Blauer, PhD1, Niina Ikonen, BS1, Juhani Sand, MD, PhD2, Johanna Laukkarinen, MD,
PhD2; 1Tampere Pancreas Laboratory; Tampere, Finland, 2Department of Gastroenterology and Alimentary Tract
Surgery, Tampere University Hospital, Finland, Tampere, US
BACKGROUND: Vitamin D is a pleiotropic secosteroid hormone with a wide range of homeostatic functions in bone
and extraskeletal tissues due to its ability to modulate master regulatory networks involving cell proliferation,
differentiation and apoptosis. Vitamin D exerts its effects via the nuclear transcription factor vitamin D receptor
expressed in target cells. The relationships between vitamin D and health and its potential therapeutic implications in
various proliferative disorders have raised considerable attention in recent years. Pancreatic stellate cells (PSCs)
have only recently been recognized as targets for vitamin D action and a regulatory role for vitamin D in fibrogenesis
has been suggested. Here, we investigated the expression of VDR in culture­activated mouse PSCs and studied
the effects of the biologically most active vitamin D metabolite 1,25(OH)2D3 on the proliferation and ECM protein
expression of PSCs in vitro.
METHODS: Mouse PSCs were grown on glass coverslips or 96­ and 6­well plates for immunocyto­chemical
staining and growth and protein analyses, respectively. The cells were exposed to 1,25(OH)2D3 concentrations
ranging from 0.1nM to 10nM for 7 days. Cell growth assessment was performed with colorimetric crystal violet
assay and protein­specific antibodies were employed to demonstrate VDR, fibronectin and collagen I in Western blot
and immunocytochemical analyses.
RESULTS: A low basal level of VDR expression was detected which was strongly induced in the presence of
ligand. Dose­dependent suppression of PSC growth by 1,25(OH)2D3 was observed, the mean percentages of
inhibition ranging from 24% at the physiologic 0.1nM concentration to around 60% at 10nM. At the clinically
achievable concentrations of 0.5nM and 1nM the mean percentages were 49% and 55%, respectively. Significant
48% and 40% reductions in fibronectin expression were observed at 0.5 and 1 nM 1,25(OH)2D3. Minor decrease in
collagen I expression was detected at concentrations at or above 5 nM. CONCLUSION: The effectiveness of physiologically and clinically relevant concentrations of vitamin D in
suppressing the activation state of PSCs suggests an important role for vitamin D in pancreatic tissue homeostasis
and proposes the pancreatic stroma as a potential target of vitamin D­based therapeutic modalities.
P077 A CASE STUDY OF SIBLINGS WITH HEREDITARY PANCREATITIS; OUTCOMES ARE SUPERIOR
FOR SIBLING WHO HAD TOTAL PANCREATECTOMY WITH ISLET AUTOTRANSPLANT Stefanie M
Owczarski, PAC, MPAS1, David B Adams, MD, FACS1, Jeffrey Borckardt, PHD2, Wendy Balliet, PHD2, Hongjun
Wang, PHD1, Katherine A Morgan, MD, FACS1; 1Medical University of South Carolina, Department of Surgery,
2Medical University of South Carolina, Department of Psychiatry and Behavioral Sciences, Charleston, US
Introduction: Patients with hereditary pancreatitis suffer debilitating pain from an early age. They experience a
progressive decline in their quality of life (QOL) and ongoing parenchymal fibrosis often resulting in exocrine and
endocrine pancreatic insufficiency. In this matched case controlled study, we evaluated long term outcomes in a
two­patient brother and sister cohort with the PRSS1 genetic mutation. The brother was managed with total
pancreatectomy with islet autotransplantation (TPIAT). The sister was managed with lateral pancreaticojejunostomy
(LPJ).
Methods: We evaluated daily oral morphine equivalents, insulin requirements, pain score, physical QOL (pQOL),
and mental health QOL (mhQOL) over a 4­6 year time­period in order to compare TPIAT to LPJ. Both patients had
similar morphologic changes in the pancreas with dilated duct calcific fibrosis.
Results: The brother underwent TPIAT at age 14. His daily opioid requirement prior to surgery was 150mg. At 4­
year follow­up at age 18, he was insulin and opioid independent. His average daily pain score decreased from 5 pre­
op to zero post­op, pQOL increased from 9 to 56, and mhQOL increased from 44 to 61.
The sister underwent LPJ at 11 years of age. Her daily opioid requirements increased from 660mg at age 14 to
1,105mg at age 20. Her average daily pain score averages 8­9/10, pQOL ranges from 20­31, and mhQOL ranges
from 38­51.
Conclusion: TPIAT is the treatment of choice for adolescent patients with hereditary pancreatitis. Outcomes with
TPIAT are superior to LPJ as measured by opioid utilization, analog pain scores, and mental and physical QOL.
P079 DIFFERENCES IN IMAGING MODALITIES IN THE EVALUATION OF GALLSTONE PANCREATITIS
Naeem Goussous, Hadia Maqsood, Charlotte Horne, Guneet Kaur, Lisa Setiawan, Amanda Sautter, Stephanie
James, Hamid Ferdosi, Anne Sill, MSHS, Gopal C Kowdley, MD, PhD, FACS; Saint Agnes Hospital, Ellicott City,
US
Background: Acute pancreatitis is a common reason for inpatient admissions in the United States with around
300,000 admissions annually. Gallstones are the most common etiology for acute pancreatitis, and may be
associated with a dilated common bile duct (CBD). CBD diameter is commonly assessed with ultrasound (US),
computed tomography (CT) and magnetic resonance imaging (MRI).
Methods: Retrospective review of a prospectively maintained database of patients who underwent cholecystectomy
at our institution from 2000 to 2013 was performed. Patients with a diagnosis of gallstone pancreatitis were
identified. Statistical analyses were performed with SPSS (IBM Corp, Chigaco). A paired samples test was used to
compare measurements of CBD diameter by US, CT, and MRI. P­value <0.05 was regarded as significant.
Results: Among 6790 patients undergoing cholecystectomy, 337 (5%) were diagnosed with acute gallstone
pancreatitis. Most patients (296; 88%) has an US documented in the medical record, while half (168; 50%) had a
CT, and a third (111; 33%) an MRI. Of the 296 who underwent a preoperative US, 152 (51.3%) also underwent CT
and 101 (34.1%) also underwent MRI. The diameter of CBD was significantly larger by CT compared to US (10.1
vs 6.8, P = 0.01) in patients imaged with both modalities. CBD diameter was similar by MRI and US (6.6 vs 6.2,
respectively, P = 0.41), and by MRI and CT (7.0 vs 10.1, respectively, P = 0.1).
Conclusion: CT scan may overestimate the diameter of the CBD compared to MRI and US in the setting of acute
gallstone pancreatitis, and should therefore be interpreted with caution.
P080 PANCREATIC NECROSIS: A SINGLE INSTITUTION’S REVIEW OF PRACTICAL ADHERENCE TO A
STEP­UP APPROACH Stephanie Downs­Canner, MD, Brian A Boone, MD, Jennifer Steve, BA, Amer Zureikat,
MD, Kenneth K Lee, MD, Herbert J Zeh, MD, Melissa Hogg, MD; University of Pittsburgh Medical Center,
Pittsburgh, US
Introduction: Level 1 evidence demonstrated the “step­up” approach to infected pancreatic necrosis improves
outcomes. We examined our institution’s compliance with this approach and its impact on outcomes.
Methods: We performed a retrospective review of the index admission of all patients with necrotizing pancreatitis
treated in 2012 at a tertiary referral academic medical center.
Results: 41 patients (median age 52, 75% male) were treated for necrotizing pancreatitis. 80% were transferred a
median of 3 days from presentation. By the revised Atlanta Classification, 23 (56%) had severe acute pancreatitis
and 19 (46%) had suspected/confirmed infected necrosis. Of those infected, 18 (95%) required pancreatic drainage
procedures; 10 (56%) underwent step­up (Table). Median time from presentation to first pancreatic drainage was 1
day (step­up=0.5 days versus surgery first=4.5 days). Five (50%) of step­up went on to surgery a median of 1 day
later. Nine (47%) had multi­system organ failure (step­up=30% versus surgery first=75%). Median length of stay
was 23 days for step­up versus 21 days for surgery first. ICU admission rate was 90% for step­up versus 80% for
surgery first. In­hospital mortality was 20% in step­up and 12.5% in surgery first. Overall survival for all necrotizing
pancreatitis patients was 78% and 71% of deaths were related to pancreatitis. Pancreatitis related readmission rate
was 59% (step­up=30% versus surgery=88%).
Conclusion: Step­up adherence was 56%; however, half of those went on to surgery. Further analysis is
necessary to determine if the subset of patients undergoing surgery first represent deviations from level 1
recommendations or medically appropriate deviations. P081 ROCK­STAR AND OTHER SHOCKING CAUSES OF PANCREATITIS?* Alain Abdo, Sarina Sachdev, Urvi
Shah, Gopal C Kowdley, MD, PhD, FACS, Steven C Cunningham, MD, FACS; Saint Agnes Hospital, Ellicott City,
US
Introduction/Background: Many prior studies have identified both common and rare causes of acute pancreatitis,
whose incidence is increasing. However, many cases of truly idiopathic pancreatitis exist, suggesting unrecognized
etiologies. Use of energy drinks is also increasingly common, and increasingly a health concern. Transcutaneous
electrical nerve stimulation (TENS) devices are employed at various intensities and frequencies for pain
suppression, a controversial use.
Methods: We have observed four cases of acute pancreatitis, notable due to a strong temporal association with
exposure to energy drinks and nerve stimulators, but to none of the common etiologic factors. Therefore, we
performed a search of the literature and FDA databases for precedent cases or for mechanistic bases. Results: A 46­year­old man on two occasions consumed Rockstar™ energy drink, both times followed by acute
pancreatitis requiring hospitalization. He recovered uneventfully. A 54­year­old female during a 10­hour road trip
wore a TENS device for mild chronic back pain. Shortly after arrival she developed massive necrotizing
pancreatitis, from which she failed to thrive. She required necrosectomy and is back to her healthy baseline as of 2­
year follow­up. A 30­year­old female with an implanted ENS device developed edematous pancreatitis requiring
hospitalization. She responded to medical therapy and recovered. No clear precedent cases were found in
PubMed and only scant, weak precedent cases were found in FDA databases. There was, however, an intriguing
literature in PubMed in support of mechanistic basis for these exposures playing a role in the etiology of pancreatitis.
Conclusion/Discussion: Energy drinks such as Rockstar™ and TENS devices may, we speculate, be predisposing
factors in the development of acute pancreatitis, but there is insufficient direct evidence to support causality.
Mechanisms for TENS­associated pancreatitis include neurogenic inflammation and primary sensory neurons
working as a common final pathway for pancreatitis. Rockstar™ ingredients include sugar, caffeine, ginseng,
Ginkgo, milk thistle, and guarana. While four pancreatitis cases were listed in FDA reports of patients taking
Gingko, little evidence exists associating these in combination. The four observations presented here, coupled with
the rising use of the offending products, are cause for concern and warrant further study. Such study may include
either cellular or animal models of pancreatitis using these potentially offending agents, analyses of large databases,
and the establishment of an international registry.
P082 USE OF STREPTOKINASE FOR ENHANCEMENT OF PERCUTANEOUS DRAINAGE OF
PANCREATIC NECROSIS: A DOUBLE BLINDED RANDOMIZED CONTROLLED TRIAL Rahul Gupta, MS,
Rajesh Gupta, MCh, Mandeep Kang, MD, Deepak Bhasin, Madhu Khullar, Rajinder Singh; Post Graduate Institute
of Medical Education and Research, Chandigarh, IN
Background
Percutaneous catheter drainage [PCD] has been used as one of the invaluable tool for drainage of acute necrotic
collections and walled off pancreatic necrosis in patients with acute necrotizing pancreatitis.Fibrinolytic agents like
streptokinase, urokinase have been used locally to facilitate percutaneous drainage of empyemas, liver abscesses.
This prospective randomized controlled trial has been designed to study the safety and efficacy of streptokinase in
enhancing percutaneous drainage of pancreatic necrosis.
Methods
All the patients of severe acute pancreatitis (SAP) managed with PCD in the Division of Surgical Gastroenterology,
Department of General Surgery and Department of Gastroenterology, PGIMER, Chandigarh from April 2013 to
December 2014 will be included in the trial. Currently, twenty seven patients have been recruited and randomized to
placebo group& streptokinase group.The response to treatment was assessed by serial measurements of APACHE
II score, Modified Marshall score [MMS], size of collections by CT/MRI abdomen, CRP, IL­6 and serum albumin
levels. Patients not responding to PCD underwent necrosectomy. The primary endpoints were sepsis reversal and
death. The secondary end points were streptokinase related complications, catheter related complications, need for
surgical necrosectomy, duration of hospital stay.
Results
Among the twenty seven patients recruited, twenty were males and seven female. The mean age was 35.9 years
[range 24­52]. The most common etiology was alcohol [65.38%]. The modified CTSI, APACHE II and MMS of the
two groups at randomization were similar. The primary end point [sepsis reversal] was seen in seven out of twelve
patients in placebo group and in four out of thirteen patients assigned to streptokinase group (p = 0.247). Twelve
patients expired, seven in streptokinase group and five in placebo group (p = 0.551). Major complications [bleeding,
enterocutaneous fistula] occurredin eleven patients, six in streptokinase and five in placebo group (p = 0.825). There
were no streptokinase related complications. Ten patients required surgical necrosectomy for worsening sepsis, five
in each group while two patients in placebo group required surgery for bleeding and two patients in streptokinase
group for colonic perforation. The mean duration of hospital stay after randomization in placebo group was 25.3 days
while it was 19.9 days in streptokinase group (p = 0.26).
Conclusion
This is first study of its kind to demonstrate use of streptokinase in pancreatic necrosis and its safety. Analysis of
the results on completion of this trial will give insight into its therapeutic efficacy. (Clinicaltrials.gov Identifier:
NCT01977118)
P083 EFFICACY OF MINNELIDE AND PACLITAXEL COMBINATION AGAINST PANCREATIC CANCER
Shrey Modi, MD, Kaustav Majumder, MD, Vikas Dudeja, MD, Sulagna Banerjee, PhD, Ashok Saluja, PhD;
University of Minnesota, Minneapolis, US
INTRODUCTION: Chemoresistance is a hallmark of pancreatic cancer. Targeting any single pathway leads to
adaptive upregulation of multiple other non­canonical pathways. Thus, therapies targeting multiple pathways
simultaneously are need of the hour, which can be achieved by effective combinations of anticancer agents.
Minnelide (triptolide prodrug, currently in phase I trials) and nab­paclitaxel have been shown to be effective against
pancreatic cancer in preclinical and clinical trials respectively. In the current study, we evaluated the efficacy of
lower doses of triptolide and paclitaxel and their combination in pancreatic cancer.
METHODS: In vitro: Highly aggressive metastasis derived pancreatic cancer cells (S2VP10) were treated with low
doses of paclitaxel (0­25nM), low dose triptolide (25nM) and a combination of both for 24­72h. Cell proliferation was
measured using ECIS. Cell viability was measured by WST­8 cell counting assay. Cell death by apoptosis, was
evaluated by measuring caspase 3/7 activity and proportion of cleaved PARP positive cells. Cell cycle analysis was
done using propidium iodide staining.
In vivo: In s/c model, 6­week­old athymic nude mice were injected with 1*106 S2VP10 cells in right flank. In
orthotopic model using a separate set of mice, 200,000 S2VP10 cells were injected into the tail of pancreas.
Animals were randomized in four groups (N=6) as: DMSO (vehicle); Minnelide 0.21 mg/kg/day ip; paclitaxel
10mg/kg/week in DMSO ip and combination of Minnelide and paclitaxel. S/c model was terminated after 21 days
and orthotopic model was followed as a survival study. Tumors were harvested and tissues were used for various
experiments.
RESULTS: In vitro, proliferation of pancreatic cancer cells was markedly inhibited by combination treatment as seen
on ECIS by increasing the fraction of cells trapped in G2/M phase of cell cycle by 3 fold as compared to paclitaxel
alone. Low doses of triptolide (25nM) potentiated apoptotic cell death induced by paclitaxel in S2VP10 cancer cells
as measured by cleaved PARP (48h): triptolide (25nM): 266 ±31%, paclitaxel (10nM): 318 ±70%, combination: 1268
±144% and caspase 3/7 activity (48h): triptolide (25nM): 116 ±22%, paclitaxel (10nM): 184 ±45%), combination: 798
±135%. All data are expressed as % of control, mean ±SEM.
In vivo: In s/c model, combination of low doses of Minnelide and paclitaxel inhibited tumor progression; tumor
volumes [data expressed as % of control group, mean ±SEM: Minnelide: 75.4 ±25 %, paclitaxel: 50.0 ±3%,
Minnelide + paclitaxel: 11.0 ±1%] were significantly reduced in the combination group. In metastasis model, median
survival [in days] of animals in different treatments groups was vehicle (13.0), Minnelide (20.5), paclitaxel (21.5) and
combination (all mice were alive). Besides decreasing tumor burden, combination treatment significantly reduced
cancer related morbidity by decreasing ascites and metastasis.
CONCLUSION: Combination of triptolide and paclitaxel at low doses has immense potential to emerge as novel
therapeutic strategy against pancreatic cancer.
P084 FACTORS ASSOCIATED WITH FAILURE TO REACH SURGICAL RESECTION IN PATIENTS
UNDERGOING NEOADJUVANT CHEMOTHERAPY FOR RESECTABLE AND BORDERLINE RESECTABLE
PANCREATIC HEAD ADENOCARCINOMA Ana L Gleisner, MD, PhD, Jennifer Miller, MD, Mura Assifi, MD,
Jennifer Steve, David L Bartlett, MD, Melissa E Hogg, MD, Herbert J Zeh, MD, Amer H Zureikat, MD; Division of
Surgical Oncology, UPMC, Pittsburgh, US
Introduction: Neoadjuvant chemotherapy (NAC) is being increasingly administered to patients with resectable and
borderline resectable pancreatic head adenocarcinoma (PDAC). A significant portion of these patients do not
undergo resection despite a lack of disease progression. We sought to determine the factors associated with the
inability to resect PDAC in patients receiving neoadjuvant therapy in the absence of disease progression.
Methods: Patients with resectable or borderline resectable (SSO/NCCN criteria) PDAC who received NAC at a
tertiary referral center were identified. Univariate (UVA) and multivariate (MVA) analysis was performed to identify
factors associated with failure to undergo surgical resection in the absence of tumor progression.
Results: Between 2005­2013, 188 patients underwent NAC for resectable or borderline resectable PDAC: 69.7%
proceeded to surgery, 18.1% had disease progression and 12.2% were not resected due to clinical deterioration or
death despite lack of progression on imaging. On UVA, age >70 years (RR 3.88; 95%CI 1.82­8.32), ECOG
performance status > 2 (RR 7.27; 95% CI 3.38­15.62), vessel involvement at baseline (RR 9.64; 95% CI 1.34­
69.38), ≥1 episode of cholangitis (RR 2.74; 95% CI 1.22­6.14) and hospitalization for any cause during
chemotherapy (RR 4.30 95%CI 1.86­9.94) were associated with a higher risk of derailment from surgery in the
absence of disease progression (all p<0.05). Other factors such as diabetes, BMI, baseline CA19­9, type of
chemotherapy (including modern regimens) as well as presence and type of biliary stent were not associated with
inoperability. On MVA, age >70 years old (OR 4.38; 95% CI 1.52­12.63), any hospitalization (OR 3.72; 95% CI 1.29­
10.69) and vessel involvement (OR 15.10 95% CI 1.84­124.05) remained independently associated with failure to
undergo surgical resection (all p<0.05).
Conclusion: In the absence of disease progression, age >70 years, borderline resectable disease and those requiring
any hospitalization during neoadjuvant treatment are at higher risk of not undergoing curative surgical resection after
receiving neoadjuvant chemotherapy for PDAC.
P085 IMPACT OF SARCOPENIA ON SHORT­ AND LONG­TERM OUTCOMES IN PATIENTS UNDERGOING
CURATIVE RESECTION FOR PANCREATIC ADENOCARCINOMA: A NEW TOOL Neda Amini, Rohan Gupta,
Georgios A Margonis, Yuhree Kim, Gaya Spolverato, Neda Rezaee, Matthew J Weiss, Christopher L Wolfgang,
Martin A Makary, Ihab R Kamel, Timothy M Pawlik; Johns Hopkins Hospital, Baltimore, US
Background: Sarcopenia, defined as loss of muscle mass, may be a more objective means to determine peri­
operative performance status. Traditionally, sarcopenia has exclusively been characterized using total psoas area
(TPA). Defining sarcopenia using only a single axial cross­sectional image may, however, be inadequate. We sought
to evaluate total psoas volume (TPV) as a new tool to define sarcopenia and compared TPV with traditional TPA.
Method: Sarcopenia was assessed in 763 patients who underwent pancreatectomy for pancreatic adenocarcinoma
between 1996­2014. Sarcopenia was defined as the TPA and TPV in the lowest sex­specific quartile. The impact of
TPA and TPV sarcopenia on overall morbidity and mortality was assessed using multivariable analysis.
Result: Median patient age was 67 years and 54.8% (n=418) was male. Median TPA and TPV were both lower in
women (506.3mm2/m2 and 22.4 cm3/m2, respectively) versus men (685.1mm2/m2 and 33.0 cm3/m2, respectively)
(both P<0.001). 192 (25.1%) patients had sarcopenia by TPA, while 152 patients (19.9%) had sarcopenia according
to TPV. Post­operatively, 369 (48.4%) patients had a complication. While TPA­sarcopenia was not associated with
the risk of morbidity (OR=1.06; P=0.72), sarcopenia defined by TPV was associated with a higher odds of
complications (OR=1.79; P=0.002). In fact, patients with sarcopenia according to TPV had a higher risk of a major
complication (Clavien grade ≥3)(22.4% vs. 15.1%; P=0.03) and a longer length­of­stay (10 days vs. 8 days;
P=0.002). On multivariable analysis, TPV­sarcopenia remained independently associated with an increase risk of
post­operative complication (OR=1.69; P=0.006). In addition, after controlling for competing risk factors, sarcopenia
defined by TPV was associated with a higher odds of long­term death (HR=1.46; P=0.006).
Conclusion: The use of TPV to define sarcopenia was associated with both short­ and long­term outcomes following
resection of pancreatic cancer. Assessment of the entire volume of the psoas muscle (TPV) may be a better means
to define sarcopenia rather than a single axial image (TPA).
P086 OUTCOMES FOLLOWING TREATMENT OF PANCREATIC ADENOCARCINOMA WITH SMA
INVASION Pragatheeshwar Thirunavukarasu, MD, Emmanuel Gabriel, MD, Boris Kuvshinoff, MD, Steven
Hochwald, MD, Steven Nurkin, MD; Roswell Park Cancer Institute, Buffalo, US
Background: Pancreatic adenocarcinoma (PDAC) invading the Superior Mesenteric Artery (SMA) is traditionally
considered unresectable, with surgery offered to select patients. We evaluated outcomes of surgery versus non­
operative management i.e., systemic therapy with or without radiation (CR).
Methods: Data for patients with SMA­invading PDAC without distant metastases was extracted from the National
Cancer Database, 1998­2006.
Results: Of 3,445 patients, 227 (6.9%) had surgery, specifically pancreaticoduodenectomies (73.2%), distal
pancreatectomies (14.5%) and total pancreatectomies (12.3%). The overall R0 resection rate was 43.7%. The 30­
day mortality was 7.0%, median hospital stay ­ 9 days, and 30­day readmission rate ­ 8.7%. In patients who
underwent surgery plus CR, the median OS was 21 months when margin­negative and 13.6 months when margin­
positive (p < 0.001). In contrast, among patients who underwent surgery alone, survival was uniformly poor
regardless of margin status (8.3 vs 6.7 months, p = 0.09). Patients with R0 resection plus CR had significantly
better OS compared to patients in whom surgery was not recommended or given CR alone despite surgery being
recommended (Figure 1). Among patients without comorbidities, R0 resection plus CR had significantly better
survival than CR alone (22.2 months vs. 11.4 months, p < 0.01). The operative mortality of patients with ≤ 1
comorbidity was lower than with ≥ 2 comorbidities (6.0% vs. 12.6%, p = 0.05)
Conclusion: For appropriate patients with SMA­invading PDAC, a combination of R0 resection with systemic
therapy offers the best outcome. Given the higher operative mortality, these operations should be offered selectively
in specialized centers.
Table 1
Surgery
Treatment undergone
Surgical
Margins
Median overall survival
(months)
P­value*
No
None
5.6
No
Chemotherapy (+/­ Radiation)
10.9
0.001
Yes
None
6.1
Yes
Chemotherapy (+/­ Radiation)
12.3
0.03
Yes
Surgery only
Negative
8.3
Yes
Surgery only
Positive
6.7
0.09
Yes
Surgery + Chemotherapy (+/­
Radiation)
Negative
21.0
Yes
Surgery + Chemotherapy (+/­
Radiation)
Positive
13.6
<0.001
Recommended
*compares the value of the line in which the P­value is listed to the line above
P087 COMPARISON OF PANCREAS­SPARING DUODENECTOMY (PSD) AND
PANCREATODUODENECTOMY (PD) FOR THE MANAGEMENT OF DUODENAL POLYPOSIS SYNDROMES.
Gareth Morris­Stiff1, Matthew Dong1, Noaman Ali1, Subhash Reddy1, Colin O'Rourke2, R Matthew Walsh1; 1HPB
Surgery, Cleveland Clinic Foundation, 2Quantitative Heath Sciences, Cleveland Clinic Foundation, Cleveland, US
Introduction: Duodenal adenomas are a common finding in patients with familial adenomatous polyposis (FAP) and
individuals with Spigelman stage IV adenomas are at high risk of developing duodenal carcinoma. Patients are
traditionally treated by pancreatoduodenectomy (PD) though an alternate approach is pancreas­sparing
duodenectomy (PSD). We report present a 22­year experience comparing PSD with PD for the treatment of
duodenal polyps in FAP.
Methods: A retrospective review was performed of a prospectively maintained departmental database to identify
patients undergoing PSD and PD for duodenal polyposis. Outcome measures compared included: blood loss;
complication rates (overall, pancreatic fistula, enteric/biliary leaks, delayed gastric emptying (DGS), surgical site
infections [SSI]); and the development of endocrine and exocrine dysfunction. Pancreatic leaks and DGE were
classified according to ISGPS criteria. Phone interviews were conducted to confirm current status of patient at
follow­up. Analysis was performed using R software and significance assumed at a 5% level.
Results: There were 39 patients in each group. Patients undergoing PSD were younger (52.6 versus 64.3 years;
p<0.001). There was predominance of females in the PSD cohort and of males in the PD group. There was no
difference in operative time of blood loss between the 2 groups. Furthermore, there was no difference in the overall
complication rate, the pancreatic leak rate, or the SSI rates (organ­space or superficial) between the 2 groups.
However, there were a greater number of clinically relevant grade B and C leaks in the PSD group (5 versus 2;
p=0.011) though this did not translate to an increased re­operation rate (8 versus 3; p0.37). There were no
perioperative deaths in either group, and only a single long­tem mortality in each, in neither case related to the
operative procedure. The prevalence of new­onset diabetes was higher in the PD group though not significantly so (5
versus 2; p=0.43) whilst pancreatic insufficiency requiring enzyme replacement was seen more frequently following
PD (12 versus 2; p=0.008).
Conclusion: PSD and PD are both acceptable options for the management of duodenal polyposis arising in FAP. For
PSD, the early complication rate is higher but with fewer long­term endocrine and exocrine sequelae than PD.
P088 PANCREATICODUODENECTOMY FOR PANCREATIC NEUROENDOCRINE TUMORS: ARE
COMBINED PROCEDURES JUSTIFIED? Cornelius A Thiels, DO, MBA, Kristopher Kroome, MD, Danuel V Laan,
MD, Jay R Bergquist, MD, Kristine Thomsen, Mark J Truty, MD; Mayo Clinic, Rochester, US
Introduction: The efficacy and outcomes of pancreatic resection for pancreatic neuroendocrine tumors (PNET) are
well established. Few data exist looking specifically at pancreaticoduodenectomy alone for PNET or combined with
vascular reconstruction and/or distant/adjacent organ resection. We aimed to clarify the outcomes of patients with
PNET undergoing PD alone or with combined procedures.
Methods: A retrospective review of all PD’s performed from 1998 to 2014 at a single institution was conducted.
Only pathologically confirmed PNET were included in the study. Patients were categorized into either standard PD
(SPD) or combined PD (CPD) with CPD defined as those patients undergoing concurrent vascular reconstruction
and/or additional organ resection in order to achieve complete tumor removal. Patient demographics, procedural and
pathological data, and perioperative and long­term survival outcomes were collected. Kaplan­Meier curves were used
to assess survival.
Results: We identified 107 patients who underwent PD for PNET. Median patient age was 57 years. Thirty patients
underwent CPD. This included 10 patients who underwent vascular resection and reconstruction, 21 patients
underwent additional organ resection for either synchronous metastasis or adjacent locoregional tumoral
involvement, and one patient with both. The 90­day perioperative mortality was 1.47% and 3.45% for SPD and CPD,
respectively (p = 0.51). Median follow­up was 37 months. The overall 5­year survival after PD for PNET was 80.9%
and 74.8% respectively; this was comparable (p=0.65) between groups (Figure 1). Overall 5­year survival was
94.9% for low­grade PNET vs. 43.4% for high­grade PNET undergoing PD with grade, not extent of resection, being
associated with poor survival (p<0.001)(Figure 2).
Conclusion: Combination PD with vascular reconstruction and/or additional organ resection for PNET appears
justified, particularly for patients with low­grade tumors. Tumor grade, rather than extent of resection, was more
associated with poor overall long term survival. The need for combinatorial procedures during PD is not a
contraindication alone for otherwise resectable patients with PNET.
Figure 1. Kaplan­Meier survival curves of SPD and CPD showing equivalent survival at 60 months.
Figure 2. Kaplan­Meier survival curves of PD for PNET by tumor grade showing worse survival at 60 months in high
grade tumors.
P089 SURVIVAL OUTCOMES AND TREATMENT FAILURE AFTER METAL BILIARY STENT AND OPEN
SURGICAL BILIARY BYPASS AMONG PATIENTS WITH ADVANCED PANCREATIC ADENOCARCINOMA
RECEIVING CHEMOTHERAPY Alessandra Storino, MD1, Rohan Maydeo, MD2, Ammara A Watkins, MD1,
Manuel Castillo­Angeles, MD1, William E Gooding, MS3, Tara S Kent, MD1, Mandeep S Sawhney, MD2, A. James
Moser, MD1; 1Institute of Hepatobiliary & Pancreatic Surgery ­ Beth Israel Deaconess Medical Center, 2Advanced
Endoscopy and Gastroenterology ­ Beth Israel Deaconess Medical Center, 3Biostatistics Department ­ University of
Pittsburgh Cancer Institute, Boston, US
Background: Recent modeling data indicate that rapid initiation of chemotherapy may improve survival among
patients diagnosed with advanced pancreatic ductal adenocarcinoma (PDA). Biliary obstruction, and associated
treatment­complications, may delay treatment and diminish survival. We hypothesized that self­expanding metal
biliary stents (SEMS) permit earlier initiation of chemotherapy by comparison to open surgical biliary bypass (OSBB)
with equivalent patency.
Methods: Retrospective analysis of 127 subjects with PDA and biliary obstruction diagnosed between 2003­2014. 65
subjects (OSBB) underwent surgical staging and OSBB for unresectable cancer or radiographically­occult
metastases. 62 subjects received SEMS for locally­advanced disease or radiographically­occult metastases
identified during surgical staging. Subjects with radiographically­detectable metastases were excluded. Dual primary
endpoints included overall survival and cumulative incidence of treatment failure for biliary obstruction.
Results: The OSBB and SEMS cohorts did not differ with respect to age, gender, Charlson comorbidity Index, tumor
size, or preoperative Ca19­9 (p>0.05). Although NCCN classification demonstrated a higher initial proportion of
resectable/borderline lesions in the OSBB (83%) vs. SEMS (45%) groups, the proportion of locally­advanced and
occult metastatic PDA in the two cohorts was similar after surgical staging (p=0.37). Median time to initiation of
chemotherapy was earlier in the SEMS (median 12 days, IQR 5­36) than OSBB group (47, IQR 32­63; (p=0.0033)
but had no impact on median overall survival (11 months in both cohorts, p=0.81). Allocation to stent or surgery was
not associated with overall survival (HR: 1.09, 95% CI: .62 – 1.95, p=.7604). Conversely, median time to treatment
failure was 15 months in the SEMS cohort and 29 months after OSBB (p=0.018, Figure 1) and was significant after
adjusting for the cumulative risk of death (p=0.0055). 24% of subjects required re­intervention after SEMS compared
to 8% after OSBB (p=0.014). The median Comprehensive Complication Index was 0 (0­26.2) after SEMS and 8.7 (0­
28.5) after OSBB and (p=0.1198), corresponding to longer median total hospital stay from intervention to death/loss
to follow up after OSBB (median 8 days, IQR 7­11) compared to SEMS (median 1 day, IQR 0­5; p < .0001). The
cumulative incidence of treatment failure for the two groups is presented below (Image). Conclusion: SEMS was associated with faster initiation of chemotherapy and reduced total hospitalization but had
no beneficial impact on survival compared to OSBB. OSBB demonstrated longer hospitalization but superior patency
that may become clinically­relevant with improving chemotherapy. We conclude that enhanced postoperative
recovery after minimally­invasive surgery warrants a prospective randomized clinical trial between SEMS and biliary
bypass with the dual endpoints of survival and incidence of treatment failure.
P090 THE CHARACTERIZATION AND PREDICTION OF ISGPF GRADE C FISTULAS FOLLOWING
PANCREATODUODENECTOMY Matthew T McMillan, BA1, Charles M Vollmer, MD1, Jeffrey A Drebin, MD,
PhD1, Michael H Sprys, MS1, Pancreas Fistula Study Group1, Stephen W Behrman, MD2; 1University of
Pennsylvania Perelman School of Medicine, 2University of Tennessee Health Science Center, Philadelphia, US
Introduction: ISGPF Grade C postoperative pancreatic fistulas (POPF) are the greatest contributor to major
morbidity and mortality following pancreatoduodenectomy (PD); however, their infrequent occurrence (~2% of all PD)
has hindered deeper analysis. This study sought to develop a predictive algorithm, which could facilitate effective
management of this challenging complication.
Methods: Data was accrued from 4,301 PDs, performed by 55 surgeons at 15 institutions worldwide (2003­2014).
Demographics, postoperative management, and microbiology characteristics of Grade C POPFs were evaluated.
ACS­NSQIP preoperative variables were compared between Grade C POPFs and a 427­case sample of non­Grade
C POPFs (including no­POPF and Grade A/B POPF cases) drawn from the overall cohort. Risk factors for Grade C
POPF formation were identified using regression analysis and subsequently validated using resampling
methodology.
Results: Grade C POPFs developed in 79 patients (1.8%). Deaths (90 Day) occurred in 2.0% (N=88) of the overall
series, with 35% (N=25) occurring in the presence of a Grade C POPF. A similar proportion of Grade C POPFs
resulted in death (37%). Reoperations were necessary 73% of the time, with 30% of these requiring multiple
reoperations. The rates of single and multi­system organ failure were 28% and 40%, respectively. Mortality rates
escalated with certain types of organ failure, but they were unaffected by reoperation(s) (Table 1). The median
number of complications incurred was four (IQR: 2­5), and the median duration of hospital stay was 32 days (IQR:
21­54). Grade C POPF treatment required extensive resources: antibiotics (96%), ICU use (82%), transfusions
(82%), and TPN (76%). Warning signs for impending Grade C POPFs most often presented on POD 6. Surgeons
indicated Grade C POPFs evolved from a Grade B POPF 56% of the time. The predominant genera derived from
cultures of these fistulas were: Enterococcus (42%), Staphylococcus (35.8%), and Candida (35.8%). Positive
Candida cultures were associated with a mortality rate of 50% (P=0.082). Adjuvant chemotherapy might have
benefited 56% of Grade C POPF patients, yet it was delayed or never delivered in 26% and 67% of patients,
respectively. Preoperative factors associated with Grade C POPF occurrence were identified (Table 2) and a
predictive model yielded an area under the ROC curve of 0.78 (95% C.I.: 0.71­0.84; P < 0.00001).
Conclusion: This multinational study represents the largest analysis of Grade C POPFs following PD. It
demonstrates that Grade C POPFs incur a severe burden on patients, with high rates of reoperation and infection,
while also potentially worsening overall survival by causing delay or complete omission of adjuvant therapy. The
preoperative identification of high­risk patients using the proposed risk algorithm may facilitate optimal management
and improve outcomes.
P091 200 ROBOT­ASSISTED PANCREATIC RESECTIONS Niccolò Napoli, Emanuele Federico Kauffmann, Sara
Iacopi, Francesca Costa, Fabio Vistoli, Ugo Boggi; Division of General and Transplant Surgery, University of Pisa,
Pisa ­ Italy, Pisa, IT
INTRODUCTION / BACKGROUNDS: Acceptance of laparoscopy for pancreatic resections, other than distal
pancreatectomy, was probably slowed by both the inherent technical limitations of laparoscopy and the lack of
agreed standards for subspecialty training in advanced laparoscopic techniques. The da Vinci Surgical System
(dVss) improves surgeon dexterity and could eventually permit safe laparoscopic pancreatectomy in a greater
proportion of patients. We herein report our initial experience in 200 consecutive patients.
METHODS: All patients met the general selection criteria for major laparoscopic surgery. Patients diagnosed with
pancreatic cancer, initially discarded, were progressively accepted in the absence of obvious vascular involvement
and if clear surgical margins were evident all around the tumor. Unfortunately, some eligible patients could not be
operated robotically because of the lack of timely availability of the dVss.
Data were prospectively entered into a database and retrospectively analyzed.
RESULTS: Between October 2008 and October 2014, 200 consecutive patients underwent robotic pancreatic
resection. There were 123 females (61.5%) and 77 males (38.5%), with a mean age of 58 years (range 21­84) and
mean BMI of 24.6 kg/m2 (range 16.8­43).
Conversion to open surgery occurred in 3 patients (1.5%), because of intollerance to pneumoperitoneum (n= 2) and
troublesome dissection (n= 1), despite 10 patients underwent en­bloc resection of portomesenteric vein, 1 en­bloc
resection of the celiac trunk, and 1 reconstruction of the splenic vein.
Pancreaticoduodenectomy (PD) was performed in 83 patients (41.5%), total pancreatectomy (TP) in 17 (8.5%),
distal pancreatectomy (DP) in 83 (41.5%), including 46 patients in whom the spleen and the splenic vessels were
preserved, enucleation in 12 (6%), and central pancreatectomy in 5 (2.5%). Mean operative time was 527 min (range
330­960) for PD, 584 min (range 390­800) for TP, and 291 min (range 130­540) for DP. Twenty percent of the
patients underwent associated surgical procedures, resulting in prolonged operative time. There was one death
within 30­days (0.5%), and a further death within 90 days (1%). Both deaths occurred after PD, but either patients
had received simultaneous resection of the portomesenteric vein. Complications were recorded in 60% of the
patients after PD, 65% after TP, and 58.2% of DP, but most complications were graded I or II according to
Clavien's scale. Pancreatic fistula occurred in 32.5 % of the patients after PD (Grade A: 13; 15.6%) (Grade B: 9;
10.8%) (Grade C: 5; 6.0%). There were 86 malignant tumors, and the mean number of examined lymph nodes was
30.
CONCLUSIONS: In selected patients robot­assisted pancreatic resections can be safely performed. The
generalizability of these results, however, remains to be established by enlarging the number of procedures.
P092 SHOULD ACUTE PANCREATITIS BE AN INDICATION TO RESECT IPMN? Jessica L Cioffi, MD, Se
Joon Lee, MD, Joshua A Waters, MD, C Max Schmidt, MD, Attila Nakeeb, MD, Michael G House, MD, Eugene P
Ceppa, MD, Nicholas J Zyromski, MD; Indiana University, Indianapolis, US
Introduction: Intraductal papillary mucinous neoplasms (IPMN) cause acute pancreatitis (AP) more commonly than
generally appreciated. The natural history of IPMN with AP is unclear, and whether an episode of AP should be an
indication for surgery in the setting of IPMN is controversial. We sought to determine the natural history of IPMN
causing AP.
Methods: 348 patients with pathologically proved, resected IPMN were analyzed. Patients with single versus
multiple episodes of AP were compared to determine clinical and pathological differences between these groups.
Results: 114 (33%) IPMN patients had at least 1 episode of AP. Among IPMN/AP patients, 22 (19%) had more than
one episode of AP. IPMN type (main duct, branch duct, mixed type), location (head versus body/tail), and size were
similar between patients with single versus multiple AP episodes. Duration of symptoms related to IPMN was
significantly longer for multiple AP patients (single episode 17 months, multiple episodes 37 months). Invasive
carcinoma was present in 13 (11%) of all IPMN/AP patients, and was more common in patients with multiple AP
episodes (23%) than those with a single AP episode (9%).
Conclusions: Acute pancreatitis occurs in 33% of patients with resected IPMN, 19% of whom had multiple AP
episodes prior to resection. Patients with multiple AP episodes were more likely to harbor invasive carcinoma
compared to those with a single episode of AP. These data support early resection of IPMN patients who develop
acute pancreatitis.
P093 A GENOME­WIDE LOSS­OF­FUNCTION CRISPR SCREEN TO IDENTIFY MECHANISMS OF
CISPLATIN­RESISTANCE IN PANCREAS CANCER Mathew M Augustine, John Mansour, MD, Adam Yopp,
MD, Patricio Polanco, MD, Sam Wang, MD, Matt Porembka, MD, Michael Choti, MD, Joshua Mendell, MD, PhD;
UT Southwestern Medical Center, Dallas, US
Despite our best attempts to treat pancreatic adenocarcinoma with a combination of surgery, chemotherapy, and
radiation most patients inevitably develop recurrent disease refractory to most available therapies. Patients and
oncologists are therefore resigned to sub­optimal treatment options, selected with minimal rationale for the affected
pathways and compensatory mechanisms involved in therapeutic drug resistance. Included among the agents used
in treating pancreatic cancer are platinum­based compounds. Understanding the mechanisms essential to platinum
resistance provides an opportunity to identify the molecular basis and signaling pathways contributing to treatment­
refractory disease with the ultimate goal of developing novel strategies to target that resistance in patients. The
recent introduction of the engineered clustered regularly interspersed short palindromic repeats (CRISPR)­Cas9
endonuclease technology provides a distinct advantage by targeting gene disruption at the level of genomic DNA
with minimal off­target effects. We describe the use of this cutting­edge genome­wide screening strategy based on
the CRISPR/Cas9 genome editing system to identify mechanisms of platinum resistance in the platinum­sensitive,
BRCA2 mutant, pancreatic cancer cell line, CAPAN­1. Our objective is to use this system along with a large­scale
human library targeting over 20,000 genes and microRNAs, to identify genes and regulatory pathways that result in
resistance to the effects of the platinum­based chemotherapy, cisplatin, in CAPAN1 cells. We expect to uncover
both universal mechanisms of platinum resistance as well as tumor­specific resistance pathways that will ultimately
be clinically validated in pancreatic cancer patients who become resistant to platinum­based therapy. The discovery
of these complementary targets and pathways could provide unique opportunities for clinical translation and rational
drug delivery. Validation of these targets will undoubtedly facilitate biomarker development and prognostic
stratification. Novel targets could also provide unique opportunities for future drug development and utilization of
these agents in combination with current chemotherapy. Results from this screen have the potential to bring us
closer to personalized therapeutic decision­making. P094 AURANOFIN AS A NOVEL CHEMOTHERAPEUTIC AGENT FOR PANCREATIC DUCTAL
ADENOCARCINOMA Mayrim V Rios Perez, MD, David Roife, MD, Bing Bing Dai, PhD, Jason B Fleming, MD;
Department of Surgical Oncology, University of Texas MD Anderson Cancer Center, Houston, US
Auranofin is an FDA approved gold compound used for rheumatoid arthritis that has been shown to have anticancer
properties by inducing apoptosis in lung and ovarian cancer. However, its role as a potential chemotherapeutic
agent in pancreatic ductal adenocarcinoma (PDAC) has not been studied. Despite currently available anticancer
agents such as Gemcitabine (GEM), PDAC remains to be a feared and deadly disease awaiting the establishment
of more effective drug therapies. We performed in vitro studies using established PDAC and MD Anderson patient
derived cell lines along with patient derived tissue slices to assess both cellular and tissue growth inhibition after
treatment with Auranofin only and GEM only versus control groups. Dose­response curves were used to determine
Auranofin and GEM IC50s for each PDAC cell line, and were compared to determine relative sensitivity and
resistance among cell lines. Four of these cell lines (SW1990, AsPc1, Miapaca­2, and MDA­Patc53) were selected
for further mechanism studies. Western blot assays were performed to determine protein expression of thioredoxin
reductase (TXNRD1), nuclear factor erythroid 2­related factor 2 (Nrf2), Poly(ADP­Ribose)polymerase (PARP) and
PARP cleavage within experimental and control groups. We found relative Auranofin resistance among two PDAC
cell lines (SW1990 and AsPc1) when compared to more sensitive cell lines (Miapaca­2 and MDA­Patc53).
Moreover, SW1990 and AsPc1 were found to have higher TXNRD and Nrf2 expression compared to Auranofin
sensitive counterparts. The opposite is true for apoptosis marker PARP and PARP cleavage. Auranofin treated
patient derived tissue slices also showed lower viability compared to untreated group. Results show that PDAC cell
lines expressing high TXNRD1 and Nrf2 activity had a lower degree of apoptosis after treatment with Auranofin
when compared to control. We can conclude that Auranofin decreases PDAC cell viability in vitro by induction of
apoptosis, but its mechanism is still unknown. Our findings suggest that Nrf2 and TXNRD1 could be possible
biomarkers for PDAC chemoresistance to this compound. Our study altogether demonstrates anticancer activity of
Auranofin in vitro in cancer cells and tissue slices, which could be a novel approach for PDAC treatment.
P095 BIOBANK OF PANCREATIC DUCTAL ADENOCARCINOMA ACHIEVED FROM HUMAN PATIENTS
AND TRANSPLANTED INTO IMMUNODEFICIENT MICE Eugenio Morandi, MD1, Michela Monteleone, MD1,
David Alessio Merlini, MD1, GianAndrea Vignati, MD1, Tiziana D'Aponte, MD1, Marco Castoldi, MD1, Maria Rosa
Bani2, Raffaella Giavazzi2; 1Eugenio Morandi Foundation for the Study and Treatment of Pancreatic Cancer,
2IRCCS­Mario Negri Institute for Pharmacological Research, Rho (milan), IT
BACKGROUND Pancreatic ductal adenocarcinoma (PDA) is a lethal disease that remains one of the most resistant
to traditional therapies. This tumor is the only human malignancy for which patients' survival has not improved
substantially during the past 20 years. Despite advances in the comprehension of the molecular mechanisms
underlying pancreatic carcinogenesis, current systemic treatments offer only a modest benefit in symptoms control
and survival. The poor clinical outcome in PDA is attributed to intrinsic chemoresistance and a growth­permissive
tumor microenvironment. This changes are maybe caused by the conversion from quiescent to activated pancreatic
stellate cells that drives the severe stromal reaction that characterizes PDA.
METHODS The aim of our study is to develop a panel of patient­derived PDA­xenografts that mimic biological
heterogeneity of human pancreatic cancer. Twelve patients affected by PDA underwent cephalic
duodenopancreatectomy according to literature. PDA­xenografts were transplanted, subcutaneously and directly in
pancreatic gland, in nude, NSG and SCID mice. Three PDA­xenografts (HUPA 4, 8 and 11) took root successfully
in NSG mice and were established in pancreas and in the subcute of these rodents for five times, until a biobank is
achieved. Each mouse was treated with a combination of Abraxane and Gemcitabine, given at the optimal dose and
scheduled (Abraxane iv (25mg/kg) + Gemcitacbine iv (150mg/kg) ] once/week: 2 weeks on plus 1 week of, is
comparable to the human dosing regimen). Each tumor was then analyzed by our pathologist.
RESULTS Pancreatic xenografts were histologically and pharmacologically similar to the corresponding patient’s
tumor. Each tumor showed robust desmoplastic stroma, rare mucosal glandular component and resistance to
combined therapy. Subcutaneous pancreatic grafts showed a prevalence of glandular and mucosal component, less
stroma and responce to therapy. Each tumor had a medium time of latency (time to reach 150mm3) of 50­60 days, a
time to reach 250mm3 of 60­70 days and a doubling time of 20 days.
CONCLUSIONS This biobank of patient­derived PDA­xenografts is useful to study the biology of PDA, identify
tumor­specific molecular markers and develop novel treatment modalities designed to reengineer the pancreatic
cancer stroma and render it permissive to agents targeting cell­autonomous events or to reinstate
immunosurveillance.
P096 CXCL10 WITHIN THE TUMOR MICROENVIRONMENT INDUCES GEMCITABINE RESISTANCE IN
PANCREATIC CANCER CELLS Daniel Delitto, MD, Chelsey Perez, Brian S Black, BS, Heather L Sorenson, BS,
Andrea E Knowlton, PhD, Song Han, PhD, Dongyu Zhang, PhD, George A Sarosi, MD, Lyle L Moldawer, PhD,
Kevin E Behrns, MD, Chen Liu, MD, PhD, Thomas J George, MD, Ryan M Thomas, MD, Jose G Trevino, MD,
Shannon M Wallet, PhD, Steven J Hughes, MD; University of Florida, Gainesville, US
Background: The systemic treatment of pancreatic cancer (PC) is hindered by the rapid development of
chemoresistance to current cytotoxic therapies. Mechanisms governing the development of chemoresistance remain
poorly characterized, particularly with respect to contributions from the tumor microenvironment. Thus, the goal of
this study was to identify novel mechanisms acting within the tumor microenvironment which lead to PC
chemoresistance.
Methods: Intratumoral soluble mediator concentrations from resected PC specimens (n=26) as well as supernatants
from co­cultures of primary tumor­associated pancreatic stellate cells (PSCs) and PC cells (n=12) were evaluated
using a panel of 41 growth factors, chemokines and cytokines. The effect of CXCL10, a highly expressed soluble
mediator during co­culture, on viability, proliferation, and apoptosis of PC cells was evaluated with and without
gemcitabine treatment. In addition, the contribution of CXCL10 on migration patterns of peripheral blood mononuclear
cells (PBMCs) was assessed.
Results: Co­culture of tumor­associated PSCs with PC cells revealed increased CXCL10 levels compared to either
cell type cultured alone. In addition, high intratumoral CXCL10 concentrations correlated with reduced overall
survival (HR 6.9; P = .006). While CXCL10 treatment had a small effect on the viability of PC cells, it led to
significantly increased PC cell viability in the presence of gemcitabine. Further, gemcitabine treatment induced the
expression of the CXCL10 receptor, CXCR3, and this induction of CXCR3 was associated with the absence of
apoptotic markers in PC cells. Finally, constitutive expression of CXCL10 by PC cells preferentially led to the
migration of regulatory immune cell subsets.
Conclusion: Paracrine CXCL10 signaling between stromal, PC and immune cells may be responsible not only for
chemoresistance to gemcitabine, but also the recruitment and potential polarization of regulatory immune cell
subsets in the pancreatic cancer microenvironment.
P097 DIFFERENT CHARACTERISTICS IN HORMONAL EXPRESSION BETWEEN PRIMARY PANCREATIC
NEUROENDOCRINE TUMORS (PNETS) AND METASTATIC SITES Hideyo Kimura, MD1, Takao Ohtsuka1,
Takaaki Fujimoto1, Kenjiro Date1, Taketo Matsunaga1, Yusuke Watanabe1, Koji Tamura1, Atsushi Abe2, Yusuke
Mizuuchi2, Yoshihiro Miyasaka1, Daisuke Yamada1, Hisato Igarashi3, Tetsuhide Ito3, Shunichi Takahata1, Yoshinao
Oda2, Kazuhiro Mizumoto1, Masao Tanaka1; 1Department of Surgery and Oncology, Graduate School of Medical
Sciences, Kyushu Univ., 2Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu Univ.,
3Department of Medicine and Bioregulatory Science, Graduate School of Medical Sciences, Kyushu Univ., Fukuoka,
JP
Background: PNETs are known to have heterogeneity in terms of the ability to produce multiple hormones, and we
have sometimes experienced different hormonal expression patterns between the primary tumors and metastatic
lesions. The aim of this study was to evaluate the heterogeneity of PNETs from the viewpoint of hormonal
expressions.
Method: Among 105 patients who underwent pancreatectomy for PNETs at our institution between 1987 and 2014,
20 patients with metastatic PNETs (6 gastrinomas, one insulinoma, one glucagonoma, and 12 non­functioning
PNETs (NF­PNETs)) were evaluated. Expressions of representative 4 hormones such as gastrin, insulin, glucagon,
and somatostatin were analyzed in both primary and metastatic lesions including lymph nodes in all of 20 patients
and hepatic metastasis in 7 patients (2 gastrinomas, and 5 NF­PNETs) by immunohistochemical staining.
Immunohistochemical results were considered as positive when the final score (FS) (= the intensity score (IS: 0­3) ×
the proportional score (PS: 0­4)) was 4 to 12.
Results: The concordance rate (CR) of the hormonal expression pattern between primary tumors and metastatic
lesions was 50%. Gastrin was positive in 8 of 20 patients, insulin in 3, glucagon in 3, and somatostatin in 10.
Positive expression of 2 or more hormones was found in 6 patients, and that of one or more hormones was found
even in 9 of 12 patients whose primary tumors were diagnosed as NF­PNETs. All 8 patients with gastrin­positive
primary tumors had gastrin­positive metastatic lymph node (CR 100%), while concordant hormone expression was
present in one of 3 with insulin­positive primary tumors (CR 33%), one of 3 with glucagon (CR 33%), and 2of 10 with
somatostatin (CR 20%). Two patients had somatostatin­positive metastatic lymph node in spite of negative
somatostatin expression in primary tumor. All 7 patients with hepatic metastasis had some hormonal expression in
primary tumors; gastrin in 4 patients, insulin in one, glucagon in one, and somatostatin in 3. On the other hand,
positive gastrin expression was found in hepatic metastasis in only 2 of 7 patients, one of whom had a gastrin­
negative primary tumor, and any hormonal expression was negative in hepatic metastasis in the remaining 5
patients.
Conclusion: It should be recognized that hormonal expressions are often different between primary and metastatic
lesions of PNETs.
P098 EMT­MARKER, TUMOR BUDDING AND COLLECTIVE MIGRATION ­ 3­DIMENSIONAL
RECONSTRUCTION OF THE INVASION FRONT IN HUMAN PANCREATIC ADENOCARCINOMA K C
Honselmann, MD1, P Bronsert, MD2, M Werner, MD2, M Pross, MD1, D Bausch, MD1, T Keck, MD1, U Hopt,
MD3, U F Wellner, MD3; 1Department of General Surgery, University Medical Center Schleswig­Holstein, Campus
Luebeck, Germany, 2Institute of Pathology, University Medical Center Freiburg, Germany, 3Clinic for General and
Visceral Surgery, University Medical Center Freiburg, Germany, Luebeck, DE
Introduction
Cancer cell invasion takes place at the cancer­host interface and is a prerequisite for distant metastasis. Ductal
adenocarcinoma of the pancreas (PDAC) is characterized by diffuse growth patterns and bad prognosis.
Mechanistic invasion and metastasis models based on experimental data for epithelial­mesenchymal transition
(EMT) postulate that single cells undergoing EMT detach from the main tumor mass to disseminate to metastatic
sites. This concept continues to evoke skepticism among clinical pathologists. The aim of this study was to gain
insights in the relationship of EMT, tumor budding and mode of cell migration at the microscopic level by
examination of tumor budding and expression of EMT markers by three­dimensional (3D) reconstruction of human
PDAC.
Methods
Serial tissue slices from resected human pancreatic adenocarcinoma were stained with Pan­Cytokeratin, E­Cadherin
and ZEB1. These slides were subjected to 3D reconstruction with FreeD Software and statistic analysis was
performed with Medcalc Software.
Results
Budding tumor cells display a shift towards spindle­like as well as rounded morphology. Spindle­like cells were rare,
but their proportion was significantly higher in tumor buds (5.6%) than in main tumor branches (0.6%). Rounded cells
were more frequent than spindle­like cells and also significantly increased in tumor buds (11.4% vs 3.2% in main
tumor branches, p<0.001). This is associated with decreased E­Cadherin staining intensity and a shift from
membranous to cytoplasmic staining as well as increased nuclear ZEB1 expression (22.5%) in tumor buds (vs.
3.2% in main tumor branches). Most tumor buds on 2D histological slides were artefacts when analyzed in 3D,
where they retained out­of slice connections with an adjacent main tumor branch. Single cell migration could not be
identified although 46932 cells were analyzed. All cancer cells were part of a cell­cluster.
Conclusion
Collective cancer cell invasion associated with signs of EMT in a small subgroup of invasive cells is the main
invasion­migration mode in human PDAC. Single cell migration seems to be exceedingly rare or absent.
P099 MUCINOUS CYSTIC NEOPLASM IN MALE PATIENT: A CASE REPORT Duangpen Thirabanjasak, MD;
Faculty of Medicine, Chulalongkorn University, Bangkokok, TH
Mucinous cystic neoplasm (MCN) is almost always arising in female. There are a limited number of articles reported
about MCN in male patients. The diagnosis is based on cystic lesion of pancreas, simple lining epithelium, and
typical ovarian stroma. This is the first case report of the mucinous cystic neoplasm of pancreas in male patient in
the archive. The mass is at the tail of pancreas. No communication between bile duct and the mass is present both
in imaging study and pathologic study. Mucosal epithelium demonstrates low grade dysplasia but no malignant
transformation. Ovarian like stroma is highlighted by immunohistochemical studies. Resection is complete.
Mucinous Cystic Neoplasm with no invasive malignant part is usually benign, but there is uncertain outcome of long
term outcome in male patients.
P100 NUTRITIONAL STATUS REPRESENTS A NEGATIVE PROGNOSTIC FACTOR FOR POSTOPERATIVE
COMPLICATIONS AFTER PANCREATICODUODENECTOMY Carla Cappelli1, Ugo Boggi2, Rosa Cervelli1,
Emanuele Federico Kauffmann2, Niccolò Napoli2, Andrea Morandi2, Carlo Bartolozzi1; 1Diagnostic and
Interventional Radiology, Department of Translational Research and New Technologies i, 2General and Transplant
Surgery, Department of Translational Research and New Technologies in Medicin, Pisa, IT
Background. Since the 1970s postoperative mortality after pancreaticoduodenectomy (PD) has decreased, but
morbidity still remains a critical problem. Classical prognostic factors are tumor type, tumor stage, surgeon
expertise, and hospital volume. Among newer factors obesity, fat distribution and cachexia seem to play an
important role, although conflicting results have been reported. Further, it is not defined how to measure body fat and
fat distribution.
The aim of our study was to evaluate if body fat and fat distribution, measured at preoperative CT exam, can predict
postoperative morbility and mortality.
Methods and Materials. Seventy­seven (43 females; 34 males) consecutive patients undergoing open (40) or robotic
(37) PD with (31) or without (46) vein resection, were examined. All patients had a pre­operative abdominal CT
examination at our Institution.
Preoperative parameters included sex, age, American Society of Anesthesiologists (ASA) score, and body mass
index (BMI). Total, visceral and abdominal wall subcutaneous fat volume were measured with a semi­automatic
algorithm (Synapse 3D Fat analysis) applied on unenhanced CT scan. Postoperative complications were graded
according to Clavien’s score.
Univariate and multivariate analyses were performed to identify the factors associated with the development of post­
operative complications.
Results. Sixty patients had malignant tumors and 17 benign pancreatic diseases. Mean visceral fat volume was
2479±1588 cm3 and 1734±1174 cm3 in patients with malignant and benign lesions, respectively; corresponding
mean subcutaneous fat volume was 2422±1106 cm3 and 2411±1073 cm3. Fat distribution was mainly visceral in
men (3180±1596; 58.9±10.5% of total fat volume), and subcutaneous in women (2687±1186; 64.5±11.1%).
Patients with high visceral fat or low subcutaneous fat volume had a greater Clavien’s score (p<0.0001) or multiple
post­operative complications (p<0.0001). Clavien’ score was not statistically related to BMI, robotic/open surgery,
and venous resection. Patients with low BMI, low visceral and subcutaneous fat, had an increased risk of
developing pancreatic fistula.
Overall number of complications and Clavien’s score were not influenced by tumor type. Tumor type, instead, was
associated with the risk of pancreatic fistula (p=0.0212).
Conclusions. Our study confirms that obesity and cachexia are associated with an increased risk of postoperative
complications after PD. Nutritional support could improve these figures.
P101 PATIENT­DERIVED XENOGRAFT MODELS FOR PANCREATIC ADENOCARCINOMA
DEMONSTRATE RETENTION OF TUMOR MORPHOLOGY THROUGH THE INCORPORATION OF MURINE
STROMAL ELEMENTS Daniel Delitto, MD, Kien Pham, PhD, Adrian C Vlada, MD, George A Sarosi, MD, Ryan M
Thomas, MD, Kevin E Behrns, MD, Chen Liu, MD, PhD, Steven J Hughes, MD, Shannon M Wallet, PhD, Jose G
Trevino, MD; University of Florida, Gainesville, US
Direct implantation of viable surgical specimens provides a representative preclinical platform in pancreatic
adenocarcinoma (PC). Patient­derived xenografts consistently demonstrate retained tumor morphology and genetic
stability. However, the evolution of the tumor microenvironment over time remains poorly characterized in these
models. This work specifically addresses the recruitment and incorporation of murine stromal elements into
expanding patient­derived PC xenografts, establishing the rapidity by which murine cells are integrated into networks
of invading cancer cells. In addition, we provide methodology and observations in the establishment and
maintenance of a patient­derived PC xenograft model. A total of 25 histologically confirmed pancreatic
adenocarcinoma specimens were implanted subcutaneously into NOD­SCID mice. Patient demographics, staging,
pathologic analysis and outcomes were analyzed. After successful engraftment of tumors, histologic and
immunofluorescent analyses were performed on explanted tumors. PC specimens were successfully engrafted in 15
of 25 (60%) of attempts. Successful engraftment does not appear to correlate with clinicopathologic factors or
patient survival. Tumor morphology is conserved through multiple passages and tumors retain metastatic potential.
Interestingly, despite morphologic similarity between passages, human stromal elements do not appear to expand
with invading cancer cells. Rather, desmoplastic murine stroma dominates the xenograft microenvironment after the
initial implantation. Recruitment of stromal elements in this manner to support and maintain tumor growth represents
a novel avenue for investigation into tumor­stromal interactions.
P102 PD­1+CD28H+ SELECTIVELY IDENTIFIES FOR TUMOR­REACTIVE T LYMPHOCYTES IN
PANCREATIC CANCER Michelle R Koenig, BS1, Alessandro Paniccia, MD, PhD1, Joshua T Byers, MS1, Nate
Kahn, PhD1, Alexander Cenciarelli Schulick1, Justin Merkow, MD1, Lieping Chen, MD, PhD2, Richard Schulick,
MD, MBA2, Barish Edil, MD1, Yuwen Zhu, PhD1; 1Department of Surgery, University of Colorado Anschutz Medical
Campus, Aurora, CO, 2Department of Immunobiology, Yale University School of Medicine, New Haven,
Connecticut, Aurora, US
Tumor­infiltrating lymphocytes (TILs) are enriched in tumor­specific T cells and have been used as a source for
adaptive cell therapy (ACT) of cancer. TILs are a mixture of different T cell subsets with both anti­tumor and pro­
tumor capacities. It is valuable to isolate potent tumor­reactive T cells while eliminating suppressor T cells from
TILs to maximize anti­tumor immune response. CD28H is a newly discovered co­receptor of the B7 family that
interacts with its ligand B7­H5 to costimulate human T cells. CD28H is constitutively expressed on native T cells. However, repetitive antigenic exposure induces the loss of CD28H expression in T cells. T cells that have lost
CD28H have the phenotypic characteristics of terminally differentiated and replicative senescent cells. Here we
characterized PBMCs and TILs from patients with pancreatic cancer based on the expression of CD28H and PD­1.
We found that a subset of T cells, PD­1+CD28H+ (DP) cells, are exclusively enriched in TILs. Based on
phenotypic and functional analysis, our preliminary results suggest that PD­1+CD28H+ cells represent a subset of
superior tumor­reactive cells in TILs. By targeting these DP cells by CD28H agonistic mAb and/or with a PD­1
blocking antibody, we are expecting to reprogram TIL functions in human pancreatic cancer.
P103 THE RNA BINDING PROTEIN­HUR­ IS A MASTER REGULATOR OF PANCREATIC CANCER CELL
METABOLISM Mahsa Zarei, PhD1, Fernando F Blanco, PhD1, Jonathan R Brody, PhD1, Laszlo G Boros, MD2,
Jordan M Winter, MD1; 1Department of Surgery,The Jefferson Pancreas,Thomas Jefferson University, 2David
Geffen School of Medicine,UCLA, Philadelphia, US
Introduction: Cancer cells reprogram cellular metabolism to 1) satisfy the demands of growth, and 2) to overcome
their harsh tumor microenvironment. To date, the molecular mechanisms of metabolic reprogramming in pancreatic
ductal adenocarcinoma (PDA) have not been elucidated. We recently demonstrated that the regulatory RNA binding
protein, HuR, binds to different mRNA transcripts that encode proteins central to metabolism, and regulates
their expression. Additionally, HuR silencing with small interfering RNAs (siRNAs) sensitized cells to
glucose deprivation, which is a hallmark of PDA. Herein, we demonstrate HuR’s importance for cell survival when
deprived of the other key nutrient in the tumor microenviroment­ glutamine. In addition, we utilized state­of­the­art
metabolomics profiling with isotope tracers to understand the mechanism of HuR­regulated metabolic
reprogramming. Methods: HuR expression was modulated using siRNA oligos (scrambled or specific against HuR) and cell death
was measured by trypan blue staining in two PDA cell lines (MiaPaCa2 and Panc1 cells) under normal tissue culture
(4mM glutamine) and low glutamine conditions (0.5 mM glutamine). Next, 13C­labeled nutrients (glucose and
glutamine) were added to the culture media under both normal and low glucose conditions. Gas
chromatography/mass spectrometry was performed in order to dynamically and precisely map the flow of carbon
through numerous metabolic pathways. Results from these metabolic flux experiments were compared in isogenic
PDA cells with normal and silenced HuR expression. The Pearson correlation between the levels of 13C­labeling in
metabolites under normal and low glucose conditions was determined. Results: HuR silencing was estimated to be >70% by immunoblot. In trypan blue cell death assays, HuR silencing
resulted in increased death with glutamine deprivation at 24 and 72 hours, in both tested cell lines (Figure). In the
metabolic pathway flux analysis, HuR silencing directly affected the flow of carbon from glutamine into numerous
metabolites including ribose (pentose phosphate pathway, Table, line 1), glucose (gluconeogenesis, line 2) and
myristate (fatty acid synthesis, line 3). Increased futile carbon exchange fluxes were prevalent , as evidenced by
increased lactate labeling (line 4). Conclusions: HuR regulates the flow of carbon directly from glutamine into numerous metabolic pathways, and
likely achieves this by regulating key metabolic RNA binding targets. HuR’s regulation of cellular metabolism likely
accounts for the RNA binding protein’s pro­survival effects on PDA cells under both glucose and glutamine
deprivation. These mechanistic insights provide a strong rationale to pursue pharmacologic inhibition of HuR (using
small compound inhibitors, Novartis) in combination with metabolic pathway inhibitors as a novel therapeutic
strategy to treat PDA.
Altered metabolites with HuR silencing in BxPC3 PDA cells using [U­13C5]­L­glutamine tracer. Metabolite
Metabolic Pathway
+HuR
­HuR
­HuR
(Normal
(Low Glucose)
glucose)
Pearson’s
Correlation
1 Ribose
Pentose phosphate pathway
100%
77.9
74.7
0.996
2 Glucose
Gluconeogenesis
100%
28.6
13.9
0.991
3 Myristate
Fatty Acid synthesis
100%
54.9
46.3
0.992
4 Lactate
Futile Carbon Cycling
100%
102
102.4
­0.990
P104 USING CYST FLUID PROTEASE ACTIVITY SIGNATURES TO DIFFERENTIATE BENIGN FROM
MALIGNANT CYSTIC PANCREATIC LESIONS (CPLS) Sam L Ivry, BS1, Kimberly S Kirkwood, MD1, Charles
Craik, PhD1, Dana Dominguez, BS1, Anthony O'Donoghue, PhD1, Randall E Brand2; 1University of California San
Francisco, 2University of Pittsburgh, San Francisco, US
Introduction/Background: With increasing use of abdominal MRI, CPLs are being detected at an alarming rate,
whereas biological indices that predict cyst behavior are insufficient to guide clinical decision­making. Previous
reports showed differential expression of individual proteases and protease inhibitors among cysts with varying
malignant potential. This candidate­based approach can both miss important proteases that were not the focus of
the study, and typically fails to account for the effects of endogenous protease inhibitors that are ubiquitous in the
pancreas. In this study, we characterized differences in protease activity using our Multiplex Substrate Profiling by
Mass Spectrometry (MSP­MS) technology, in which cyst fluid is screened against a physicochemically diverse
peptide library to generate a cleavage signature.
Methods: Pancreatic cyst fluid was collected during EUS or resection from 32 patients, all of whom underwent
pancreatic resection with pathologic evaluation. Protease activity signatures were determined using equal
concentrations of cyst fluid protein in our MSP­MS assay. This assay uses a library of 228 tetradecapeptides and
mass spectrometry to identify preferred protease cleavage sites. From the resulting cleavage pattern, we can infer
the relative activities of virtually all known proteases identified within a given sample.
Results: Cluster analysis of protease cleavage sites determined by MSP­MS revealed three distinct cyst
populations. Cysts with the highest risk of becoming cancerous showed the least overall proteolytic activity by
MSP­MS. Among IPMNs and MCNs, those with moderate to high grade dysplasia showed a preference for bulky
hydrophobic residues in the P1 position. This may reflect an abundance of chymotrypsin or overexpression of
pancreatic trypsin inhibitors. Carboxypeptidase activity was highly prevalent in pseudocysts and barely detectable in
MCNs and IPMNs. Pseudocysts retained a trypsin­like specificity and preferentially cut after arginine and lysine
residues.
Conclusions: IPMNs and MCNs with moderate/high grade dysplasia show unique cleavage specificity and are less
proteolytically active than benign/low grade cysts. This is in agreement with previous reports indicating that the
protease inhibitor SPINK1 is overexpressed in higher risk cysts. Following validation and refinement in a larger
cohort of patients, protease activity may be a useful tool for risk stratification among CPLs. MSP­MS is a powerful
tool for the comprehensive evaluation of proteolysis in complex clinical fluids.
P105 A RARE CASE OF AMPULLARY GANGLIOCYTIC PARAGANGLIOMA WITH LYMPH NODE
METASTASIS Jacob E Dowden, MD, Ramsay Camp, MD, Eric T Kimchi, MD, Katherine A Morgan, MD, David B
Adams, MD, Kevin F Staveley­O'Carroll, MD, PhD; Medical Univeristy of South Carolina, Charleston, US
Gangliocytic paraganglioma (GP) is an uncommon neuroendocrine tumor that is usually found in the periampullary
region of the second portion of the duodenum. They can be difficult to accurately diagnose preoperatively due to
their submucosal location. A 59 year old African­American female presented with a one month history of abdominal
pain and weight loss. Abdominal CT scan revealed marked pancreatic and biliary ductal dilatation extending to an
enhancing ampullary mass. Endoscopic ultrasound with fine needle aspiration confirmed a neuroendocrine
neoplasm with gross invasion of the main pancreatic duct and common bile duct as well as gross invasion of the
submucosa. Pylrous­preserving pancreaticoduodenectomy was performed without issue. A 2.8 cm ampullary
gangliocytic paraganglioma with metastasis to 2 of 22 lymph nodes was diagnosed on final pathologic review. Generally, these tumors are felt to be benign, but rarely can metastasize to regional lymph nodes. Less than 20
cases have been reported in the worldwide literature with such lymph node involvement. Endoscopic ultrasound is a
helpful preoperative procedure as it aids in diagnosis and allows for detection of local invasion or concerning
lymphadenopathy. Although rare, the potential for gangliocytic paragangliomas to metastasize to regional lymph
nodes should be kept in mind when considering therapeutic options.
P106 BEYOND THE NUMBERS: THE PATTERN OF LYMPH NODE METASTASIS ALLOWS FOR
DISTINGUISHING DIFFERENT N+ CATEGORIES ­REAPPRAISAL OF THE JAPANESE CLASSIFICATION
OF LYMPH NODE METASTASES IN PANCREATIC ADENOCARCINOMA­ Laura Maggino, MD, Giuseppe
Malleo, MD, Francesco Gulino, MD, Giovanni Butturini, MD, PhD, Roberto Salvia, MD, PhD, Claudio Bassi, MD,
FRCS, FACS; Department of Surgery, The Pancreas Institute, University of Verona, Verona, Italy, Verona, IT
Introduction: Lymph node (LN) involvement is a major prognostic factor in pancreatic adenocarcinoma (PDAC).
However, in contrast to most other gastrointestinal cancers, there is not enough evidence to allow a proper
distinction of several LN­positive categories (as previously proposed by the Japanese Pancreas Society1) and the
current TNM­based staging system is therefore inaccurate to predict prognosis. It has been recently shown that the
number of positive LNs (PLN) enables to distinguish different N categories and therefore improves prognostic
accuracy in resected PDAC. However, the impact of LN metastases site on survival remains unclear.
Our aim was to reappraise the prognostic impact of the pattern of LN spread in a subgroup of patients undergoing
pancreaticoduodenectomy (PD) with “standard” lymphadenectomy according to the International Study Group of
Pancreatic Surgery (ISGPS).
Methods: Patients fulfilling the inclusion criteria (LN sampling of the stations included in the ISGPS definition, R0/1,
M0) were extracted from our electronic database and retrospectively analyzed. Pathologic reports were thoroughly
re­examined. LNs were classified into three groups according to the Japanese Pancreas Society1 (i.e. N1:
peripancreatic, N2: regional, N3: distant). Predictors of survival were analysed using univariate and multivariate
models.
Results: Among the patients who underwent PD for PDAC between January 2000 and December 2011, 204 were
enrolled in the study. The mean number of harvested LNs was 30,2 (SD 12,3; range 9­78). Factors with a significant
impact on survival in N+ patients at univariate analysis were: LN metastasis group (N1­N2­N3), lymph node ratio
(LNR), PLN, tumor grading, R status and adjuvant therapy. However, the sole LN­related parameter being significant
at multivariate analysis was the LN group (p=0,01; HR 2,2 IC 95%: 1,2­4).
Conclusion: The pattern of LN spread allows for distinguishing three N groups associated to different survival
outcomes, thus improving prognostic accuracy in LN positive patients. This parameter appears superior to other
“numeric” LN factors (such as the number of PLN and the LNR) in predicting the prognosis of resected PDAC and
could be taken into account for further revisions of the TNM staging system.
1. Japan Pancreas Society. Classification of pancreatic carcinoma. 2nd English edition Tokyo: Kanehara & Co. Ltd;
2003
P107 BILIARY STENOSIS AND GASTRIC OUTLET OBSTRUCTION: COMPLICATIONS AFTER ACUTE
PANCREATITIS Motokazu Sugimoto, Gregg Flint, Cody Boyce, John Kirkham, Tyler Harris, Sean Carr, David
Sonntag, Brent Nelson, Joshua Barton, L W Traverso; St. Luke's Health System, Boise, US
Introduction: Common bile duct (CBD) stenosis and gastric outlet obstruction (GOO) during acute pancreatitis are
not often reported although these conditions have been recognized with chronic pancreatitis. The aim of this study
was to observe the frequency, duration, and treatment of CBD stenosis and GOO.
Methods: Between June 2010 and June 2014, 871 patients were hospitalized with clinical diagnosis of acute
pancreatitis at the St. Luke’s Health System. Of those 139 cases had pancreatic and/or peripancreatic collections
by CT scan and were included in our study. Severity was evaluated using the CT severity index (CTSI) scoring
system. Percutaneous catheter drainage (PCD) was performed in 52 patients with persistent or enlarging collections
by CT scan, systemic inflammatory response syndrome, organ failure, and/or refractory abdominal pain. All patients
were followed until resolution with median follow­up of 483 days [range, 47­1355] after index discharge. CBD
stenosis was defined as anatomic narrowing of the CBD, intrahepatic biliary dilatation, and usually elevation of liver
function tests while GOO was defined as gastric dilatation, narrowing of duodenum shown by upper gastrointestinal
contrast study, and the inability to handle gastrointestinal secretions (nausea and vomiting). In these patients with
pancreatic and/or peripancreatic collections, the clinical and pathological findings were compared between those
who did and did not develop CBD stenosis and/or GOO.
Results: Of the 139 cases there were 13 cases with CBD stenosis and/or GOO (9%) – 7 with CBD stenosis­only, 2
with GOO­only, and 4 with both CBD stenosis and GOO. Comparing these 13 cases to the 126 patients without
CBD stenosis or GOO the former had higher CTSI scores (P < 0.001), higher incidence of pancreatic head necrosis
(P < 0.001), and higher incidence of portal vein occlusion (P = 0.002). They required PCD more frequently (P <
0.001). For those treated with PCD, amylase­rich drain fluid and culture­positive drain fluid were observed more
often (P < 0.001 and P = 0.006, respectively).
CBD stenosis occurred 65 days [11­231] after onset, whereas GOO occurred 88 days [22­117] after onset.
In 11 the patients with CBD stenosis, 6 were treated with endoscopic stenting and 5 underwent percutaneous
transhepatic biliary drainage. Median duration of biliary decompression was 180 days [36­231]. All 6 patients with
GOO underwent percutaneous gastric drainage for a median of 117 days [41­176]. Five patients had simultaneous
jejunal feeding. All 13 cases recovered from these inflammatory complications without surgical intervention.
Conclusions: The anatomic proximity of the CBD and the duodenum to the severe inflammatory process of acute
pancreatitis results in the late onset of reversible inflammatory stenosis. Especially in patients with pancreatic head
necrosis, development of these complications should be anticipated. Percutaneous and endoscopic methods
successfully managed these complications although prolonged decompression was required.
P108 CAN AETHIOLOGICAL FACTORS GIVE SOMETHING NEW IN THE TREATMENT OF PANCREATIC
HEAD TUMORS? Gyula Farkas Jr, PhD1, Peter Hegyi, DSc2, Gyorgy Lazar, DSc1; 1University of Szeged,
Department of Surgery, Szeged, Hungary, 2University of Szeged, First Department of Medicine, Szeged, Hungary,
Szeged, HU
Introduction: Pancreatic head tumors are irresecables in mainly 80% when patients occurs in GI departments.
Aim: we investigated 200 inoperable pancreatic tumor patients (2010.09­2014.09.) in the view of anamnestic datas.
We checked if these datas can give something new information to increase the resecability rate.
Materials and methods: we investigated the age, the sex, the alcohol and nicotine consumption, the period of
jaundice, the type and period of pain. Abdominal US, CT scan, and hystological verification was performed in all
cases. In most cases EUS was carried out. The inoperable state was declaired either by CT and or EUS, but in 50
cases the inoperability were detected only during surgical exploration due to vessel propagation; in those cases
bypass operation was performed. If jaundice was on stage, ERCP and stent implanting was performed. Results: 114 male and 86 female patients were involved in. The mean age was 65.3 vs. 69.1 years accordingly.
Neither alcohol nor nicotine consumption were relevant. In 40 % of cases jaundice was presented, the mean onset
was 1.5 week long. Indefinite upper abdominal or back pain was found in 65%. The patients claimed about 16.5 kg
of weight loss in 70%. About 1/3 of the patients had nor jaundice nor weight loss, only prolonged upper gastric
indefinite pain. We found painless patients in about 35 %, they just had relevant weight loss ­ more than 6kg/month.
Discussion: From these datas we conlude that if persistent or returning upper gastric indefinite pain with relevant
weight loss occurs an early gastroenteroligical examination with CT scan and or EUS is necessary for the succesful
and better treatment of pancreatic head tumors.
P109 CIRCULATING MICRORNAS AS RESPONSE INDICATORS FOR THE TREATMENT OF PATIENTS
WITH PANCREATIC CANCER Eveline E Vietsch, MD1, Jeroen W Versteeg2, Narayan M Shivapurkar1, Niels F
Kok2, Mustafa Suker2, Casper H van Eijck2, Anton Wellstein1; 1Lombardi Comprehensive Cancer Center,
Georgetown University, 2Erasmus Medical Center, Rotterdam, the Netherlands, Washington Dc, US
Introduction/Background
Predicting which patient with pancreatic ductal adenocarcinoma (PDAC) will respond to therapy is challenging and
many patients will not benefit from treatment due to late detection, chemo/radiotherapy resistance or organ
metastasis. MicroRNAs (miRs) are short RNAs that control cellular pathways in physiology and pathology. Many of
the miRs expressed in different organs are shed into the circulation, can be isolated from serum and can serve as
stable biomarkers of both physiologic state and malignant progression (Refs. 1, 2). Monitoring responses to therapy
using changes in the expression patterns of circulating miRs could be useful in treatment decision making. Here we
sought to establish circulating miR signatures indicative of tumor recurrence, metastasis and treatment responses.
Materials and Methods
To develop a panel of miRs, we screened serum for circulating miRs by analyzing expression levels with
quantitative real time PCR (qRT­PCR). Serum samples from patients with resectable PDAC were collected before
and after surgery or chemo/radiotherapy. All patients that underwent surgery received postoperative gemcitabine.
Moreover, we included serum samples of patients that received neoadjuvant gemcitabine before surgery.
Three comparisons are used: (a) miRs indicative of cancer presence are derived by a comparison of serum samples
collected before and after surgery. Serum of patients who underwent comparable surgery for benign pancreatic
disease serve as controls. (b) To identify miRs indicative of organ metastasis, miRs in serum from patients with
early stage disease are compared with patients with known organ metastases. (c) To assess the effect of
chemo/radio therapy miR patterns in patients with postoperative adjuvant gemcitabine versus pre­operative neo
adjuvant treatment are compared. Principal Component Analyses are used to reveal distinct patterns indicative of
disease progression and/or response to treatment.
Results
The preservation of miRs in serum samples was assessed and confirmed by measuring the levels of five miRs that
were previously studied in samples from transgenic animals and patients with PDAC (Ref. 2). Subsequently,
genome wide expression analysis was conducted on samples from various time points after treatment, to identify
informative miRs in an unbiased fashion.
Expression patterns of circulating miRs were distinct when comparing serum samples from patients before and after
surgery. This was confirmed by PCA. Also, the extent of changes in miR levels suggest distinct efficacy of surgical
tumor removal. Analysis of samples from patients with chemotherapy is ongoing and we expect to see distinct
changes based on studies in transgenic mice with PDAC treated with gemcitabine (Ref 2).
Discussion/Conclusion
Circulating miRs can be used as biomarkers and can indicate response to surgery in patients with PDAC. This
approach is minimally invasive and allows to assess individual patient responses repeatedly during treatment. The
circulating miRs indicate molecular aspects of cancer, and provide information about the host response to
therapeutic intervention.
References
1. Shivapurkar N et al. (2014) Recurrence of early stage colon cancer predicted by expression pattern of circulating
microRNAs. PLoS ONE 9:e84686.
2. LaConti JJ et al. (2011) Tissue and Serum microRNAs in the Kras Transgenic Animal Model and in Patients with
Pancreatic Cancer. PLoS ONE 6:e20687.
P110 COMPARISON OF THE PROGNOSTIC IMPACT OF PERIOPERATIVE CA 19­9, SPAN­I AND DUPAN II
LEVELS IN PATIENTS WITH RESECTABLE PANCREATIC CARCINOMA Naru Kondo, MD, Yoshiaki
Murakami, MD, Kenichiro Uemura, MD, Yasushi Hashimoto, MD, Naoya Nakagawa, MD, Taijiro Sueda, MD;
Institute of Biomedical and Health Sciences Applied Life Sciences Surgery,Hiroshima University, Hiroshima, JP
Background: Although serum carbohydrate antigen 19­9 (CA19­9), s­pancreas antigen­1 (SPan­1) and duke
pancreatic monoclonal antigen type 2 (DUPAN­2) are commonly utilized tumor markers in pancreatic ductal
adenocarcinoma (PDAC), it is still unclear which is the most useful tumor marker for predicting prognosis after
surgical resection. Purpose: The aim of this study was to compare the prognostic impact of perioperative serum CA19­9, SPan­1 and
DUPAN II levels in patients with resectable PDAC. Methods: Of a total of 230 consecutive patients who underwent surgical resection for PDAC, preoperative CA19­9,
SPan­1 and DUPAN II levels were available in 189 patients, and both pre­ and postoperative CA19­9, SPan­1 and
DUPAN II levels were available in 142 patients. Preoperative CA19­9, SPan­1 and DUPAN II levels were analyzed
to compare the diagnostic value for resectable PDAC. Moreover, the relationships of clinicopathological factors
including pre­ and postoperative CA19­9, SPan­1 and DUPAN II levels with overall survival (OS) were analyzed with
univariate and multivariate analyses in 142 patients.
Results: Preoperative Span­1 levels were significantly correlated with preoperative SPan­1 levels (r = 0.85, p <
0.001), whereas preoperative DUPAN II levels were not (r = 0.12, p = 0.10). Of the 189 patients with resectable
PDAC, elevated preoperative CA19­9 (> 37 U/ml), SPan­1 (> 30 U/ml) and DUPAN II (> 150 U/ml) levels were
found in 113 (60%), 96 (51%) and 82 (43%) patients, respectively. Univariate analysis revealed that absent of
postoperative adjuvant chemotherapy (p = 0.0002), R1 resection (p = 0.01), higher histological grade (p = 0.007),
more advanced UICC pT stage (p = 0.04) and lymph node metastasis (p = 0.004) were significantly associated with
worse OS. In addition, significant worse OS were found in patients with higher preoperative CA19­9 (>200 U/ml, p =
0.002), SPan­1 (> 50 U/ml, p = 0.0005) and DUPAN II (> 300 U/ml, p = 0.001), and in those with elevated
postoperative CA19­9 (>37 U/ml, p < 0.0001), SPan­1 (> 30 U/ml, p = 0.004) and DUPAN II (> 150 U/ml, p =
0.006). In multivariate analysis, absent of postoperative adjuvant chemotherapy (hazard ratio [HR], 4.47: 95%
confidence interval [CI], 1.83 – 10.04; P = 0.001), higher histological grade (Grade 2/3) (HR, 2.71; 95% CI, 1.41 –
5.45; p = 0.002), R1 resection (HR, 2.14; 95% CI, 1.19 – 3.78; p = 0.01) and elevated postoperative CA19­9 (> 37
IU/ml) (HR, 4.70; 95% CI, 1.99 – 10.71; p = 0.0006) were identified as independent predictors for worse OS. Conclusion: When the prognostic impacts of perioperative serum CA19­9, SPan­1 and DUPAN II levels in patients
with resectable PDAC were compared, elevated postoperative CA19­9 (> 37 IU/ml) would be the strongest
predictive marker of poor survival in the perioperative period, which may contribute to establishment of new
therapeutic strategy.
P111 CONVERSION SURGERY IS A VITAL OPTION FOR LOCALLY­ADVANCED PANCREAS CANCER
Keita Wada, MD, Keiji Sano, Hodaka Amano, Fumihiko Miura, Naoyuki Toyota, Hiromichi Ito, Yoshiko Aoyagi,
Makoto Shibuya; Teikyo university school of medicine, Tokyo, JP
BACKGROUND: Patients with locally­advanced unresectable pancreatic cancer (LAUR­PDAC) are not candidate
for surgical resection. However, recent advances of chemotherapy for PDAC gives us unique opportunities to see
patients with initially unresectable disease who respond to down­staging chemotherapy (DCT) very well, i.e. tumor
shrinkage or downstage. For those patients is there any role for surgical resection? We herein reported our initial
experience with conversion surgery following a favorable response to DCT for patients with initially LAUR­PDAC.
METHODS: A retrospective comparison was performed between 30 patients with LAUR­PDAC who received
upfront surgery during 2005­09 (Surgery­first) and 22 patients who received DCT using gemcitabine with S1 (GS) as
a first­line anticancer treatment during 2010­13 (Chemo­first). Primary endpoint was survival, and secondary
endpoint included conversion rate, short­term and long­term outcome of surgery.
RESULTS: There was no difference between the two groups in age, gender, location of tumor and involved artery.
Objective response rate based on RECIST of in Chemo­first group was 36% and disease control rate was 68% at 6
months after the initial chemotherapy. During DCT CA19­9 reduction was significantly associated with favorable
response. Among 15 patients who respond to DCT (RECIST: PR/SD) for longer than 6 months, conversion surgery
was attempted in 7 and completed in 5 (23%) at 9 (6­12) months after the initial treatment. All 5 patients with surgery
achieved R0 with pathological anti­tumor response ranging 70­90%, and 3 of 5 were node­negative. Intention­to­treat
overall survival was 11.5 months in Surgery­first group and 15.5 months in Chemo­first group (p=0.95). In Surgery­
first group, 4 out of 30 patients live longer than 5 years without recurrence, whereas in Chemo­first group all patients
without surgery died within 3 years but among 5 patients with conversion surgery 3 of 5 patients are alive with a
follow­up period of 30 (16.5­44.3) months.
CONCLUSIONS: Even for patients with LAUR­PDAC, surgical resection has potential to achieve cure of disease.
Chemo­first strategy followed by conversion surgery is better than Surgery­first by selecting those who can benefit
from this kind of aggressive surgery.
P112 HAS SURVIVAL FOLLOWING PANCREATICODUODENECTOMY FOR PANCREAS
ADENOCARCINOMA IMPROVED OVER TIME? Ahmed Salem, Mina Alfi, Emily Winslow, MD, FACS, Clifford S
Cho, MD, FACS, Sharon M Weber, MD, FACS; University of Wisconsin, Madison, US
Background: Survival following resection of pancreas cancer is poor, and it is uncertain whether improvements in
outcome have occurred over time. Due to the recent advances in surgical techniques, diagnostic evaluation, and
systemic treatment of pancreas cancer, we hypothesize that pancreas cancer outcome has improved over time.
Methods: Prospectively collected data on patients who underwent pancreaticoduodenectomy for pathologically
confirmed pancreatic adenocarcinoma from (1999 to 2012) were analyzed. Patients were divided into era 1 (1999 ­
2005), and compared to era 2 (2006 ­ 2012). Patient demographics, clinicopathological data and operative outcomes
were analyzed.
Results: A total of 216 patients were evaluated, including 76 in era 1 and 140 in era 2. Overall operative mortality
(30 d) was 1.4%, (1.3%, era 1, vs 1.4%, era 2, p=0.946). Patients in era 2 were at increased risk for a number of
poorer pathological characteristics, although margin positivity decreased with the concomitant increased use of
venous resection in era 2 (Table 1). There was no difference in median survival between era 1 and 2 on univariate
analysis (18 mo., vs 21 mo., p=0.830). After adjusting for perineural invasion, lymphovascular invasion, margin
status, EBL and venous resection, there was no association of improved survival in era 2 compared to era 1
(OR=1.036, p=0.848, CI=0.722 – 1.486).
Factor
Table 1. Univariate Analysis of Factors Influencing Survival
n (%)
p Value
Era 1
Era 2
• Advanced Stage (IIB III)
35 (64)
97 (71)
0.333
• Perineural Invasion
27 (40)
95 (68)
<0.001
• Lymphovascular Invasion
7 (10)
42 (30)
0.002
• Lymph Node Positivity
45 (60)
97 (70)
0.148
• Mean Tumor Size (cm, mean ± SD)
3.1±1.2
3.2±2.6
0.628
• Margin Positivity
29 (39)
25 (18)
0.001
• Estimated Blood Loss (EBL in ml) (mean ± SD)
990 ± 1599
640 ± 591
0.021
• Venous Resection
9 (12)
36 (26)
0.018
• Neoadjuvant Therapy
8 (12)
23 (16)
0.376
• Adjuvant Therapy
34 (50)
79 (56)
0.383
Pathological Features:
Operative Features:
Therapeutic Features:
Conclusion: Patients with more advanced and more aggressive tumors are undergoing definitive resection. After
adjusting for clinicopathological features, there was no association of improved outcome over time. However,
despite an increasing prevalence of anatomically advanced and histologically aggressive tumors, perioperative
outcomes such as blood loss and margin negativity improved over time, with no increase in 30 day mortality.
Strategies designed to improve systemic treatment of pancreas cancer are essential to improving outcome.
P113 HIGH­GRADE INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM IS NOT MALIGNANCY Neda
Rezaee, MD, Jin He, MD, PhD, Bulent Salman, MD, Ralph H Hruban, MD, John L Cameron, MD, Nita Ahuja, MD,
Matthew J Weiss, MD, Laura Wood, MD, PhD, Anne Marie Lennon, MD, PhD, Christopher L Wolfgang, MD, PhD;
Johns Hopkins Medical Institutions, Baltimore, US
Background: Since identification of intraductal papillary mucinous neoplasm (IPMN) in 1996, high­grade dysplasia
and IPMN­associated invasive carcinoma was used frequently under the umbrella term “malignancy”. We aimed to
compare the pathological features and survival outcomes of high­grade IPMN to invasive carcinoma.
Patients and Methods: From 1996 to 2013 data of 616 patients who underwent pancreatic resection for an IPMN
were reviewed. IPMNs were classified as low/intermediate­, high­grade dysplasia (HGD), and invasive carcinoma.
Results: A total of 293 (48%) patients diagnosed with low/intermediate­grade dysplasia, 140 (23%) with HGD, and
183 (30%) with invasive carcinoma. Actual 5­year survival was 55% for the entire cohort. The median overall
survival was 94 months for HGD, which was similar to low/intermediate­grade IPMN (118 months, p=0.07), and
superior to invasive carcinoma (29 months, p<0.001) (figure). Invasive carcinoma was associated with regional
lymph node metastasis in 34%, perineural invasion in 38%, and vascular invasion in 38%. In contrast no lymph node
metastasis, perineural or vascular invasion was observed after resection of HGD. Compared to invasive carcinoma,
HGD was associated with a lower rate of positive margin (38% vs. 24%, p=0.007). Among patients who had more
than 6 months follow­up, the recurrence rate after resection of HGD (16%) was similar to low/intermediate dysplasia
(19%, p=0.50); and was lower compared to invasive IPMN (29%, p=0.03).
Conclusion: IPMN with high­grade dysplasia has a favorable survival outcome and a lower rate of recurrence after
resection compared to IPMN­associated invasive carcinoma, and thus should not be considered a malignant entity.
P114 IMPACT OF HISTOLOGICAL EFFECT OF CHEMORADIOTHERAPY (CRT) ON CLINICAL OUTCOME
FOR PANCREATIC ADENOCARCINOMA, PAYING ATTENTION TO INTRATUMORAL EXPRESSION OF
TENASCIN C AS A POTENTIAL SURROGATE MARKER Yasuhiro Murata1, Masanobu Usui1, Shugo Mizuno1,
Hiroyuki Kato1, Akihiro Tanemura1, Naohisa Kuriyama1, Yoshinori Azumi1, Masashi Kishiwada1, Hiroyuki Sakurai1,
Toshimichi Yoshida2, Shuji Isaji1; 1Department of Hepato­biliary Pancreatic and Transplant Surgery, Mie University,
Tsu, Mie, Japan, 2Department of Pathology, Mie University, Tsu, Mie, Japan, Tsu, Mie, JP
Background: The impact of histological effect of chemoradiotherapy (CRT) on clinical outcome for pancreatic
adenocarcinoma (PDAC) has remained uncertain, and the histological effect cannot be monitored during CRT
because of the lack of useful surrogate marker. Tenascin C (TNC) is an extracellular matrix that is not expressed in
normal pancreatic tissue but is upregulated in PDAC and that is associated with cell­matrix interaction facilitating
epithelial tumor cell invasion and metastasis. The primary objective of the current study is to determine impact of
histological effect of CRT on clinical outcome for PDAC patients. The secondary objective is to assess usefulness
of TNC as a surrogate for histological effect of CRT by evaluating relationship between histological effect of CRT
and intratumoral expression of TNC after CRT.
Methods: From February 2005 to December 2013, 124 consecutive PDAC patients (resectable (R): 11, borderline
resectable (BR): 76, unresectable (UR): 37) who underwent curative­intent pancreatectomy after CRT (gemcitabine
or gemcitabine plus S­1 based) and 12 patients (R: 9, BR: 3) who underwent pancreatectomy without receiving CRT
were enrolled in the study. Histological effect of CRT was evaluated according to Evans criteria, and the patients
were divided into two groups: high responder (grade IIb, III, IV) and low responder (grade I, IIa). The results were
correlated with rate of R0 (no residual tumor) resection and patient prognosis. Among the enrolled patients, the
intratumoral expression of TNC for 22 patients (R: 1, BR: 17, UR: 4) with CRT and 12 patients without CRT was
evaluated by immunohistochemistry of resected specimens. The staining of intratumoral TNC was scored as
negative (<=5%), weak (>5% to 20%), intermediate (>20% to 60%), and strong (>60%), and intermediate or strong
(>20%) was defined as positive. The results were correlated with histological effect of CRT.
Results: The rate of high responder was 34.7%. In 100% of R cases that underwent CRT, R0 resection was
achieved. For BR and UR cases who underwent CRT, the rates of R0 resection were higher in high responder than
that in low responder (High vs. low responder in BR, UR: 93.3 vs. 82.6%, p=0.159, 88.9 vs. 46.4%, p=0.017). The
3­year survival rates did not differ statistically significantly between high and low responders (High vs. low responder
in BR, UR: 50.9 vs. 38.1%, p=0.12, 29.2 vs. 6.4%, p=0.23). The positive rate of intratumoral expression of TNC
was 61.8%. The positive rate of intratumoral expression of TNC was significantly lower in the patients with CRT
than in those without CRT (with vs. without CRT: 45.5 vs. 91.7%, p=0.005). Among the 22 patients with CRT, the
positive rate of intratumoral TNC after CRT was significantly lower in high responder (n=10) than in low responder
(n=12) (High vs. low responder: 20 vs. 66.7%, p=0.036).
Conclusion: Effective histological effect of CRT contributes to increasing rate of R0 resection even for locally
advanced PDAC. The intratumoral TNC expression will be able to serve as a useful surrogate marker of histological
effect of CRT, if it can be monitored properly during CRT. P115 IS PARTIAL PANCREATECTOMY THE CORRECT OPERATION FOR DIFFUSE INTRADUCTAL
PAPILLARY MUCINOUS NEOPLASM (IPMN) INVOLVING THE ENTIRE MAIN DUCT? Neda Rezaee, MD, Jin
He, MD, PhD, John L Cameron, MD, Martin A Makary, MD, MPH, Timothy M Pawlik, MD, MPH, PhD, Nita Ahuja,
MD, Mathew J Weiss, MD, Anne Marie Lennon, MD, PhD, Ralph H Hruban, MD, Christopher L Wolfgang, MD,
PhD; Johns Hopkins Medical Institutions, Baltimore, US
Background: Mixed­type and main­duct Intraductal Papillary Mucinous Neoplasms (IPMN) commonly involve the
entire main pancreatic duct. Despite the diffuse nature and the significant risk of malignancy, these subtypes are
most often resected by a pancreaticoduodenectomy (PD) leaving a residual dilated main pancreatic duct. The risk of
progression to cancer in the pancreatic remnant is unknown.
Methods: 460 patients underwent resection of a non­invasive IPMN. The entire main pancreatic duct was diffusely
dilated in 15 main­duct IPMN (MD­IPMN) and 60 mixed type IPMN (MT­IPMN). A pancreaticoduodenectomy was
performed in 70 and a total pancreatectomy in 5. Results: The pathological findings included 36 (51%) patients with low/intermediate­grade dysplasia and 34 (49%)
with high­grade dysplasia. At a median follow­up of 29 months, 8 of 70 (11%) had progression of IPMN (n=6) or
developed cancer (n=2) within their remnant. Completion pancreatectomy was performed in 4 (6%; median of 55
months from the first operation). Two (3%) were found to have high­grade dysplasia and 2 (3%) had invasive cancer.
The degree of dysplasia in the pathology of the original operation among patients with progression included low or
intermediate­grade dysplasia in 5 and high­grade dysplasia in 3. Univariate and multivariate analysis did not identify
any clinical or pathological factor associated with recurrence.
Conclusion: Progression of disease within the remnant for a diffusely dilated duct in main and mixed type IPMN
following a pancreaticoduodenectomy occurs in approximately 10% over a median follow­up of only 29 months. PD
for this disease may be performed, but close and frequent long­term follow­up is required.
P116 LOW COMPLETION RATE OF ADJUVANT CHEMOTHERAPY AFTER ONCOLOGIC RESECTION OF
PANCREATIC CANCER IN CLINICAL ROUTINE CARE Guido Alsfasser, MD, Johanna Bochow, Anna L
Kutsch, Ernst Klar, Bettina M Rau; University of Rostock, Rostock, DE
Introduction Adjuvant chemotherapy (Cx) has evolved as integral part of multimodal treatment in resected ductal
pancreatic cancer. The German S3 guide­line recommends adjuvant Cx since its first implementation in June 2007
irrespective of tumor stage and R­status. In the present study we investigated the clinical impact of this guide­line in
terms of recommended, initiated, and completed adjuvant Cx in our institution. Patients and Methods Between 09/2003 and 12/2013 a total of 382 pancreatic resections were performed at our
institution. There were 141 patients undergoing oncologic pancreatic resection for ductal adenocarcinoma. Complete
Follow­up data could be obtained from 133 patients (94%). Follow­up information derived from patients directly, the
local University cancer registry, and from house practitioners or oncologists.
Results In our study group of 133 patients we had performed 104 Kausch­Whipple operations, 13 total
pancreatectomies, 15 left resections, and 1 segmental resection. The rate of R1 and R2 resections was 41% and
2%, respectively. Based on TNM­stage, R­status and individual postoperative patient performance adjuvant Cx was
recommended in 125 patients (94%). Follow­up revealed that only 103 patients (77%) acutally started Cx. Main
reasons for declining Cx were lack of patient consent or prolonged postoperative recovery. Completion of adjuvant
Cx was reached in 62 patients (60%) only, which equals 47% of all resected pancreatic carcinomas. Main reasons
for discontinuing of Cx were side effects or cancer recurrence.
Conclusion Our results indicate a high acceptance rate of adjuvant Cx as integral part of integrated pancreatic
cancer treatment among physicians in a clinical routine setting. However, there are relevant problems in realization
and completion of adjuvant Cx resulting in complete treatment in only less than 50% of patients outside of clinical
studies.
P117 LYMPHANGIOMA: A RARE BUT CURABLE TUMOR INVOLVING THE PANCREAS Owen Young, MD,
Thomas Biehl, MD, Adnan Alseidi, MD, Flavio Rocha, MD; Virginia Mason Medical Center, Seattle, US
Background: Lymphangioma is a rare tumor characterized by aberrant growth of lymphatic channels that can occur
in the abdomen and involve the pancreas. Although these lesions are benign, they can grow locally, causing pain,
early satiety and pancreatitis and may warrant complete resection if symptomatic.
Methods: We performed a retrospective review of all patients undergoing pancreatectomy at our institution since
2010 and identified three patients with peri­pancreatic lymphangioma. Clinical and treatment information was
obtained from electronic medical records.
Results: Case 1: A 35­year­old man presented with 3 years of increasing abdominal discomfort and a palpable
mass. He underwent abdominal CT imaging that demonstrated a 15 cm cystic tumor arising between the stomach
and pancreas. Endoscopic ultrasound and fine­needle aspiration revealed low CEA and elevated triglycerides. Case
2: A 22­year­old woman presented with an enlarging left abdominal mass, intermittent pain, and a CT scan revealing
a 20 cm cystic mass in the tail of the pancreas. Both had uncomplicated resections and, on final pathology, had
tumors with positive immunohistochemical staining for CD31 and D2­40, consistent with a lymphangioma. Case 3:
A 41­year­old man with recurrent idiopathic pancreatitis and CT images demonstrating an ill­defined fluid­filled cystic
mass around the head and neck of the pancreas extending into the porta hepatis was diagnosed with lymphangioma
by laparoscopic biopsy. Resection was not attempted given the resolution of symptoms with conservative
measures. All patients are currently alive without complaints.
Conclusions: The differential diagnosis for a cystic lesion involving the pancreas should include lymphangioma.
Complete surgical resection can be curative for symptomatic patients
P118 MODIFIED APPLEBY PROCEDURE WITH ARTERIAL RECONSTRUCTION: A LITERATURE REVIEW
AND REPORT OF 3 UNUSUAL CASES Jessica A Latona, MD1, Kathleen M Lamb, MD1, Daniel M Relles, MD2,
Warren R Maley, MD1, Charles J Yeo, MD1; 1Thomas Jefferson University Hospital, 2Morgan Stanley Children's
Hospital, Philadelphia, US
BACKGROUND
Pancreatic body and tail carcinomas are often diagnosed with local vascular invasion of the celiac axis (CA) and its
branches. With such involvement, these tumors have traditionally been considered unresectable. The modified
Appleby procedure allows margin negative resection of such locally advanced tumors. This procedure involves
distal pancreatectomy with en bloc splenectomy and CA resection and relies on the presence of collateral arterial
circulation via an intact pancreaticoduodenal arcade and gastroduodenal artery (GDA). When the resultant collateral
circulation is inadequate to provide sufficient hepatic and gastric arterial inflow, arterial reconstruction (AR) is
necessary to “supercharge” the inflow. Herein, we review all reported cases of AR with modified Appleby
procedures and report our experience of 3 recent cases with arterial reconstruction including 2 cases with arterial
bypasses not requiring interposition grafting.
METHODS
A PubMed search was systematically completed of studies relating to distal pancreatectomy with CA resection and
subsequent AR.
RESULTS
Eleven reports involving 27 patients were identified of distal and total pancreatectomy with AR after CA resection
(Table 1). The most common AR, performed in 11 patients, was a bypass from the aorta to the common hepatic
artery (CHA) using a variety of interposition conduits. In our experience, patient #1 had a primary side to end aorto­
CHA bypass, patient #2 had a primary end to end bypass of the distal CHA to the left gastric artery (LGA) in the
setting a replaced left hepatic artery (rLHA), and patient #3 required an aortic to proper hepatic artery bypass with
saphenous vein graft and venous reconstruction. Patient #1 received adjuvant chemo­radiation therapy. Patient #2
received neoadjuvant chemotherapy and proton therapy with a near complete tumor response. Patient #3 received
neoadjuvant chemo­radiation therapy with a near complete tumor response. All patients recovered well and they are
currently 6, 5 and 2 months post­op, respectively.
CONCLUSIONS
Criteria for resectablilty in patients with locally advanced pancreatic body and tail neoplasms are expanding due to
increasing experience with AR in the setting of the modified Appleby procedure. When performing AR, primary
anastomosis may be considered preferable to interposition grafting as it decreases the potential for the infectious
and thrombotic complications associated with conduits and it reduces the number of vascular anastomoses from
two to one. Consideration must also be given to normal variant anatomy of the hepatic circulation (as seen in patient
#2) during operative planning as the LGA is resected with the CA. The modified Appleby with AR, when used in
appropriately selected patients, offers the potential for safe, margin negative resection of locally advanced tumors.
P119 MULTIDISCIPLINARY MANAGEMENT OF PATIENTS AFFECTED BY POST­PANCREATECTOMY
HEMORRHAGES Domenico Borzomati, MD, FACS1, Rosario Francesco Grasso, MD2, Sergio Valeri, MD1,
Eliodoro Faiella, MD2, Gennaro Nappo, MD1, Giacomo Luppi, MD2, Pasquale Scognamiglio1, Roberto Coppola, MD,
Ph, FACS1; 1General Surgery, Campus Bio­Medico University of Rome, 2Interventional Radiology, Campus Bio­
Medico University of Rome, Lyon, FR
Introduction
Post­Pancreatectomy Hemorrhage (PPH) is a life­threatening complication after Pancreatico­Duodenectomy (PD).
According to severity, ISGPS classified PPH into grades A, B and C. Grade B­C cases require operative treatment
with a 20­30% treatment related mortality. Based on clinical conditions and local expertise, the treatment of PPH
ranges from re­surgery to Interventional Radiology (IR) procedures. In spite of a significant mortality rate, re­surgery,
ranging from hemostasis to completion of pancreatectomy, is still considered the first line therapeutical option in
case of Grade C PPH, with IR usually considered in case of stable patients. However, recently published papers
report that IR can successfully treat also unstable PPH patients. The aim of this retrospective analysis was to verify
if the implementation of IR utilization in our institution for the treatment of severe PPH determined a variation of
patients’ outcome.
Materials and Methods
We retrospectively evaluated all PD performed at Campus Bio­Medico University of Rome from 2004 to 2014. In all
cases, the incidence of PPH was evaluated and classified according to the ISGPS classification. According to the
kind of treatment, PPH were classified as: a) Surgery Alone; b) Surgery Included (Surgery + IR and/or Endoscopy)
c) Surgery Excluded (IR and/or Endoscopy). Outcome of the treatment in terms of efficacy (absence of re­bleeding
after treatment) and mortality for each group was also evaluated.
Results
From 2004 to 2014, 230 patients underwent PD at our Institution. PPH was recorded in 38 patients (16.7%). Sixteen
patients (42.1%) underwent conservative management while 22 patients (57,9%) (Grade A: 0%; Grade B: 36,4%;
Grade C: 63,6%) required operative treatment. Treatment consisted of: “Surgery Alone” in 8 (36,4%) cases, “Surgery
Included” in 7 cases (31,8%) and “Surgery Excluded” in 7 (31,8%) cases. An ex­post analysis of clinical conditions
and severity of PPH of the 22 treated patients showed no differences between the three groups of treatment.
Treatment’s efficacy was: a) “Surgery Alone” 7/8 cases (87.5%); b) “Surgery Included”: 6/7 (85,7%); “Surgery
Excluded”: 7/7 (100%). Treatment related mortality was: a) “Surgery Alone” 2/8 cases (25%); b) “Surgery Included”:
2/7 (28,6%); “Surgery Excluded”: 0/7 (0%).
Discussion
PPH is a relatively frequent and severe complication after PD. In case of severe PPH urgent treatment often
consisting of re­surgery is warranted. Re­surgery is an effective and time­tested option but it is affected by high
mortality rates. The introduction of IR determined an overlapping (in terms of indication) in the clinical management
of Grade B­C PPH patients. As a consequence, the use of IR in unstable cases is now an argument of debate with
no definitive guidelines stated. Our results confirm that IR is an effective treatment modality even in the
management of unstable PPH patients. In our experience no mortality in Grade B­C cases treated with IR was
recorded. Future studies should clarify indication and limits of IR in the management of PPH.
P120 NAB­PACLITAXEL PLUS GEMCITABINE VS GEMCITABINE ALONE FOR RESECTED PANCREATIC
CANCER IN A PHASE III TRIAL (APACT) Margaret Tempero1, Dana Cardin, MD2, Andrew Biankin, MD3, David
Goldstein4, Malcolm Moore5, Eileen M O'Reilly6, Philip Philip7, Hanno Riess8, Teresa Macarulla9, Lotus Yung10,
Mingyu Li10, Julie Jeane, PharmD10, Brian Lu10; 1UCSF Pancreas Center, 2Vanderbilt University Medical Center,
3Wolfson Wohl Cancer Research Center, 4Prince of Wales Hospital, 5Princess Margaret Hospital, 6Memorial Sloan
Kettering, 7Karmanos Cancer Center, 8Charité­Universitätsmedizin Berlin, Campus Virchow­Klinikum, 9Vall
d’Hebron University Hospital, 10Celgene Corporation, San Francisco, US
Background: Gem monotherapy after surgery improves both survival rates and disease­free survival (DFS) in
patients with PC. However, disease recurrence is common, suggesting a need for improved treatment. nab­P +
Gem demonstrated superior efficacy over Gem alone in a phase III trial (MPACT) of patients with metastatic PC
including the primary endpoint of overall survival (OS; median 8.7 vs 6.6 months; hazard ratio [HR] 0.72; P < 0.001).
Toxicities were manageable. Based on the activity demonstrated in the metastatic setting, nab­P + Gem will be
compared with Gem alone in the adjuvant setting.
Trial design: Approximately 800 patients with histologically confirmed PC who undergo macroscopic complete
resection (R0 or R1) with no evidence of metastasis will be randomized 1:1 to receive 6 cycles of either nab­P 125
mg/m2 + Gem 1000 mg/m2 or Gem alone 1000 mg/m2 on days 1, 8, and 15 of each 28­day cycle. Other eligibility
criteria include staging of T1­3, N0­1, M0; Eastern Cooperative Oncology Group performance status of 0 or 1;
acceptable hematologic function; and carbohydrate antigen 19­9 < 100 U/mL prior to randomization. Patients with
neuroendocrine tumors, any other malignancy within 5 years of randomization, infection with human
immunodeficiency virus or hepatitis B or C, or prior neoadjuvant treatment or radiation therapy for PC are ineligible.
Stratification factors are resection status (R0 vs R1), nodal status (LN+ vs LN−), and geographic region (North
America, Europe, and Australia vs Asia Pacific). The primary endpoint is independently assessed DFS, and
secondary endpoints are OS and safety. Exploratory endpoints include molecular profiling of tumor tissue to
correlate tumor heterogeneity with clinical outcome and quality of life as assessed by the European Organisation for
Research and Treatment of Cancer Quality of Life Questionnaires (EORTC QLQ­C30 and EORTC QLQ­PAN26). At
least 489 DFS events from 800 patients will allow 90% power to detect an HR for DFS of 0.74 at a 2­sided
significance level of 0.05. One interim safety analysis and 2 interim efficacy analyses (the first for futility and the
second for both futility and efficacy) will be performed. Patient enrollment is ongoing (ClinicalTrials.gov identifier
NCT01964430).
Conclusions: This phase III trial will determine whether nab­P + Gem is superior to Gem alone as adjuvant treatment
for patients with resected PC. Such a finding would establish nab­P + Gem as a new standard therapy in this
disease setting.
P121 NATIONAL DISPARITIES IN MINIMALLY INVASIVE SURGERY FOR PANCREATIC CANCER
Emmanuel Gabriel, Pragatheeshwar Thirunavukarasu, Kristopher Attwood, Steven Nurkin; Roswell Park Cancer
Institute, Buffalo, US
Robotic vs Open
Laparoscopic vs Open
OR
OR
COLON
Insurance
Uninsured
1.000
1.000
Private
3.818
1.635
Medicaid
2.821
1.110
Medicare
3.234
1.494
Geographic setting
Metro
1.000
1.000
Urban
0.712
0.873
Rural
0.695
0.828
Facility type
Community cancer center
1.000
1.000
Comprehensive community cancer center
1.242
1.407
Academic center
1.874
1.393
RECTUM
1.000
1.000
3.579
1.641
2.321
1.239
3.097
1.457
Insurance
Uninsured
Private
Medicaid
Medicare
P value
0.001
0.001
0.001
0.001
Geographic setting
Metro
1.000
1.000
Urban
0.797
0.864
Rural
0.716
0.843
Facility type
Community cancer center
1.000
1.000
Comprehensive community cancer center
2.433
1.298
Academic center
4.042
1.132
0.002
0.001
Introduction: Social and racial disparities have been related to differences in access to care. This study
investigated patterns in minimally invasive surgery (MIS) across different social, racial and geographic populations of
patients with pancreatic cancer.
Methods: We utilized the National Cancer Database, 2004 to 2011, to identify patients with pancreatic cancer who
underwent surgery through either an open, laparoscopic or robotic approach. Multivariate analysis was performed to
characterize differences in patient demographics in relation to surgical approach.
Results: A total of 11,464 patients were identified. The initial surgical approach included 82.5% open (9,461), 15.8%
laparoscopic (1,815) and 1.6% robotic (188). Table 1 shows the results of our analysis. Race was not statistically
significant across the different surgical approaches. There was a trend toward increased MIS in patients with private
insurance. Academic centers performed more MIS compared to community cancer programs. On multivariate
analysis, only national location was shown to be a statistically significant factor associated with increased rates of
MIS. Patients in the Middle Atlantic region of the US were most likely to have robotic surgery. Regarding
laparoscopic surgery, the Mountain and West South Central states had the lowest rates of laparoscopic procedures,
but among the other national regions there were no statistically significant differences.
Conclusions: Minimally invasive approaches for pancreatic cancer comprise 17.5% of surgical procedures. Race
and insurance status were not statistically significant factors associated with MIS. Although academic centers
performed most of the MIS, specific geographic regions comprised the only statistically significant factor on
multivariate analysis.
P122 PANCREATIC SURGERY FOR PANCREATIC ADENOCARCINOMA: A COMPARISON BETWEEN THE
LAPAROSCOPIC AND OPEN SURGICAL APPROACH. John A Stauffer, MD1, Alessandro Coppola, MD2,
Horacio J Asbun, MD1; 1Mayo Clinic Florida, 2Università Cattolica del Sacro Cuore, Rome, Italy, Jacksonville, US
Introduction: With a more widespread adoption of minimally invasive pancreatic surgery, concerns have risen
regarding the maintenance of oncologic principles when dealing with malignancy. The purpose of this study was to
assess the adequacy of laparoscopic pancreatic resection (LPR) for patients with pancreatic adenocarcinoma
(PADC).
Methods: A retrospective analysis of a prospectively maintained database was performed for all patient undergoing
distal pancreatectomy (DP) (n=70) and pancreaticoduodenectomy (PD) (n=95) at a single institution. These groups
were divided into open distal pancreatectomy (ODP) (n=29) vs laparoscopic distal pancreatectomy (LPD) (n=41),
and open pancreaticoduodenectomy (OPD) (n=57) vs laparoscopic pancreaticoduodenectomy (LPD) (n=38). Age,
Sex, BMI, Operative Time (OT), Estimated Blood Loss (EBL), Transfused Patient (TP), Vein Resection (VR),
Multivisceral Resection (MR), Fistula Rates (PF), Delay Gastric Emptyng (DGE), Post­op Hemorrhage (PH), Lymph
Node Dissection (LND), Tumor size (Ts), R0 Rate (R0r), ICU days (ICUd), Length of Stay (LOS), Clavien­Dindo
classification complications, Mortality (M), disease free survival (DFS) and overall survival (OS) were compared for
both groups.
Results: For DP the EBL, TP, LND, ICUd and LOS were significantly different between ODP and LDP, favoring
LDP. In PD the OT is shorter for OPD. There is no statistically difference regarding EBL in the PD group, however
the TP is statistically lower for LPD. No differences were observed in all other parameters in the DP or PD groups
including disease free and overall survival.
Conclusion: This limited study suggests that LPR for PADC has similar oncologic results to open surgery. In
addition, a significant decrease in TP was observed for LDP and LPD. Further studies are needed to assess if this
difference is reproducible in larger matched series and as such, if it translates into any significant change in overall
prognosis for these cancer patients.
P123 PANCREATITIS, CANCER, AND THE INTERNET: WHAT DOCTORS SHOULD KNOW TO BEST HELP
THEIR PATIENTS Isabella Guajardo, BA; University of California, San Francisco, San Francisco, US
Introduction/Background
After speaking with their physician and learning the poor prognosis for pancreatic cancer, patients will return home
and research their disease online. Physicians must be aware of the overly technical, inaccessible nature of the
material patients will find, and prepare their patients for this.
Methods
A web search of “pancreatic cancer,” “cáncer de páncreas,” “cholangiocarcinoma,” “cáncer de las vias biliares” and
“pancreatitis” were performed individually on Google. Following healthcare IT research practices, websites listed on
the first two pages were assessed using SOL, which is a frequently used tool for analyzing the readability of
Spanish language healthcare materials. The results were then double checked with three additional readability
indices: LIX, Automated Readability Index, and the Coleman­Liau index. The quality of the information was
assessed using the validated DISCERN instrument.
Results
Although the National Institutes of Health and the American Medical Association recommend that patient health
information be written at a 6th grade level, the information reviewed in this study was far more difficult to
comprehend.
The overall readability for English­language pancreatitis information was at the level of a college student. The
readability scores for 75% of the sites were “very difficult,” and the remaining were “difficult.” Quality scores using
DISCERN are given on a 1 to 5 Likert­type scale. The DISCERN result average in English language sites was 2.7,
with 6% receiving a 1, and 12% receiving a 4. In Spanish language sites, the DISCERN average was 2.6, with 12%
receiving a 1. Of the Spanish language websites, 75% of the sites were “very difficult,” and the remaining were
“difficult.”
Prior research shows that Hispanic patients develop pancreatic cancers earlier, are more likely to be diagnosed
later, and are less likely to receive chemotherapy or surgical treatment than whites. To ensure that physicians are
prepared for their Spanish speaking patients, and those patients’ typically extremely poor prognoses, the tests were
also performed for Spanish language websites. Of the English language websites, 75% of the sites were “very
difficult,” and the remaining were “difficult.” Of the Spanish language websites, 87% of the sites were “very difficult,”
and the remaining were “difficult.” The DISCERN result average in English language sites was 2.7, with 12%
receiving a 1, and 25% receiving a 4. In Spanish language sites, the average was 2.3, with 33% receiving a 1.
The overall readability for English and Spanish­language cholangiocarcinoma information was similar.
Discussion/Conclusion
Physicians should incorporate this knowledge when discussing their patients’ diagnoses. Physicians can prepare
their patients, during an already traumatic time, for the frustration and anger they are likely to encounter if they
search the internet to learn more about their disease. This will allow physicians to better set expectations and
encourage their patients to come to them with questions. As their patients deal with denial, anger and bargaining,
physicians are better able to help them through the diagnoses and treatments. P124 PREDICTORS OF PROGRESSION OF LOCALLY ADVANCED PANCREATIC CANCER ON
NEOADJUVANT CHEMOTHERAPY J B Rose, MD, MAS, F G Rocha, MD, A A Alseidi, MD, T R Biehl, MD, B
Lin, MD, V Picozzi, MD, W S Helton, MD; Virginia Mason Medical Center, Seattle, US
Introduction:
Neoadjuvant treatment for patients with locally advanced pancreatic adenocarcinoma (LAPD) is becoming more
widely utilized and may be associated with improved survival. However, even with careful patient selection, many
patients do not undergo curative resection due to disease progression. The objective of the present study is to
identify predictors of disease progression prohibiting resection.
Methods:
A retrospective review was performed on all patients with LAPD at a high volume tertiary care center between
January 2008 and August 2014 who received extended neoadjuvant gemcitabine and docetaxel chemotherapy.
Clinicopathologic predictors of disease progression prohibiting resection were determined by univariate and
multivariate logistic regression analysis.
Results:
Eighty­four patients with LAPD were initiated on neoadjuvant chemotherapy. 16 patients (19%) progressed on
treatment by RECIST criteria (9 distant, 7 local). Multivariate logistic regression analysis found that sex, age by
quartile, McGill­Brisbane score, clinical stage III, or multi­vessel involvement were not predictive of progression
prohibiting resection. However, tumor size (OR 2.6 [1.3 – 5.5]) and lack of at least a 50% decrease in CA19­9 levels
(OR 13.2 [2.5 – 69.1]) were. On sub­analysis, CA19­9 decrease <50% remained predictive for any progression while
tumor size predicted distant progression only. A receiver operating characteristic curve showed that tumor size
>3cm was 87% sensitive for preoperative progression (AUC 0.785).
Conclusion:
Early identification of LAPD patients at risk for progressing on neoadjuvant chemotherapy can be aided by
monitoring for a 50% decrease in CA19­9 and by identifying tumors >3cm. Patients with these risk factors may
benefit from additional treatment prior to an attempt at resection.
P125 READABILITY AND ACCURACY OF ONLINE PATIENT MATERIALS FOR PANCREATIC CANCER BY
TREATMENT MODALITY AND WEBSITE AFFILIATION Alessandra Storino, MD1, Manuel Castillo­Angeles,
MD1, Ammara A Watkins, MD1, Christina Vargas, MD1, Joseph D Mancias, MD, PhD1, Andrea J Bullock, MD1,
Aram N Demirjian, MA2, A. James Moser, MD1, Tara S Kent, MD1; 1Beth Israel Deaconess Medical Center,
2University Of California Irvine, Boston, US
Introduction: Online health information is frequently sought by patients but there is little control of its quality. Patient
and family understanding of presented information may depend in part upon readability and accuracy. This study
was undertaken to evaluate reading level and accuracy among commonly searched websites relating to treatment
options for pancreatic cancer.
Methods: An online search on 5 pancreatic cancer treatment modalities was conducted. For each website,
readability level was measured by 9 standardized tests and accuracy was assessed by an expert panel. Readability
and accuracy were compared by treatment modality and website affiliation using Kruskal­Wallis test.
Results: Significant differences existed by treatment modality for both readability and accuracy (see Table 1), with
surgery­related websites having the lowest reading level. Alternative therapy­related websites had significantly lower
accuracy than websites discussing other treatment modalities. Readability varied by affiliation, with non­profits
having lower readability than websites owned by media (p=0.00001) and academic centers (p=0.0001). Privately­
owned websites had lower readability than media sites (p=0.009). Accuracy was highest for government websites,
but government, academic, and non­profit sites were more accurate than privately­owned or media­owned
websites (p=0.0001). There is no association between accuracy and readability level.
Conclusions: Although variation existed in both readability and accuracy, improvement is needed throughout.
Readability level for all treatment modalities is higher than recommended, which may negatively impact
patient/family understanding of treatment options. Accuracy was generally reasonable, except for alternative therapy
websites. Website affiliation also impacted both readability and accuracy. Privately­owned and media sites had
lower accuracy. In accordance with patient­centered care, improvement is needed in the quality of online resources
in order to empower patients in the shared­decision making setting.
P126 REDUCED FIELD OF VIEW DIFFUSION IMAGING OF THE PANCREAS Lorenzo Mannelli, MD, PhD1,
Maggie M Fung, PhD2, Gregory Nyman1, Sabrina Lopez1, Richard K Do, MD, PhD1; 1Memorial Sloan Kettering
Cancer Center, 2Global MR Applications and Workflow, GE Healthcare, New York, NY, United, New York, US
Purpose:
Diffusion­weighted imaging (DWI) is routinely used in MR imaging of the pancreas for tumor detection. Respiratory
motion suppression is critical to achieve images of diagnostic quality and visualizing the fine details. Recently, a
reduced Field of View (rFOV) methodology has been introduced with the potential to achieve higher resolution DWI
in centrally located organs, such as the pancreas and prostate. This imaging approaches allows for higher image
resolution of the organ of interest. The aims of this study are:
To demonstrate the feasibility of pancreatic high resolution NT rFOV ss­DWI.
To compare image quality, presence and grade of artifacts, signal­to­noise­ratio (SNR), and apparent diffusion
coefficient (ADC) values in pancreatic tissues between NT full FOV ss­DW EPI and NT rFOV ss­DW EPI.
Method:
This retrospective study was approved by the local IRB committee. 10 consecutive patients who underwent both
large FOV and rFOV pancreas DWI with NT over a 5 month period in 2014 were included. Conventional large FOV
DWI and rFOV DWI were acquired. Navigator echo respiratory triggering technique was used in both large and rFOV
DWI.
Presence of artifacts and overall image quality were subjectively rated for both large and rFOV DWI by 2 radiologists
in consensus with a 5 point scale:
Artifacts: 1=no artifact, 2 = minimal artifact that does not interfere with diagnostic quality, 3 = artifacts that reduces
diagnostic quality, 4 = only minimal diagnostic information is still present, 5= non diagnostic images.
Overall image quality: rated from 1 to 5, with 1 = excellent image quality, and 5 non­diagnostic images.
SNR and ADC were measured in the head, body, and tail of the pancreas on a dedicated workstation (GE
Readyview, GE Healthcare, USA). Statistical analysis was performed using student t­test to compare the ADC and
SNR values and Wilcoxon Signed Rank Test to compare the scores on image qualities and artifacts.
Results & Discussion:
Large FOV and rFOV pancreas DWI were obtained from 10 patients. Pancreatic high resolution NT rFOV ss­DWI
was feasible in all patients. Average SNR and ADC values are reported in table I. Average SNR and ADC values
are reported in table I. There was no significant difference in the SNR between the two image datasets (figure 1)(all
p > 0.05, see table I). Average artifact score was 3.4 for large FOV and 2.0 for rFOV (figure 2), with a statistically
significant difference between the two image datasets p = 0.011. The ADC values in the body of the pancreas were
significantly lower when calculated from rFOV images (p=0.025). No statistically significant differences were found
between ADC values in the head and tail of the pancreas, but a trend was observed, with lower ADC values for
rFOV DWI. rFOV images had subjectively higher overall image quality (figure 3): average score for rFOV was 2.0
and for large FOV was 3.5 (p = 0.015).
Conclusion:
Our preliminary results show that rFOV DWI is feasible with similar SNR compared to large FOV DWI, and also
demonstrates higher overall image quality with reduced artifacts.
P127 SURGICAL STANDARDIZATION IMPROVES SURVIVAL IN PANCREATIC CANCER Daniel Delitto,
MD, Brian S Black, BS, Holly B Cunningham, BS, Sarunas Sliesoraitis, MD, Xiaomin Lu, PhD, Chen Liu, MD, PhD,
George A Sarosi, MD, Ryan M Thomas, MD, Jose G Trevino, MD, Steven J Hughes, MD, Thomas J George, MD,
Kevin E Behrns, MD; University of Florida, Gainesville, US
Introduction: Durable clinical gains are often associated with centralization of pancreatic surgery in the setting of
large, multi­institutional databases. We present a focused, stage­matched cohort of patients with pancreatic
adenocarcinoma (PC) who underwent pancreaticoduodenectomy (PD) prior to and after the implementation of a
pancreatic surgery partnership. We hypothesized that the standardization of surgical management would result in
improved long­term survival.
Methods: Data from 77 consecutive patients undergoing PD for PC were analyzed, representing all patients who
underwent PD and received long­term postoperative care at the University of Florida. Patients receiving neoadjuvant
therapy were excluded. Patients were divided into pre­ and post­standardization groups based on the timing of
partnership implementation and operative standardization. Primary outcomes included disease­free survival (DFS)
and overall survival (OS). Univariate and multivariate analyses were performed using Cox proportional hazards
models.
Results: Groups were similar with respect to age, BMI, comorbidities, stage and preoperative CA 19­9 levels. Major
operative differences post­standardization included a doubling in the mean number of lymph nodes obtained (17.6 vs.
9.1; P < .001), resulting in increased predictive power of nodal staging on DFS and OS. Despite similar rates of 30­
day postoperative complications, standardization resulted in a reduction in median length of stay (10 vs. 12; P =
.032). Surgical standardization of PC management resulted in increased DFS (17 vs. 11 mo; P = .017) and OS (26
vs. 16 mo; P = .004). The improvement in OS remained significant on multivariate analysis (HR = 0.46, P = .005).
Conclusion: Standardization in the surgical management of PC was associated with considerable gains in long­term
survival. These results reinforce the clinical benefit from standardization of pancreatic surgery in two controlled,
single­institution cohorts.
P128 SURVIVAL ANALYSIS FOR LOCALLY EXTENDING PANCREATIC CANCER PATIENTS Motokazu
Sugimoto, Joshua Barton, L W Traverso; St. Luke's Health System, Boise, US
Introduction: The treatment of pancreatic cancer is in evolution. What can we learn from the clinical outcomes using
NCCN Guidelines for pancreatic cancer?
Methods: Between Jun/2010 and Dec/2013 there were 226 pancreatic cancer patients seen at St. Luke’s Health
System. All had histological confirmation. Cases were categorized by tumor extension on the initial imaging studies:
locally confined, locally extending, or metastatic disease. Treatments were chosen using NCCN guidelines and the
overall survival (OS) was compared.
Results: Median OS was 9.7 mo for all cases (n = 226) ­ 16.1 mo for locally confined (n = 21), 11.6 mo for locally
extending (n = 125), and 5.0 mo for metastatic disease (n = 80). Not unexpectedly independent predictors for shorter
OS included ECOG ≥ 2, metastasis, and resection. To observe the true effect of non­operative protocols with
minimal selection bias we excluded those with ECOG ≥ 2 (n = 47), metastasis (n = 80), and resection (n = 33). In 71
cases receiving anti­cancer therapy, independent predictors for better OS was use of FOLFIRINOX (leucovorin +
fluorouracil + irinotecan + oxaliplatin) (P = 0.047), concurrent chemoradiation after induction chemotherapy (P =
0.007), and no progression of tumor ≥ 6 mo (P < 0.001). Parameters related to no tumor progression ≥ 6 mo were
age ≥ 68 yr (P = 0.046) and the initial use of FOLFIRINOX (P = 0.026). Within this group median OS for the cases
with initial use of FOLFIRINOX vs. gemcitabine­based was 17.9 mo vs. 12.2 mo (P = 0.026). Rates of patients with
ECOG 0 among those with initial treatment of FOLFIRINOX vs. gemcitabine­based regimens were 74% vs. 40% (P
= 0.010).
Conclusions: After excluding those with ECOG ≥ 2, metastasis, and resection we attempted to determine the true
outcome of non­operative treatment protocols in locally­extending pancreatic cancer. Survival was better if
FOLFIRINOX was used or if the tumor did not progress ≥ 6 months after beginning any chemotherapy treatment.
The former result should be used with caution as it appeared that the oncologist reserved the initial use of
FOLFIRINOX for ECOG 0 cases as allowed by NCCN guidelines. The 6 month no progression period is probably
reliable as the tumor can best be vetted under treatment if no progression is noted for ≥ 6 months. We speculate that
no progression after 6 months of treatment in a patient with a locally extending pancreatic cancer should be
selection criteria for potential resection.
P129 THE TRUE INCIDENCE OF MAIN DUCT INVOLVEMENT IN IPMN Trang K Nguyen, Gavin Falk, Daniel
Joyce, Gareth Morris­Stiff, R. Matthew Walsh; Cleveland Clinic, Cleveland, US
Introduction: While there are guidelines as to the management of Intraductal Papillary Mucinous Neoplasms (IPMN)
according to distribution, it is possible that a false impression of main duct involvement leading to excess resections
of side branch disease. The purpose of this study was to determine the accuracy of the pre­operative diagnosis
compared to final surgical pathology in identifying the distribution of IPMN within the pancreas.
Methods: Under IRB approval, a retrospective review of a prospectively collected database of patients who
underwent pancreatic resection for IPMNs was performed. The pre­operative assessment was made based on CT,
MRI, and/or endoscopic ultrasound evaluation. Patients with cytology or histology suggestive for malignancy either
pre­or post­operatively were excluded, as were those with IPMNs diagnosed on histology.
Results: From 2000 to June 2014, 128 patients underwent pancreatic resection for IPMNs. There was a strong
correlation between pre­ and post­operative diagnoses as to the distribution of the IPMN.
Conclusion: A pre­operative determination of distribution of IPMN based on radiographic and endoscopic findings is
accurate, or would not alter clinical management in the vast majority of cases.
P130 A GRADED EVALUATION OF OUTCOMES FOLLOWING PANCREATICODUODENECTOMY WITH
MAJOR VASCULAR RESECTION IN PANCREATIC CANCER: MAJOR VASCULAR RESECTION IS
ASSOCIATED WITH SEVERE ADVERSE POSTOPERATIVE OUTCOME AND EARLY RECURRENCE Olga
Kantor, MD1, Mark S Talamonti, MD2, Susan J Stocker, LPN2, Chi­Hsiung Wang, PhD3, David J Winchester, MD,
FACS2, Richard A Prinz, MD2, Marshall Baker, MD, MBA2; 1Department of Surgery, The University of Chicago
Medicine, 2Department of Surgery, NorthShore University HealthSystem, 3Center for Biomedical Research
Informatics, NorthShore University HealthSystem, Chicago, US
Introduction:
Recent multi­center retrospective studies in pancreatic cancer (PDAC) report disease specific survival following
pancreaticoduodenectomy with major vascular resection (PDVR) to be superior to that for palliative bypass and
comparable to that for pancreaticoduodenectomy not requiring vascular resection (PD). These studies have not
graded perioperative complications and provide incomplete assessments of the value of PDVR.
Methods:
We queried our institutional database identifying 24 patients undergoing PDVR for PDAC between 2007 and 2013.
Propensity score matching was used to match this cohort (3:1) by age, gender and tumor stage to 72 patients
undergoing PD in the same period. Charts were reviewed for all complications and 90­day readmissions. Clavien­
Dindo grade IIIb, IV, and V complications were classified as severe adverse postoperative outcomes (SAPO). Grade
I, II and IIIa complications requiring more than one interventional procedure or overall lengths of stay including
readmissions (LOS) >3 standard deviations beyond the mean for patients without complications were also classified
as SAPO. All others were considered minor adverse outcomes.
Results:
There were no statistical differences in demographics, comorbid disease, preoperative albumin, rates of R0
resection, use of neoadjuvant chemotherapy (NAC), or incidence of recurrent PDAC between groups. Patients
undergoing PDVR were more likely to have had antrectomy (75.0 vs 36.1%, p=0.001), had higher intraoperative
blood loss (1.3±1.1 vs 0.45±0.3L; p<0.001) and longer operative times (7.5±1.6 vs 5.8±1.1 hrs; p<0.001) than those
undergoing PD. PDVR patients were more likely to require readmission (41.7 vs 15.3%, p=0.01), demonstrated
longer LOS (22.2±15.8 vs 13.5±8.8 days, p=0.008), were more likely to have a SAPO (66.7 vs 19.4%, p<0.001) and
to miss adjuvant chemotherapy (33.3 vs 4.2%, p=0.001). Disease free and overall survival intervals were shorter in
the PDVR group (9.2±8.1 vs 18.9±17.1 months and 12.3±10.7 vs 24.2±17.7months; p≤0.002). Multivariate logistic
regression adjusted for age, comorbidities, hypoalbuminemia, NAC, tumor size and PDVR identified age ≥70 years
(OR 3.62 [1.04, 12.67]) and PDVR (OR 11.18 [2.98,41.89]) as independent predictors of SAPO. Cox­regression also
adjusting for SAPO identified PDVR (HR 2.11 [1.12,3.98]) and tumor size ≥3cm (HR 2.37 [1.48, 3.81]) as
independent predictors of long term overall mortality.
Conclusions:
PDVR results in a higher severity complication profile than that seen for PD. Patients requiring PDVR for PDAC are
less likely to receive adjuvant chemotherapy and demonstrate earlier disease recurrence than those undergoing PD.
Well powered trials carefully evaluating perioperative complications and long term outcomes are required to
determine the true value of PDVR for patients with resectable and borderline resectable PDAC. P131 AGING IS NOT ASSOCIATED WITH IMPAIRED PANCREATIC EXOCRINE FUNCTION FOLLOWING
PANCREATODUODENECTOMY USING NONINVASIVE 13C­MIXED TRIGLYCERIDE BREATH TEST
Masahiko Morifuji1, Kenichiro Uemura2, Yasushi Hsgimoto2, Yoshiaki Murakami2; 1Tsujinaka hospital, 2Department
of Surgery, Applied Life Sciences Institute of Biomedical & Health Sciences, Hiroshima Un, Kashiwa City, JP
Background:
Postoperative exocrine pancreatic insufficiency and resultant maldigestion is multifactorial in nature, mainly
influenced by patient­specific features of the pancreas; however, the impact of advancing age is less well
understood. The aim was to evaluate the effect of aging on postoperative digestive and fat absorptive disturbances
following pylorus­preserving pancreatoduodenectomy (PPPD).
Methods:
A prospectively collected, IRB approved database at a single institution was reviewed. This study included 51
patients who underwent PPPD from Jan 2005 to Aug 2009. Patients with an aged greater than or equal to 75 (elderly
group) were compared to those with an aged less than 75 prior to surgery (control group). An abdominal computed
tomography (CT) scan was obtained preoperatively and the diameter of the main pancreatic duct (MPD) was
measured on the presumed transection line of the pancreas. An optimized 13C­mixed triglyceride breath test [13C­
MTG­T] using a labeled long­chain triglyceride mixture was performed to assess postoperative fat absorptive
function after PPPD. Pancreatic exocrine insufficiency was defined as cumulative 7­hour 13CO2 exhalation [% dose
13C cum 7h] < 5%. Pre and postoperative HbA1c levels were measured in blood samples to assess glucose
metabolism function. Diabetic patients were identified as those treated with insulin, oral hypoglycemic medications,
or having an HbA1c level ≥ 6.9% (NGSP). Data pertaining [13C­MTG­T], HbA1c levels, oral pancreatic enzyme
requirements, and body mass index (BMI) were measured at 1 year following surgery. Post­operative fat absorptive
function was compared with pre­ and post­operative patient’s characteristics and glucose metabolism.
Results:
The elderly group (≥75 years) included 18 patients, while the remaining 33 patients were assigned as the control
group (<75 years). Main pancreatic duct was significantly low in the elderly group (4.84 ± 2.54) than the control group
(3.08 ± 2.29; P<0.05). The % dose 13C cum 7h was significantly higher in the elderly group (6.5 ± 5.1%) compared
to the control group (3.3 ± 2.4%; P<0.05). The number of patients requiring oral pancreatic enzyme was significantly
higher in the elderly group (12 of 18; 83%) comparing to the control group (11 of 33; 33%; P<0.05). The difference in
either HbA1c or body mass index (BMI) between the two groups is not statistically significant.
Conclusion:
Aging is not associated with impaired pancreatic exocrine function following PPPD compared to the younger
patients, and that cumulative 7­hour 13CO2 exhalation of 13C­mixed triglyceride breath test is an important
predictive marker of exocrine pancreatic insufficiency, even in a subclinical condition. These findings may have
potential implications for the selection of therapeutic strategies in the clinical setting.
P132 ARE THERE PREDICTORS OF INSULIN INDEPENDENCE IN PATIENTS UNDERGOING TOTAL
PANCREATECTOMY WITH ISLET AUTOTRANSPLANTATION? Katherine Engelhardt, MD, William P
Lancaster, MD, Hongjun Wang, PhD, David B Adams, MD, Katherine A Morgan, MD; Medical University of South
Carolina, Charleston, US
BACKGROUND: Selected patients with debilitating pain from chronic pancreatitis may benefit from total
pancreatectomy with islet autotransplantation (TPIAT) for pain relief but only a fraction will be insulin independent
long­term. Identifying predictors of insulin independence may aid in patient selection and counseling.
METHODS: A prospectively collected database of patients undergoing TPIAT from March 2009­May 2014 was
reviewed. Preoperative and perioperative variables were assessed in light of insulin requirement postoperatively.
RESULTS: 127 patients (76% women, mean age 40.5) underwent TPIAT, and 116 had at least 6­month follow­up
data. Twenty­two patients (19%) had diabetes preoperatively. Twenty­two patients (19%) had no insulin
requirement(NoIR), while 94 (81%) had some insulin use(IR) at last follow­up. The patients in the NoIR group were
younger (mean 36.4 v 41.7 years, p=0.05) and had a shorter disease duration (5.5 v 8.2 years, p=0.05) than those
that were IR. They had no significant difference in gender or etiology. There were trends in weight (68 v 75 kg),
tobacco use (23% v 36%), prior pancreas surgery (13% v24%), and surgeon rated gland character, although these
did not reach significance. The strongest predictor of insulin independence was islet yield (427721 v 215766 IEQ,
6338 v 3008 IEQ/kg, p=0.0001).
CONCLUSIONS: Insulin independence after TPIAT is more likely in patients that are younger and have a shorter
disease duration. Earlier intervention may be advantageous.
P133 CRITICAL SELF­ASSESSMENT OF WHIPPLE WITH VENOUS RESECTION Somala Mohammed, MD,
Amy McElhany, MPH, Charles A West, MD, Daniel Gonzales­Luna, BS, George Van Buren, II, MD, Courtney
Nalty, MPH, Eric J Silberfein, MD, Nader N Massarweh, MD, MPH, Alexander C Smith, BS, William E Fisher, MD;
Baylor College of Medicine, Houston, US
Background: Other series report no increased risk with the addition of en bloc segmental venous resection (VR) to
pancreaticoduodenectomy (PD) for venous involvement of pancreatic tumors. We analyzed our perioperative
morbidity, including vein patency rate and long­term survival following PD with VR.
Methods: 60­day postoperative outcomes for patients who underwent PD or PD+VR (2004­2013) were compared.
Two independent observers reviewed all available CT scans to determine long­term patency. The impact of VR on
OS was assessed in patients with pancreatic adenocarcinoma.
Results: 296 patients underwent PD (35 PD+VR). Patients undergoing PD+VR required longer operations (545±95
vs 426±113 mins, p<0.001) and had more blood loss (929±922 vs 432±457, p<0.001), but there was no difference in
60­day mortality, specific postoperative complication rates, graded severities of complications, reoperation,
readmission, or length of stay. VR involved portal vein (n=17) or superior mesenteric vein (n=16) or both (n=2) and
required either primary repair (n=19) or interposition graft (jugular vein) (n=16). Reconstructions remained patent in
93% of living patients at 3 months, 85% at 12 months, and 78% at 24 months postoperatively. Among 111
adenocarcinoma patients, median OS was 17.8 months. Patients in the PD+VR group were more likely to receive
neoadjuvant therapy than patients in PD­alone group (40.7% vs 6.5%, p<0.001), but there was no difference
between the groups in terms of R0 resection rate, tumor size, involved LN ratio or median OS.
Conclusions: PD+VR provides acceptable perioperative and oncologic outcomes. Venous thrombosis affects
~15% of patients by 1 year, warranting further evaluation of anticoagulation strategies.
P134 DISEASE RECURRENCE AFTER LONG­TERM SURVIVAL: THE NEW REALITY OF PANCREATIC
CANCER? Alessandra Landmann, MD, Russell G Postier, MD; University of Oklahoma Health Sciences Center,
Oklahoma City, US
Introduction­
We present two cases of long­term survivors of pancreatic adenocarcinoma who developed disease recurrence
greater than 60 months from their initial operation.
Methods­
Case 1:
JS is a 55­year­old female who initially presented with locally unresectable pancreatic cancer. She received
neoadjuvant therapy and underwent pyloric­preserving pancreaticoduodenectomy. Pathology
revealed adenocarcinoma staged T1N0. She received two cycles of adjuvant gemcitabine chemotherapy. She
returns to clinic 60 months after surgery with complaints of weight loss and abdominal pain and was found to have
locoregional recurrence. She is currently receiving gemcitabine and abraxane chemotherapy.
Case 2:
JE is a 64­year­old male who underwent pylorus­preserving pancreaticoduodenectomy and was found to have
moderately differentiated adenocarcinoma staged T2N0. He received gemcitabine, 5FU and radiation. On routine
follow up, he was found to have a right upper quadrant mass concerning for malignancy nine years after resection.
He underwent an omentectomy with en bloc removal of the mass. Pathology revealed recurrent pancreatic cancer.
He is currently receiving gemcitabine.
Results­
Long­term survival after pancreaticoduodenectomy, considered to be survival longer than 60 months, is rare, with
mean survival ranging from 8­12 months. Only 15% of patients survive 60 months and fewer patients, 5.9%, survive
10 years or longer. Unfavorable characteristics for long­term survival include increased tumor diameter, lymph node
metastases, number of malignancy­positive lymph nodes, decreased serum albumin concentration, and
intraoperative packed red blood cell transfusion. On multivariate analysis, lymph node staging was found to be the
only independent prognostic factor for long­term survival. Lymph node status has been shown to be more significant
than margin status in terms of survival.
Conclusion­
Pancreatic cancer is considered a systemic disease at diagnosis. Long­term survival after pancreatic resection does
not imply a systemic cure. With increased survival after pancreaticoduodenectomy and improved chemoradiation,
disease recurrence after long disease free interval is a distinct reality in current treatment of pancreatic cancer.
P135 EARLY DRAIN REMOVAL IS A BEST PRACTICE IN SELECTED PANCREATIC SURGERY PATIENTS
Henry A Pitt, MD1, Benjamin L Zarzaur, MD2, Stephen W Behrman, MD3, E M Kilbane, RN2, Bruce L Hall, MD,
PhD, MBA4, Abhishek Parmer, MD5, Taylor S Riall, MD, PhD5; 1Temple University School of Medicine, 2Indiana
University School of Medicine, 3University of Tennessee College of Medicine, 4Washington University School of
Medicine, 5University of Texas Medical Branch, Philadelphia, US
Background: The morbidity of pancreatic surgery remains unacceptably high. Recent reports suggest that drain
management may influence postoperative complications. In one randomized controlled trial (RCT) outcomes were
significantly worse in pancreatoduodenectomy patients managed without drains. Another RCT of patients with low
postoperative day one (POD#1) drain amylase demonstrated that pancreatic fistulas were dramatically reduced
when drains were removed early. While, approximately 30% of pancreatectomy patients have very low POD#1
drain amylase, pancreatic surgeons have been slow to adopt early drain removal. Therefore, the aim of this analysis
was to compare outcomes when drains were removed early or late after pancreatic surgery. Methods: Data were gathered through the American College of Surgeons – National Surgical Quality Improvement
Program (ACS­NSQIP) Pancreatectomy Demonstration Project (PDP). Over a 14­month period, 2,805 patients
underwent a pancreatoduodenectomy, distal pancreatectomy, total pancreatectomy or pancreatic enucleation at 43
institutions. After exclusion of patients without drains or data on drain removal, 1,841 patients were available for
analysis. Early drain removal was defined as on or before POD#3. Early drain patients (n=148, 8.1%) did not differ
from late drain removal patients with respect to multiple variables but were younger, had more women and were less
likely to undergo a pancreatoduodenectomy. Therefore, propensity score matching and sensitivity analyses were
performed. Outcomes were determined by ACS­NSQIP and PDP definitions. Standard statistical tests were applied.
Results: After propensity score matching, early (n=127) and late (n=127) drain removal patients were well­balanced
for age, gender, BMI, serum albumin, ASA class, operation type, time and approach, vascular resection, gland
texture, duct size, pathology and POD#1 drain amylase. Outcomes were: Conclusions: Early drain removal is associated with fewer pancreatic fistulas, surgical site infections, serious and
overall morbidity. Selected pancreatectomy patients should be managed with early drain removal.
P136 FEEDING JEJUNOSTOMY CONSIDERATIONS IN HIGH­RISK PATIENTS UNDERGOING
PANCREATICODUODENECTOMY Cornelius A Thiels, DO, MBA, Christopher R Shubert, MD, Daniel S Ubl, BA,
John R Bergquist, MD, Michael L Kendrick, MD, Mark J Truty, MD, Elizabeth B Habermann, PhD; Mayo Clinic,
Rochester, US
Introduction: While placement of feeding jejunostomy (FJ) during pancreaticoduodenectomy (PD) is declining,
surgeons use them selectively, particularly for high­risk patients. We aimed to determine outcomes of selective use
of FJ in high­risk patients.
Methods: ACS­NSQIP was reviewed for all PD’s from 2005­2012. Patients who underwent concurrent FJ
placement were identified by CPT4 code. Multivariable analyses controlling for peri­operative risk factors (CPT4
code, age, weight loss, chemotherapy/radiation, albumin, bilirubin, BMI, DM, smoking, and COPD) were utilized to
compare thirty­day outcomes for patients with and without FJ. High­risk diagnosis was defined as bile duct and
ampullary, duodenal, and neuroendocrine neoplasms.
Results: Of 7120 patients, 2362 were considered high­risk and 332 (14.1%) had FJs placed. In the last five years,
FJ placement decreased overall (p=0.0264) while remaining stable in the high risk group (p=0.4675). Patients in the
high­risk group with FJ were more likely to have prolonged length of stay (LOS) (OR: 2.19, 95% CI 1.72­2.80,
p<0.001), increased major complications (OR: 1.71, 95% CI 1.34­2.18, p<0.001), and higher 30 day mortality
compared to high­risk patients without FJ (OR: 2.19, 95% CI 1.27­3.77, p<0.005) on multivariable analysis.
Conclusion: Despite an overall decrease in FJ placement, its use in high­risk PD has remained stable. Our
analysis demonstrates that in patients with a high­risk diagnosis, the use of FJ is associated with patients that are
more likely to have increased LOS, complications, and mortality after PD. Reasons for worse outcomes, whether
specific to FJ­related complications or selection bias in those patients receiving FJ at PD is unknown, however the
routine use of FJ in high­risk PD should be carefully considered given these findings. P137 IMPACT OF INTRATUMORAL NERVE GROWTH FACTOR EXPRESSION ON PERINEURAL INVASION
AND PROGNOSIS IN RESECTABLE EXTRAHEPATIC CHOLANGIOCARCINOMA Kazuhide Urabe, MD,
Yoshiaki Murakami, Kenichiro Uemura, Yasushi Hashimoto, Naru Kondo, Naoya Nakagawa, Hayato Sasaki, Taijiro
Sueda; Hiroshima University, Hiroshima, JP
Background
Perineural invasion is one of the risk factors of poor survival in extrahepatic cholangiocarcinoma. Intratumoral
expression of nerve growth factor (NGF), which is a crucial neurotrophic factor of nerve growth and proliferation in
proto neural crest, has been reported to be associated with perineural invasion in several kinds of cancers.
However, it is still unclear whether intratumoral NGF expression impacts on perineural invasion and survival in
extrahepatic cholangiocarcinoma. The aim of this study was to investigate the association of NGF expression in
resected specimens with perineural invasion and survival of patients who underwent surgical resection for
extrahepatic cholangiocarcinoma.
Materials and Methods
Records of 112 patients with extrahepatic cholangiocarcinoma (including 53 with distal cholangiocarcinoma and 59
with perihilar cholangiocarcinoma) who underwent surgical resection between September 1999 and April 2014 were
reviewed retrospectively. Intratumoral NGF expression were investigated using immunohistochemical technique.
Relationships between NGF expression and clinicopathological factors including perineural invasion in resected
specimen were statistically evaluated, and risk factors for poor survival of patients with resectable extrahepatic
cholangiocarcinoma were analyzed using univariate and multivariate analyses.
Results
High and low intratumoral NGF expression was observed in 62 (55%) and 50 (45%) patients, respectively. For all
112 patients, no significant differences was found between NGF expression and presence of perineural invasion (P
= 0.94). Moreover, intratumoral NGF expression was not associated overall survival (P = 0.97). In multivariate
analysis, presence of perineural invasion (hazard ratio [HR] 3.53: 95% confidence interval [CI] 1.52­10.3; P =
0.0021) and lymph node metastasis (HR 2.47; 95% CI 1.56­5.08; P = 0.0005) are identified significant as
independent risk factors for poor overall survival.
Conclusion
The current results suggested that presence of perineural invasion in extrahepatic cholangiocarcinoma was an
independent risk factor of poor survival of patients with resectable extrahepatic cholangiocarcinoma, however, NGF
expression was probably not associated with perineural invasion, and did not impact on their survival.
P138 INCREASED MORBIDITY AND MORTALITY OF CONCOMITANT COLECTOMY DURING
PANCREATICODUODENECTOMY: A NSQIP PROPENSITY SCORE MATCHED ANALYSIS Jennifer W
Harris, MD, Jeremiah T Martin, MD, Erin C Maynard, MD, Patrick C McGrath, MD, Ching­Wei D Tzeng, MD;
University of Kentucky, Lexington, US
Background: Select patients with locally aggressive periampullary cancers require concomitant colon resection (CR)
during pancreaticoduodenectomy (PD) for margin­negative resections. Past single­institution studies have suggested
that major morbidity and mortality rates are not higher in these patients compared to those undergoing standard PD.
This study sought to analyze the impact of CR on post­PD major morbidity and mortality in a broader patient
sampling using a large national dataset.
Methods: All National Surgical Quality Improvement Program (NSQIP) patients undergoing PD for periampullary
cancers (with/without CR) from 2005­2012 were initially screened. A 4:1 propensity score matched analysis was
constructed to identify the impact of CR upon PD. Risk factors for 30­day major morbidity (defined using NSQIP
parameters for pneumonia, re­intubation/ventilator dependence, renal insufficiency/failure, cardiac events,
neurological events, sepsis/septic shock, return to the operating room, dehiscence, organ space infection, and
venous thromboembolism) and mortality were analyzed to determine the postoperative sequelae of adding CR to
PD.
Results: Of 10,965 PD and 159 PD+CR patients in total, 624 and 156, respectively, were selected for the 4:1
matched analysis. PD+CR resulted in significantly higher major morbidity and mortality (50.0% and 9.0%) vs. PD
alone (28.8% and 2.9%, respectively, p<0.001).
Multivariate analysis identified the following independent risk factors for major morbidity after PD: concomitant CR
(odds ratio, OR­3.19, p<0.001), smoking history (OR­1.92, p=0.005), lack of functional independence (OR­3.29,
p=0.018), cardiac disease (OR­2.39, p=0.011), decreased albumin (per g/dL, OR­1.38, p=0.033), and longer
operative time (vs. median time, OR­1.56, p=0.029). Independent predictors of mortality included concomitant CR
(OR­3.16, p=0.010), ventilator dependence (OR­13.87, p<0.001), and septic shock (OR­6.02, p<0.001).
Conclusions: Contrary to previous single­institution studies, this propensity score matched analysis using the
NSQIP dataset showed that adding CR to PD significantly increased the magnitude of surgery and was an
independent predictor of both major morbidity and mortality. To improve surgical outcomes, patients who may need
PD+CR should be preoperatively identified using high­resolution imaging, maximally optimized with prehabilitation,
and referred to expert centers.
P139 LAPAROSCOPIC PANCREATICODUODENECTOMY FOR PANCREATIC CANCER – A NATIONWIDE
ANALYSIS Pragatheeshwar Thirunavukarasu, MD1, Emmanuel Gabriel, MD1, Kristopher Attwood, Phd2, Steven
Nurkin, MD1; 1Roswell Park Cancer Institute, 2University at Buffalo, Buffalo, US
Background: Laparoscopic pancreaticoduodenectomy (LP) is an increasingly adopted alternative to traditional open
resection for pancreatic cancer. LP has been demonstrated to be safe by few high volume centers specializing in
minimally invasive surgeries.
Methods: Data for patients who underwent laparoscopic pancreaticoduodenectomy between years 2010 and 2011
for pancreatic cancer were extracted from the National Cancer Database. Patients who underwent open
pancreaticoduodenectomy (OP) during the same time period were used for comparison.
Results (Table 1): Of 6,298 patients who underwent pancreaticoduodenectomy for pancreatic cancer, 6,130 (97.3%)
had invasive cancer, and 168 (2.7%) had in situ cancer. Median age was 66 years, 50.8% were male, and 86.1%
were Caucasian. In the entire study cohort, 11.3% and 7.0% had undergone chemotherapy and radiation prior to
surgery, respectively. 831 (13.2%) underwent laparoscopic pancreaticoduodenectomy and 5,467 (86.8%) underwent
open resection. The conversion rate for laparoscopic surgery was 29.7% . The proportion of patients undergoing
laparoscopic surgery increased from 10.8% in 2010 to 15.5% in 2011 (P<.001). LP and OP patient cohorts were
similar in terms of median age, sex distribution, racial distribution, insurance status, percentage of in­situ cancers,
histological grade, TNM stage, and in the proportion of patients who underwent neoadjuvant chemotherapy or
radiation. OP cohort had more patients who had 2 or more medical comorbidties than LP cohort (7.4% vs. 6.6%, P =
0.002). Positive margin resection rate was lower with LP than with OP (19.7% vs. 23.3%, P = 0.02). Median hospital
length of stay was lower with LP compared to OP (8 vs. 9 days, P <.001), but there was no difference in 30­day
unplanned readmission rate (8.0% vs. 9.2%, P = 0.23). Univariate analysis showed that patients advanced age (OR
1.04, P<.001), patients with 2 or more medical comorbidities (OR 2.15, P<.001), and margin­positive resection rate
(OR 1.45, P = 0.01) were associated with higher risk for 30­day mortality. There was a non­statistically significant
trend towards higher 30­day mortality with LP compared to OP (4.8% vs. 3.5%, P = 0.06). On multivariate analysis,
approach of surgery (laparoscopic vs. open) did not emerge as a significant factor affecting 30­day mortality (LP vs.
OP, OR 1.3, 95% CI 0.89 – 1.98, P = 0.17).
Conclusion: Laparoscopic approach is an increasing performed alternative for pancreaticoduodenectomy.
Laparoscopic pancreaticoduodenectomy is safe, and offers comparable operative outcomes compared to open
approach, with a potential for shorter length of stay. This is the first nationwide report of laparoscopic
pancreaticoduodenectomies.
Table 1
Outcome
Lapaproscopic
Pancreaticoduodenectomy
N = 5,467
Open Pancreaticoduodenectomy
N = 831
P­value
Margin positive rate
19.7%
23.3%
0.02
Median length of stay
8.0 days
9.0 days
<.001
30­unplanned readmission
rate
8.0%
9.2%
0.23
30 ­ day mortality
4.8%
3.5%
0.06
P140 LONG TERM FOLLOW UP AFTER RESECTION OF RENAL CELL CARCINOMA METASTASIS TO THE
PANCREAS Marius Distler, MD1, Felix Rückert, MD2, David Ollmann, MD2, Patrick Teoule, MD2, Thorsten
Wilhelm, MD2, Robert Grützmann, Prof1; 1Department of General, Thoracic and Vascular Surgery, University
Hospital Carl Gustav Carus, Technis, 2The Department of Surgery, University Medical Centre Mannheim, Medical
Faculty Mannheim, Heidelberg, Dresden, DE
Background
Metastases from renal cell cancer (RCC) have a high affinity for the pancreatic gland. Previous reports showed an
excellent survival for patients after resection of such metastases to the pancreas and several predictive factors were
reported. Although RCC is the most common primary tumor metastasizing to the pancreas this is a rare event and
only about two to three hundred cases were reported in the last 60 years. The present study aims to give more
evidence to reported risk factors by analysing the biggest cohort of patients with pancreatic resection due to
pancreatic renal cell carcinoma metastases (pRCC) so far. Patients and Methods
We retrospectively analyzed all pancreatic resections due to pRCC between January 1993 and October 2014 in two
German centres for pancreatic surgery. Predictive factors were analysed using chi square test.
Results
Surgery was performed in 40 patients, as follows: 15 pancreatic head resections 12 distal pancreatectomies, 9
pancreatectomies, three segmental resections and one papillectomy. The mean age was 65.5 (SD 9.0). Mean time
of between resection of the primary tumor and the diagnosis of the metastases was 125.4 months (SD 77.4). Mean
survival was 147.9 months (SD 25.6 months). Statistical analysis showed that none of the analysed parameters had
predictive value.
Conclusion
Retrospective studies give the opportunity to study large patient cohorts even in rare diseases. Although our
analysis comprised the biggest cohort of patients with pRCC it rendered no significant predictor. This might be due
to the excellent prognosis of our patients in connection with the relatively small number of patient. According to our
data, an aggressive approach to pRCC can be recommended as prognosis after resection is excellent.
P141 METABOLIC SYNDROME IS ASSOCIATED WITH INCREASED POST­OPERATIVE MORBIDITY AND
HOSPITAL RESOURCE UTILIZATION IN PATIENTS UNDERGOING ELECTIVE PANCREATECTOMY: A
NSQIP STUDY OF 16,562 CASES May C Tee, MD, MPH1, Daniel S Ubl, BA2, Elizabeth B Habermann, PhD,
MPH2, David M Nagorney, MD1, Michael L Kendrick, MD1, Michael G Sarr, MD1, Mark J Truty, MD, MS1,
Florencia G Que, MD1, KMarie Reid­Lombardo, MD, MS1, Rory L Smoot, MD1, Michael B Farnell, MD1; 1Mayo
Clinic, Department of Surgery, Division of Subspecialty General Surgery, 2Mayo Clinic Robert D. and Patricia E.
Kern Center for the Science of Health Care Delivery, Rochester, US
Purpose:
Obesity, insulin resistance, and cardiovascular risk factors are associated with poorer peri­operative outcomes for
many surgical procedures. In patients undergoing elective pancreatectomy (distal, proximal, and total), our aim was
to evaluate the effect of the Metabolic Syndrome (MS) on postoperative mortality, morbidity, and hospital resource
utilization. Our hypothesis was that MS is associated with worse surgical outcomes following pancreatectomy.
Methods:
16,562 patients undergoing elective pancreatectomy from 2005 through 2012 were identified in the Participant Use
File of the American College of Surgeons National Surgical Quality Improvement Program (NSQIP). Univariable and
multivariable analyses were performed examining the association of MS (defined as body mass index ≥ 30 kg / m2,
hypertension requiring medications, and diabetes requiring medications and / or insulin) and risk of 30­day mortality,
morbidity, and hospital resource utilization (risk of blood transfusion in the first 72 hours after surgery and prolonged
hospital stay, defined as ≥ 13 days, which was the 75th percentile of this cohort). Multivariable logistic regression
models controlled for age, sex, race, procedure, smoker status, ethanol use, COPD, functional status, steroid use,
albumin, INR, creatinine, bilirubin, hematocrit, and any cardiac or vascular disease.
Results:
1,113 (6.7%) patients had MS. MS was statistically significantly associated with increased postoperative morbidity,
major morbidity, surgical site infection, sepsis, cardiac event, respiratory failure, pulmonary embolism, blood
transfusion, and prolonged duration of hospital stay (p<0.05 for all analyses). After controlling for potentially
confounding variables, there was a 19% increased odds of major morbidity (p=0.017), 33% increased odds of
surgical site infection (p<0.001), 38% increased odds of respiratory failure (p=0.011), 26% increased odds of blood
transfusion (p=0.016), and 20% increased odds of prolonged hospital stay (p=0.015) in patients with MS compared
to patients without MS. MS was not significantly associated with 30­day mortality following elective pancreatectomy
(p=0.340). Subgroup analysis was performed by malignant diagnosis. Patients with MS and benign disease had 37%
increased odds of major morbidity (p=0.016) and 52% increased odds of blood transfusion (p= 0.027) on
multivariable analyses. For patients with malignant disease, MS was not significantly associated with increased risk
of major morbidity (p=0.189) nor blood transfusion (p=0.107).
Conclusion:
Metabolic Syndrome is an under­emphasized predictor of increased postoperative morbidity and hospital resource
utilization in patients undergoing elective pancreatectomy. The effect of MS on these postoperative outcomes
appears to be more pronounced for patients with benign rather than malignant indications for pancreatectomy, which
may help inform patient selection and optimization of co­morbidities prior to elective surgery.
P142 ORTHOGRADE PLANIMETRY OF THE PANCREATIC TRANSSECTION PLANE BY USING
PREOPERATIVE COMPUTERTOMOGRAPHY IMAGING FOR RISK EVALUATION OF POSTOPERATIVE
PANCREATIC FISTULA AFTER PANCREATIC HEAD RESECTION Ulrich Adam, Prof, Dr1, Colin M Krüger,
Dr1, Karsten Krüger2, Frank Makowiec, Prof, Dr3, Hartwig Riediger1; 1Humboldt­Klinikum, Klinik für Chirurgie,
2Humboldt­Klinikum, Klinik für Radiologie, 3Universitätsklinikum Freiburg, Chirurgische Klinik, Berlin, DE
Introduction/Background
Postoperative pancreatic fistula is a relevant complication after pancreato­duodenectomy. Therefore, preoperative
detection of high risk patients may be important. We evaluated preoperative CT­imaging by planimetry at the
expected resection plane along the superior mesenteric vein and correlated the results with the incidence of
postoperative pancreatic fistula.
Methods
From 2009 to 2013, 123 patients with pancreatoduodenectomy underwent uniform preoperative imaging and
reconstruction of the pancreatojejunostomy. Planimetry was performed at a mulitplanar reconstruction of the
pancreatic transsection plane (diameter, range, duct width, area) as well as the calculation of ratios (duct
width/pancreatic diameter; D/P­ratio). The measured values were correlated with the incidence of postoperative
pancreatic fistula.
Results
Planimetry showed a significant difference of the pancreatic transsection plane in relation to the incidence of
postoperative pancreatic fistula. A thick parenchyma and a tiny duct are significant factors of risk. Up to 94% of
patients with postoperative pancreatic fistula had a duct width of less than 20% of the pancreatic diameter (D/P­ratio
<0,2; p<0,01).
Discussion/Conclusion
The incidence of postoperative pancreatic fistula correlates significantly with the morphology of the pancreatic
transsection plane. The risk increases significantly with a D/P­ratio of <0,2.
P143 PANCREAS SURGERY IN GERMANY ­ ANALYSIS OF ALL GERMAN PANCREATIC RESECTION
2009­2012 Robert Grützmann1, Christian Krautz1, Marius Distler, MD1, Ulrike Nimptsch2, Thomas Mansky2;
1University Hospital Dresden, 2Technical University Berlin, Dresden, DE
Introduction:
Pancreatic resection are complex operations with high morbidity and mortality. In most publications published
mortality in centers of pancreatic surgery is below 5%. On the other hand Birkmeyer et al., showed a clear
correlation between mortality and hospital and surgeon volume. There are no German­wide data until now. The
German DRG­system allows to analyse every signle case of pancreatic operation in Germany.
Patients and methods:
The DRG­statistics of Germany has been analysed for the years 2009 to 2012 kumulatively. All cases with
pancreatic resection have been identified using coded procedures. The following procedure have been included: total
and partial pancreatectomies and local excisionsen. In house mortaility of the hospital has been determined as
inidicator of quality.
Results:
From 2009 to 2012 45,678 resections of the pancreas have been performed in Germany. Among them were 78%
partial Pankreatektomien, 10% local resections, 8% total pancreatectomies und 4% others operation. The most
common diagnosie are malignant tumors of the pancreas, duodenum or bile duct (54%), benign or uncertain tumors
(8%) and chronic pancreatitis with 8%. In those years in hospital mortality after pancreatic resection in Germany
was 10%. Detailed data will be presented.
Discussion:
For the first time complete data concerning pancreatic surgery in Germany have been evaluated. Main indication for
pancreatic surgery are: malignant tumors of the pancreas, duodenum or bile duct (54%), benign or uncertain tumors
(8%) and chronic pancreatitis with 8%. In hospital mortality after pancreatic resection in Germany was 10%.
P144 PANCREATIC RESECTION FOR MUCINOUS NEOPLASM INTRADUCTAL PAPILLARY ­
EXPERIENCE OF 10 YEARS S Corado, MD, E. Vigia, E Filipe, A Nobre, L Bicho, J Paulino Pereira, A Martins, E
Barroso; Hospital de Curry Cabral, Lisbon, PT
INTRODUCTION ­ Pancreatic resection for cystic tumors has been increasing in recent years, mainly due to
mucinous neoplasm Intraductal Papillary (IPMN) that are broad spectrum dysplastic lesions.
MATERIALS AND METHODS ­ We analyzed our series of patients undergoing surgery for pancreatic last 10 years
for the clinical and pathological variables relevant for prognosis. SPSS was used 20 for statistical analysis.
RESULTS ­ Between June 2004 and June 2014 were operated 59 patients with the diagnosis of IPMN. Of these,
26 (44%) had invasive component and 38 (64%) were located in the head of the pancreas; 54% of surgeries
performed (n = 32) were Duodenopancreatectomias cephalic. The Overall Survival (OS) was 58.9% and 53% at 3
and 5 years; in the subgroup of patients with IPMN Invasive SG decreases to 43% and 32% at 3 and 5 years,
respectively. In univariate analysis, influenced the SG: the presence of invasive disease (HR 2.81; 95% CI 1.07 to
7.4), males (HR 4:49; 95% CI 1.3­15.4), age (HR 1.08, 95% CI 1.02­ 1.15) and vascular resection (HR 3.27, 95% CI
1.17 to 9.15). The adenocarcinoma subtype did not prove statistically significant nor for SG or to the disease­free
survival.
DISCUSSION ­ The prognosis of IPMN is better than that of pancreatic adenocarcinoma. Older age, male sex,
presence of invasive disease and vascular need for resection can influence Survival.
P145 PERCEPTION IS REALITY: QUALITY METRICS IN PANCREATIC SURGERY­ A CENTRAL
PANCREAS CONSORTIUM (CPC) ANALYSIS OF 1399 PATIENTS De Abbott1, Da Kooby2, Nb Merchant3, Mh
Squires2, Sk Maithel2, Sm Weber4, Er Winslow4, Cs Cho4, Dj Bentrem5, Hj Kim6, Cr Scoggins7, Rc Martin7, Aa
Parikh3, Wg Hawkins8, G Martin1, Sa Ahmad1; 1University of Cincinnati, 2Emory University, 3Vanderbilt University,
4University of Wisconsin, 5Northwestern University, 6University of North Carolina, 7University of Louisville,
8Washington University, Cincinnati, US
Introduction/Background: Because variability in outcomes exists between centers performing pancreatic surgery,
several groups have defined quality metrics that identify centers delivering quality care. Although these metrics are
perceived to be associated with good outcomes, their relationship with actual outcomes has not been established. Methods: We surveyed a national cadre of pancreatic surgeons regarding perceived quality metrics. The
performance of these metrics were then evaluated against the database of eight high volume institutions to
determine how often they were being performed and when possible their relationship with long­term outcomes. Results: Overall, 103 pancreatic surgeons responded to the survey. Based on this, the top five important metrics
were perceived to be multidisciplinary care, case volume, mortality rates, margin status, and complications rates. Other factors included rate of LN harvest and timing of adjuvant therapy. Subsequent analysis using the multi­
institutional dataset of 1399 patients demonstrated that all institutions had the availability of a multidisciplinary team
and supporting infrastructure, including institutional monitoring of surgeon and center outcomes. For the entire
cohort, median survival was 19.7 months and perioperative mortality was 2.9%. A R0 retroperitoneal and neck
margin was obtained in 81% (n=1109) and 91.4% (n=1278) of cases, respectively. 78% of patients (n=1091) had
greater than 10 lymph nodes harvested, and LN positivity was present in 71% (n=902). 74% (n=960) of patients
received adjuvant therapy within 60 days of surgery. Multivariate analysis demonstrated margin status, identification
of greater than 10 lymph nodes, nodal positivity, and delivery of adjuvant therapy within 60 days to be associated
with improved overall survival. Discussion/Conclusions: These analyses demonstrate that systematic monitoring of surgeons’ perceived quality
metrics provides critical prognostic information, which is associated with improved patient survival. Conducting and
documenting such metrics can identify centers delivering high quality care. P146 POST­OPERATIVE OMENTAL INFARCTION IN PATIENTS UNDERGOING DISTAL
PANCREATECTOMY: CT IMAGING APPEARANCE, ETIOLOGY AND MANAGEMENT Ammar A Javed,
MBBS, Fabio Bagante, MD, Ralph H Hruban, MD, Matthew J Weiss, MD, Martin A Makary, MD, MPH, Kenzo
Hirose, MD, Christopher L Wolfgang, MD, PhD, Elliot K Fishman, MD; Johns Hopkins Hospital, Baltimore, US
Introduction
The clinico­radiological characteristics and the natural history of post­operative omental infarction in patients who
underwent distal pancreatectomy and splenectomy (DP) have not been defined. We here describe both of these
along with the important features in diagnosis and management of post­operative omental infarction.
Methods
Twelve patients treated with pancreatic surgery for pancreatic cancers over a period of 3 years were diagnosed with
omental infarction based on the findings on a computer tomography (CT). The clinical data and radiological imaging
were retrospectively collected and analyzed. Additionally, an extensive literature search was performed to identify
previous reports on omental infarction.
Results
A total of 12 patients were diagnosed with omental infarction based on their post­operative imaging. Seven (61.5%)
patients underwent laparoscopic DP and spleenectomy, one (7.7%) patient underwent robotic DP and spleenectomy
and in one (7.7%) patient laparoscopic DP was converted to an open procedure. The remaining three (23.1%) were
treated with open DP and spleenectomy. There was no difference in the surgical technique used in the procedures
that were performed via the open and the minimally invasive approach. In seven (58.3%) patients the diagnosis of
omental infarction was made during their regular follow­up. The post­operative CT scans were performed prior to the
regular follow in five patients because of abdominal pain (two patients), nausea and emesis (two patients) and fever
(one patient). At the time of diagnosis the mean size of the omental infarction was 10.6±5.14 cm as compared to
10.7 ± 3.6 cm at the time of last follow up (8.6±6 months after the diagnosis). One patient underwent surgical
resection of the omental infarction, two patients were readmitted and a drain was placed in the mass. The remaining
nine patients underwent conservative management. A systematic review helped identify nine articles that reported a
total of 34 patients who were diagnosed with omental infarction post­operatively after having abdominal surgery.
Conclusion
Omental infarction following distal pancreatectomy can be either asymptomatic and be found incidentally during the
regular follow­up or present with symptoms. The radiological findings of a large mass can be confused with disease
recurrence or post­operative abscess. If possible, the management of a postoperative omental infarction should be
conservative and operative procedure should be taken into account only in patients who are symptomatic or have
infected omental infarction.
P147 PREDICTORS OF POSTOPERATIVE OUTCOME AFTER DISTAL PANCREATECTOMY: THE ANSWER
FROM TWO HIGH­VOLUME INSTITUTIONS Giovanni Marchegiani1, Rafael Pieretti­Vanmarcke2, Giuseppe
Malleo1, Francesca Panzeri1, Tiziana Marchese1, Giovanni Butturini1, Roberto Salvia1, Andrew L Warshaw2, Keith
Lillemoe2, Carlos Fernandez­del Castillo2, Claudio Bassi1, Cristina R Ferrone2; 1Università di Verona,
2Massachusetts General Hospital, Verona, IT
Introduction/Background: Distal pancreatectomy (DP) is the treatment of choice for the surgical excision of
masses in the body­tail of the pancreas and its outcomes correlate with surgical volume. There is limited information
about reliable predictors of postoperative complications in this setting.
Methods: All patients undergoing DP at two high­volume Institutions between 2004 and 2013 were considered for
the analysis. Retrospective analysis of predictors of postoperative outcomes was performed. In particular, we
investigated factors associated with major complications (MC, Clavien­Dindo >=3)1 and with clinically relevant
pancreatic fistula (CRPF, grade B/C according to the ISGPF).
Results: Nine hundred forty­six DPs were performed during the study period. Of these, 59% were female, median
age was 57 (8­93) and median BMI 25 (15­49). DPs were associated with splenectomy in 81% of cases and
performed with minimally invasive techniques in 21%. Overall, median operative time was 200 minutes (54­660). The
rate of MC and CRPF were 16% and 17%, respectively. Mortality was 1.4%. At multivariate analysis, independent
predictors of MC were operative time (P=0.004), ASA (P=0.001) and intraoperative transfusion (P=0.001), while
minimally invasive surgery and multi­visceral resections were not. Predictors of CRPF were BMI (P=0.001),
intraoperative transfusion (P=0.03) and pancreatic ductal adenocarcinoma as final diagnosis (P=0.03). Minimally
invasive surgery and the stump closure technique did not affect the rate of CRPF.
Discussion/Conclusion: This large bicentric series confirms that DP can be safely performed at high­volume
Institutions, with satisfactory morbidity and mortality rates. Postoperative complications can be predicted by BMI,
ASA, operative time, intraoperative transfusions and pathology diagnosis. Minimally invasive surgery, stump closure
technique and multi­visceral resections do not significantly affect the outcome. P148 PRE­OPERATIVE PHYSICAL STATUS AND PERI­OPERATIVE MORBIDITY AND MORTALITY IN
PATIENTS UNDERGOING MAJOR PANCREATIC SURGERY Camilla Cena, MD1, Davide Cigolini, MD1,
Roberto Salvia, PhD1, Vittorio Schweiger, MD1, Paolo Regi, MD2, Walter Mosaner, MD2, Enrico Polati, FACS,
PhD1, Claudio Bassi, FACS, PhD1; 1University of Verona, 2Casa di Cura Pederzoli, Peschiera del Garda, Verona,
IT
Introduction: The aim of our study was to verify what are the more important pre­operative variables that affect the
patient outcomes after major pancreatic surgery. Preoperative risk stratification can be useful in choosing the whole
therapeutical strategies for patients with pancreatic disease and allows to calculate for each individual patient risk­
benefit balance of surgery. Furthermore a prediction of the postoperative risk is fundamental to the planning of
measures to be taken in the post­operative period.
Methods: We collected data of patients operated for major pancreatic surgery during 2013­2014 from clinical
database of Verona University Hospital and Peschiera Pederzoli Clinic. Outcome data included incidence of
postoperative complications, hospital mortality, ICU admission and days of hospitalization. The level of risk was
determined by considering BMI, smoking habitus, presence of jaundice, pre­operative drainage of jaundice, recent
chemiotherapy, anamnesis positive for cardiovascular, pulmonary, renal, neurological disease and an American
Society of Anesthesiologist (ASA) score ≥3.
Results: A total of 639 (321 women and 318 man, mean±sd age 61.41±12.52 years) were operated of major
pancreatic surgery during 2013 and 2014 in the two hospitals. 18.8% of these patients were patients with ASA
status ≥3. ICU admission was 8.9%. Total of post­operative complications were 57.6%, mortality was 2.2%. We
observed abdominal complications in 43.5% and general complications (cardiac, respiratory, infectious and renal) in
34.1% of total patients. Mean surgical duration was 332.81±106.05 minutes, and mean duration of hospitalizations
was 15.94±17.83 days. In ASA≥3 patients the percentage of peri­operative general complications was higher than in
low ASA group, and also the in­hospital mortality was higher in high ASA group (5.8%vs1.3%, p. <0.01, chi­square
test). Also intraoperative blood losses, the need of ICU admission (26.7%vs4.8%, p.<0.01) and the mean length of
in­hospital stay (20.90±25vs14.81±15.55 days, p.<0.01) were higher in high ASA group as well. The percentage of
complication with DGE, biliary­fistula (7.1%vs 3.8%, p. <0.01), mean duration of surgery, ischemic cardiovascular
complications were higher in group with pre­operative jaundice. ASA status was not related to strictly surgical­
related abdominal complications. No difference was found in group undergoing or not pre­operative chemiotherapy.
Conclusion: We found good correlation between ASA status and development of peri­operative general
complications (cardiovascular, renal and infectious). Also intraoperative blood losses and length of hospitalization
was well related to ASA status. Presence of pre­operative jaundice is strongly related to development of post­
operative abdominal complications like enteric or biliary fistula. Overall mortality was higher in high ASA status
group. Our data enforce the utility of the development of an easily applied scoring system with convincing accuracy
for identifying high­risk patients, based on preoperatively assessable characteristics, like ASA status, should be
very useful in choosing the right therapeutic strategy, expecially for low­malignant risk lesions. P150 SURGICAL MANAGEMENT OF COMPLICATED PANCREATIC PSEUDOCYSTS FOLLOWING ACUTE
PANCREATITIS Stephen W Behrman, MD, Katy M Marino, MD, Leah E Hendrick, BS; University of Tennessee
Health Science Center, Memphis, US
Endoscopic drainage of pancreatic pseudocysts (PP) is considered first line management but may not be effective
or safe in certain circumstances such as locations in areas outside the lesser sac, infection or when portal venous
(PV) occlusion is present causing perigastric varices.
Methods: Patients having primary internal drainage of PP (defined according to the revised Atlanta classification)
following acute pancreatitis from 2004­14. Management and outcome were assessed relative to location, presence
of infection and/or PV occlusion. Anatomic areas of pseudocyst involvement outside the lesser sac were
categorized by preoperative computed tomography (CT) and included the right and left paracolic gutters, base of
mesentery and the subhepatic space. Infection was defined as those previously stable PP that developed signs of
sepsis preoperatively and had culture positive fluid obtained at the time of surgery. Splenic and/or portal vein
occlusion with associated perigastric varices was identified when observed on preoperative CT imaging and at the
time of surgical exploration. The need for any post­operative therapeutic intervention, radiologic surveillance or
readmission was recorded. Post­operative morbidity and length of stay (LOS) was noted. Results: Forty­eight patients had internal drainage of PP during the study period including 9 with PV occlusion, 11
with infection and 24 that extended to anatomic regions beyond the lesser sac. No patient required transfusion,
reimaging was performed in 1, median post­operative length of stay was 6 days and there were no readmissions and
no procedure related morbidity in those with PV occlusion. Five infected PP extended beyond the lesser sac. Six
had postoperative imaging, 4 readmission and 2 required adjunct postoperative percutaneous drainage for definitive
management. Overall morbidity was 45% and median post­operative LOS was 10 days (range 5­32). Gram positive
and fungal organisms predominated. All but 2 with PP beyond the lesser sac had Roux­en­y cystjejunostomy to
maximize dependent drainage with 4 requiring 2 separate anastomoses. Eight and 4 required reimaging and
readmission respectively. Five patients required intervention beyond the index procedure for definitive management
of their initial pseudocyst: 3 percutaneous drainage, 1 endoscopic drainage and 1 both percutaneous and repeat
operative drainage. Median LOS was 7 days (range 5­75) and 29% suffered at least 1 post­operative complication. Conclusions: 1) Open PP drainage in the face of PV occlusion confers a low risk of bleeding and a minimal need
for reimaging or readmission; 2) Internal drainage of infected PP is a viable option to external drainage. 3) PP that
extend beyond the lesser sac can most often be managed successfully by Roux­en­y drainage but may require
additional intervention for definitive PP resolution
P151 THE LAPAROSCOPIC APPROACH TO DISTAL PANCREATECTOMY FOR DUCTAL
ADENOCARCINOMA RESULTS IN SHORTER LENGTHS OF STAY WITHOUT COMPROMISING
ONCOLOGIC OUTCOMES Susan M Sharpe, MD1, Mark S Talamonti, MD2, Edward Wang, PhD2, David J
Bentrem, MD3, Kevin K Roggin, MD1, Richard A Prinz, MD2, Robert D Marsh, MD2, Susan J Stocker, CCRP2,
David J Winchester, MD2, Marshall S Baker, MD, MBA2; 1University of Chicago Pritzker School of Medicine,
2NorthShore University HealthSystem, 3Northwestern University Feinberg School of Medicine, Chicago, US
Introduction/Background: The oncologic equivalence of laparoscopic distal pancreatectomy (LDP) to open
pancreatectomy (ODP) for ductal adenocarcinoma (DAC) is not established.
Methods: The National Cancer Data Base was used to compare perioperative outcomes following LDP and ODP for
DAC between 2010 and 2011. Chi square and student’s t­tests were used to evaluate differences between the two
approaches. Multivariable logistic regression modeling (MVR) was performed to identify patient, tumor, or facility
factors associated with lymph node count, margin­positive resection, length of stay, 30­day readmission, and
perioperative mortality.
Results: 144 patients underwent LDP; 625 underwent ODP. Compared to ODP, patients undergoing LDP were
older (68±10.1 vs 66±10.5 years, p=0.027), more likely treated in academic centers (70 vs 59%, p=0.01), and had
shorter hospital stays (6.8±4.6 vs 8.9±7.5 days, p<0.001). Other demographic data, lymph node count, 30­day
unplanned readmission, and 30­day mortality were identical between groups. Multivariable regression identified a
lower probability of prolonged length of stay with LDP (OR 0.51, [0.327, 0.785], p=0.0023). LDP was associated
with a decreased risk of having positive margins (OR 0.54, 95% CI [0.302, 0.960], p=0.0358). There was no
association between surgical approach and lymph node count, readmission, or 30­day mortality.
Discussion/Conclusion: LDP for DAC provides shorter post­operative lengths of stay, and rates of 30­day
readmission and 30­day mortality similar to OPD without compromising perioperative oncologic outcomes. P152 THE NEED FOR HEPATOPANCREATOBILIARY SURGEONS: ARE THE COMMUNITY HOSPITALS
UNDERSERVED? Sandeep Anantha Sathyanarayana, MD, Simran Randhawa, MD, Priyanka Annigeri, MD,
Giselle Marshall, Edsa Negussie, Michael Jacobs, MD, Janak Parikh, MD; Providence Hospital Medical Center,
Southfield, US
Introduction: Surgical educators have recently questioned if too many Hepato­Pancreato­Biliary (HPB) surgeons are
being trained. While academic centers may be saturated, many community hospitals may be underserved. Thus, we
sought to determine the need for an HPB surgeon at a tertiary care community hospital.
Methods: All abdominal computed tomography (CT) scans from February 2014 to May 2014 performed at a
community teaching hospital were reviewed and scans with pertinent HPB pathology were isolated.
Results: A total of 389 CT scans having pertinent HPB pathology were identified from 3500 scans, for which an
HPB surgeon consultation would be appropriate (Table 1). Out of the 291 patients with liver specific pathology, 17
patients had hepatic cysts>4cm in size, 76 had a solid mass. Eighty three patients were found with pancreatic
pathology, out of which 17 patients had cystic lesions >1cm, 21 of them had a solid mass, 13 had non­specific main
duct dilatation and 9 had chronic pancreatitis. Fifteen patients had biliary pathology including 11 with biliary ductal
dilatation, one choledochal cyst and one extrahepatic bile duct stricture. For the 3 month study period, a total of 178
patients with significant HPB pathology were identified and the projected volume of patients will be over 700 for a
period of 1 year that would require an HPB surgeon consultation.
Conclusions: The national need for HPB surgeons should be re­evaluated based on the workload at the community
hospital setting and cannot be based on saturation at the academic hospitals.
P153 A MULTICENTER RANDOMIZED CONTROLLED TRIAL TO COMPARING PANCREATIC LEAKS
AFTER TISSUELINKTM VS SEAMGUARD® AFTER DISTAL PANCREATECTOMY (PLATS) Christopher R
Shubert, MD1, Cristina R Ferrone, MD2, Carlos Fernandez­del Castillo, MD2, Daniel S Ubl1, Karla V Ballman,
PhD1, Michael J Ferrara1, Michael L Kendrick, MD1, Michael B Farnell, MD1, KMarie Reid­Lombardo, MD1, Michael
G Sarr, MD1, David M Nagorney, MD1, Rory L Smoot, MD1, Mark J Truty, MD1, Florencia G Que, MD1; 1Mayo
Clinic, 2Massachusetts General Hospital, Rochester, US
Introduction:
Pancreatic leak is one of the most common complications following distal pancreatectomy. The primary objective of
this trial is to compare the effectiveness of TissueLink TM closure of the pancreatic stump after distal
pancreatectomy to that of SEAMGUARD®.
Study Design:
This study was a multicenter, prospective randomized trial of patients undergoing distal pancreatectomy randomized
to either TissueLinkTM or SEAMGUARD® closure of the pancreatic stump at Institutions A and B. A priori power
analysis revealed to reach 80% power and to identify a 10% difference in leak rate between groups, 446 patients
were needed for enrollment.
Results:
Enrollment was closed early due to poor accrual. Overall 67 patients were enrolled, 32 SEAMGUARD® and 35
TissueLinkTM. Enrollment period extended from January 2010 to March 2014. There were no differences in
preoperative patient demographics or risk factors (P>0.05).
Overall clinically significant leak rate was 17.9%; 12.5% for SEAMGUARD® and 22.9% for TissueLinkTM (p=0.27).
There were no differences in clinically significant or major complications; any pancreatic fistula related morbidity,
postoperative length of stay, total length of stay, postop day 3 amylase, drainage occurring longer than 3 weeks,
readmission, percutaneous drainage, duration of operation, estimated blood loss, grade of complication,
pseudoaneurysm formation, ICU stay, or Grade C leak between the two treatment groups.
Post­hoc power analysis revealed, at the current difference in leak rates between the two treatment groups, it would
require 422 patients to reach statistical significance.
Conclusion:
This is the first multi­centered randomized trial to evaluate the incidence of leak rate after distal pancreatectomy
between two separate pancreatic transection methods. The trial closed early due to poor accrual. Given the accrual
rate and the observed difference in leak rates it would have been impractical and neither financially nor technically
feasible to continue the study. Even though there was a difference in leak rates between arms, statistical
significance was not reached. Both treatment options will continue to represent the current standard of care and
should be chosen based on surgeon comfort, experience and pancreas related factors.
P154 ANTERIOR APPROACH TO THE SUPERIOR MESENTERIC ARTERY BY USING NERVE PLEXUS
HANGING MANEUVER FOR BORDERLINE RESECTABLE PANCREATIC HEAD CARCINOMA WITH OR
WITHOUT ABUTMENT OF THE SUPERIOR MESENTERIC ARTERY OR HEPATIC ARTERY Shugo Mizuno,
Shuji Isaji, Masashi Kishiwada, Akihiro Tanemura, Yasuhiro Murata, Hiroyuki Kato, Naohisa Kuriyama, Yoshinori
Azumi, Masanobu Usui, Hiroyuki Sakurai; Department of Hepatobirially Pancreas and Tranplant surgery, Mie
University, Tsu, JP
BACKGROUND: To achieve R0 resection for pancreatic ductal adeno­carcinoma (PDAC) of the pancreatic head,
adequate resection of the retropancreatic nerve plexus around the superior mesenteric artery (SMA) is required.
Since 2010, in an attempt to increase the R0 resection rate, we have undergone the nerve plexus hanging maneuver
using an anterior approach to the SMA for borderline resectable (BR) PDAC of the pancreatic head following
neoadjuvant chemoradiotherapy (NACRT) (J Gastrointest Surg, 2014).
SURGICAL TECHNIQUE: The superior mesenteric veins (SMV) and the SMA are dissected and taped caudal to
the transverse mesocolon. The root of the middle colic artery and the surrounding lymph nodes are dissected and
cut. The nerve plexuses along the SMA was dissected toward the root of SMA meanwhile inferior
pancreaticoduodenal artery was ligated and cut without undergoing Kocher maneuver. A tape for guidance was
passed cranially through the space ventral to the SMA behind the pancreatic parenchyma including the splenic vein
(SV) toward the root of splenic artery, followed by sagital resection of the pancreas parenchyma with the SV.
Another tape was passed behind the nerve plexus lateral to the hepatic artery (HA) and the SMA ventral to the
inferior vena cava and the nerve plexus was dissected, resulting in adequate resection of the nerve plexus around
the SMA. RESULTS: The 35 patients with BR PDAC between January 2010 and June 2014 were divided into two groups:
patients with PDAC invading only portal vein (PV) (BR­PV group: n=19) and those with PDAC invading PV and
SMA/HA (BR­A group: n=16). In BR­A group, the abutted arteries were SMA in 10 patients, HA in 5, and celiac
artery in 1. There were no differences between BR­PV and BR­A groups in the mean duration of surgery,
intraoperative blood loss, and duration of hospital stay (593 min, 1685 ml, 34 days vs. 562, 1489, 36). Pathological
findings of the resected specimen revealed that R0 resection rate was 100% in BR­PV group and 81.3% in BR­A
group. There were no significant differences in 2­year survival rates and 2­year disease­free survival rates between
BR­PV and BR­A groups: 83.0 % vs. 61.5 %, and 51.3 % vs. 60.9 %, respectively. The median survival time was
25.9 vs. 24.3 months. CONCLUSIONS: The nerve plexus hanging maneuver for BR PDAC of the pancreatic head with combination of
NACRT is useful for obtaining adequate resection of the retropancreatic nerve plexus around the SMA to enhance
R0 resection rate even in BR­A.
P155 CENTRAL PANCREATECTOMY WITH PANCREATICOGASTROSTOMY FOR THE TREATMENT OF A
SOLID­PSEUDOPAPILLARY NEOPLASM Jacob E Dowden, MD, Ramsay Camp, MD, Eric T Kimchi, MD,
Katherine A Morgan, MD, David B Adams, MD, Kevin F Staveley­O'Carroll, MD, PhD; Medical University of South
Carolina, Charleston, US
Solid­pseudopapillary neoplasms of the pancreas are rare tumors that primarily affect young women (median age of
diagnosis, 32 years.) They are indolent in nature and most patients are cured with surgical resection. Uncommonly,
solid­pseudopapillary neoplasms have aggressive behavior, but long term survival is still anticipated with metastatic
disease. A 21 year old African­American female was referred with complaints of constant lower abdominal pain. An
abdominal CT scan revealed a 5 cm mixed cystic and solid mass in the pancreatic neck and body junction with
abutment of the celiac axis bifurcation, superior mesenteric artery, and the portal confluence. A central
pancreatectomy with radial celiac node dissection was performed with a pancreaticogastrostomy reconstruction. The patient has had no complications at 11 months post­operatively. Central pancreatectomy provides excellent
long­term functional outcomes (2­6% endocrine/exocrine insufficiency), but can have significant morbidity (30­40%
pancreatic fistula and 40­60% overall morbidity rates.) As such, the procedure is well suited for the young, healthy
patient population that is typical affected by solid­pseudopapillary pancreatic neoplasms.
P156 DOMAIN­BASED ASSESSMENT OF THE LEARNING CURVE FOR NEW SURGICAL TECHNOLOGY:
ROBOT­ASSISTED VS. OPEN DISTAL PANCREATECTOMY. Sjors Klompmaker, MD1, Ammara A Watkins,
MD1, Wald J Van Der Vliet, BSc2, Stijn J Thoolen, BSc2, Manuel Castillo­Angeles, MD1, Jennifer F Tseng, MD,
MPH1, Tara S Kent, MD, MPH1, Arthur J Moser, MD, PACS1; 1Beth Israel Deaconess Medical Center/ Harvard
Medical School, 2Maastricht University, Rosmalen, NL
Introduction: The Learning Curve for minimally­invasive surgery monitors operating time, conversion rate, and
incremental cost as surrogate markers of proficiency and outcome. We expanded this concept to include four
aggregate domains for new technology assessment (Table 1) based on Institute of Medicine principles to evaluate
overall risk/benefit. The initial Learning Curve for robot­assisted distal pancreatectomy (RADP) was compared to
unmatched consecutive open DP (ODP) at an expert center.
Methods: Unmatched comparison between 29 RADP and 169 consecutive ODP performed between 2006­2012 prior
to implementation of RADP. Cumulative treatment burden at 90 days was assessed. Propensity scoring controlled
for selection bias.
Results: No differences in age, gender, race, Charlson Comorbidity Index, suspected pathology, tumor location, or
size were observed between the RADP and ODP cohorts. Within the efficiency domain, RADP patients had a
reduced 90 day total hospital stay (6 vs. 7 days), but longer mean operative time compared to ODP patients. Cumulative morbidity and oncological efficacy for malignancy was similar to ODP within the limits of sample size.
Safety, as measured by blood loss and laparotomy rate (3.3%), was improved following RADP compared to ODP.
Propensity­scored sensitivity analysis did not alter these results.
Conclusion: Domain­based evaluation of the initial RADP learning curve was comparable to the established phase
of ODP in consecutive patients at an expert center. Operating time and associated costs should be re­evaluated in
the context of reduced total hospital stay and increased patient eligibility for the minimally­invasive approach.
Prospective validation of these metrics is required.
Safety
Table 1. Domain­Based Assessment of RADP Compared to ODP
ODP (n=169) RADP (n=29) P Value
Estimated blood loss (ml), median (IQR)
250 (150­500) 50 (35­100)
<.001
Unplanned ICU admissions within 90 d, No. (%)
6 (4)
2 (8)
.34
Mortality within 90 d, No. (%)
2 (1)
­
>.99
Reoperation within 90 d, No. (%)
6 (4)
1 (4)
>.99
Morbidity
90 d Cumulative Complication Index, median (IQR) 20.9 (0­30.8) 18.2 (6.5­34.4) .62
ISGPF Grade B/C fistula, No. (%)
24 (14)
7 (24)
.17
Operative Time (min), median (IQR)
235 (186­303) 332 (279­386) <.001
90 d Readmission, No. (%)
51 (34)
6 (23)
.29
90 d Total hospital stay (d), median (IQR)
7 (6­11)
6 (5­7)
.002
17/17 (100)
.03
25 (17­29)
.19
Efficiency
Oncological Efficacy R0 Resection rate (pre­)malignant disease, No. (%) 41/54 (76)
Lymph node harvest in PDA (n), median (IQR)
16 (10­25)
P157 FIRST 100 TOTAL LAPAROSCOPIC PANCREATODUODENECTOMY. Igor E Khatkov, MD, PhD, prof,
Viktor V Tsvirkun, MD, Prof, Roman E Izrailov, MD, PhD, prof, Pavel S Tyutyunnik, MD, Artur A Khisamov, MD,
Aleksey A Andrianov, MD; Moscow Clinical Scientific Center, Moscow, RU
Background:
Nowadays progress in pancreatic surgery is related with successful applying of minimally invasive technology for
periampullare area’s cancer patients. Total laparoscopic pancreaticoduodenectomy (LPD) is still technically
challenging. As a result just several centers in the world have experience more than 30 – 40 procedures. Aim: to estimate our results of performing LPD.
Methods:
Single surgical team performed all procedures. 100 patients underwent LPD. There were 58 females and 42 males.
Mean age was 61 (range, 45­82) years. In a retrospective study we analyzed the main outcome measures: blood
loss, operative time, length of hospital stay, number of lymph nodes, TNM stage, postoperative morbidity and
mortality. A risk­adjusted Cumulative Sum (CUSUM)model and spline regression applied to operative time (OT)were
used for evaluating the learning curve.
Results:
Having adjusted for case­mix the CUSUM analysis demonstrated a learning curve of 48 cases. Mean operative time
declined with operative experience (p=0.001) and plateaued after 48 cases at 345 min.
Median operative time was 395 min (range, 255 – 705 min) and median blood loss was 225 ml (range, 10 – 2100 ml).
Diagnosis: benign and malignant 15% (n=15) and 85% (n=85) respectively.
Mean number of lymph nodes – 19. TNM: I – 30%; II – 46,1%; III – 17,3%.
The complication rate (Clavien­Dindo Classification) 43%: I or II – 11%; IIIA – 20%; IIIB – 5%; IV – 0%; V – 7%.
Conclusion:
Laparoscopic approach permits to perform pancreaticoduodenectomy as open procedure. Time of procedure is
decreasing with growing of experience.
P158 AGED­DEPEND VULNERABILITY TO EXPERIMENTAL ACUTE PANCREATITIS IS ASSOCIATED
WITH PREVIOUS LIVER MITOCHONDRIAL DAMAGE Ana Maria M Coelho, PhD, Sandra N Sampietre, Marcel
C Machado, MD, PhD, Jose Eduardo M Cunha, MD, PhD, Eleazar Chaib, MD, PhD, Luiz C D'Albuquerque, MD,
PhD; Department of Gastroenterology (LIM/37), University of Sao Paulo, Sao Paulo, Brazil, Sao Paulo Brazil, US
Introduction/Background: It has been widely accepted that the functional impairment of mitochondria is central to the
multifactorial process of ageing. Acute pancreatitis (AP) in elderly patients in spite of similar occurrence of local
complications is followed by a substantial increase in multiple organ failure, including liver failure. We have
previously demonstrated a disruption of liver mitochondrial function in rats with AP. However, studies of the effects
of ageing on liver mitochondrial function after AP induction have not been previously reported. The aim of the present
study was to evaluate the effect of ageing on liver mitochondrial function after AP induction.
Methods: Wistar rats were divided into two groups: Young (3 months old rats, n= 20) and Aged (18 months old rats,
n= 20). Both groups were subdivided into two experimental groups: (1) Sham group: rats submitted to the operative
procedure without induction of AP and (2) AP group. AP was induced by intraductal 2.5% taurocholate injection. Two
hours after AP or sham­operation, blood samples were collected for determinations of amylase, AST and ALT. Liver
tissue was evaluated for mitochondrial function and malondialdehyde (MDA) content. Mitochondrial oxidation and
phosphorylation were measured polarographically by determining oxygen consumption.
Results: A significant increase in serum amylase, AST, ALT was observed in the Aged group compared to the
Young group (p<0.05). Two hours after AP a transient liver mitochondrial dysfunction occurred in young animals,
mainly due to uncoupling of oxidative phosphorylation, and that was partially recovered. Liver mitochondrial
dysfunction did not occur in the Sham group of young animals. However, in aged animals two hours after AP there
was a liver mitochondrial dysfunction that was also noted in sham aged animals, suggesting a previous degenerative
process similar to that found in cellular ischemia. Likewise, it was observed an increase of MDA content in young
animals two hours after AP in comparison to the sham group. The aged animals showed an increase of MDA
content both in the sham group and in the AP group.
Conclusion: This study demonstrates that liver mitochondrial function is transiently compromised in young animals
submitted to AP. However, in aged animals unexpectedly the pre­existing severe mitochondrial dysfunction
remained unchanged after induction of AP associated with a sustained oxidative stress. These findings may have
significant therapeutic implications in the clinical setting.
P159 ERYTHROCYTE AGEING AND GLYCATED MARKERS OF DIABETES MELLITUS IN CHRONIC
PANCREATITIS. Manuel Beltran del Rio, PhD, George Georgiev, MSc, Leo Amodu, MD, Horacio Rilo, MD;
Feinstein Institute For Medical Research, Manhasset, US
We present a model of erythrocyte ageing and hemoglobin glycation based on previously published in vivo data from
our own and external centers. This model was validated through direct comparison with known empirical values.
The model was used to assess possible discrepancies in glycated haemoglobin (HbA1C) levels arising in Chronic
Panreatitis due to differential erythropoietic and plasma­glucose patterns. We also calculate the age distribution of
red­blood cells, and with it we estimate a realistic reaction time for HbA1C percentages to reflect a change when
blood sugar levels are altered in different degrees.
P160 BIODEGRADABLE BILIARY STENTS MAY HAVE A BENEFICIAL EFFECT OVER COVERED METAL
STENTS ON EXPRESSION OF PROTEINS ASSOCIATED WITH TISSUE HEALING IN BENIGN BILIARY
STRICTURES Antti Siiki, MD3, Ralf Jesenofsky, MD1, Matthias Löhr, Md, PhD2, Isto Nordback, Md, PhD3, Juhani
Sand, MD, PhD3, Johanna Laukkarinen, MD, PhD3; 3Tampere University Hospital, Finland, 1University of
Heidelberg, Germany, 2Karolinska University Hospital, Sweden, Tampere, FI
Background: Benign biliary strictures (BBS), commonly caused by chronic pancreatitis are primarily treated
endoscopically, increasingly with covered self­expandable metal stents (CSEMS). Biodegradable biliary stents
(BDBS) have shown promising results in animal models and in percutaneous use in humans. Tissue response in
BBS to different types of stents is largely unknown. The aim was to assess the expression of proteins related to
tissue healing in BBS compared to the intact bile duct (BD) and to study the protein expression after therapy with
CSEMS or BDBS.
Methods: Swine with experimental ischemic BBS model were endoscopically treated either with poly­lactide BDBS
or CSEMS. Tissue samples were harvested from swine with intact BD (n=5), untreated BBS (n=5) and after six
months of therapy with BDBS (n=4) or CSEMS (n=5). Two­dimensional electrophoresis with protein identification
was performed to evaluate protein expression patterns at these stages.
Results: Compared to the intact BD, in BBS the expression of Galectin­2 and Annexin­A4 decreased. BDBS
treatment normalized Galectin­2 level, but with CSEMS therapy it remained low. Annexin­A4 expression remained
low after both treatments. Transgelin expression, which was low in intact BD and in the BBS, remained low after
BDBS treatment but increased after CSEMS therapy.
Conclusion: The expression of proteins related to tissue healing is different in BBS compared to intact bile duct.
Furthermore, the protein expression pattern is different after treatment with BDBS and CSEMS. Response to BDBS
resembles intact bile duct perhaps suggesting a beneficial effect of BDBS over CSEMS in BBS. These findings
warrant further studies for their potential therapeutic implications. P161 CHRONIC PANCREATITIS AND ASSOCIATED FACTORS: A SINGLE CENTER CASE CONTROL
STUDY Milena Di Leo, MD1, Raffaella A Zuppardo, MD, PhD1, Alberto Mariani, MD1, Margherita Bianco2, Oliva B
Morrow1, Teresa M Rogger1, Gioacchino Leandro, MD2, Pier Alberto Testoni, MD1, Giulia Martina Cavestro, MD,
PhD1; 1Gastroenterology Unit, IRCCS San Raffaele Scientific Institute, Vita­Salute San Raffaele University,
2Gastroenterology Unit 1, Gastroenterological Hospital ‘S. De Bellis’ IRCCS, Castellana Grotte, Italy, Milano, IT
BACKGROUND:Effect of smoking intensity on CP development is unknown.AIMS:To investigate the contribution of
smoking, alcohol, pancreas divisum and CFTR mutations in CP risk.
METHODS:CP patients undergone to S­MRCP were consecutively enrolled. This group was compared with
consecutively subjects examined by S­MRCP for irritable bowel syndrome(IBS) with upper abdominal pain, without
history, signs or S­MRCP findings suggesting pancreatic disorder.
RESULTS:From 2010 to 2014 we enrolled 145CP and 103IBS patients. In univariate analysis, statistical
significance differences in gender, mean age, smoking and alcohol habitus(duration and consumption) were
found(Table 1). We estimated the cigarettes and grams of alcohol cutoff for CP risk using ROC curve analysis. For
cigarette’s intake the AUC was 0,756(95%CI 0,695–0,817;p<0.001). The cut off point for cigarette for day was 5,5.
The AUC for alcohol intake was 0,766 (95%CI 0,708–0,824;p<0,001), and the cut off point for alcohol was
13,5grams/day. In CP group we confirmed a significant statistical prevalence in CFTR gene. In the multivariate
analysis, the independent risk factors for chronic pancreatitis were: pancreas divisum(OR 2,21 95%CI 1.01­4,84);
male gender(OR 2,89 95%CI 1.52­5,43); previous or current smoking habitus(OR 2,31 95%CI 1.17­4,56); previous or
current alcohol habitus(OR 3,15 95%CI 1.54­6,41).
CONCLUSIONS:We confirmed smoking and alcohol are co­factors that increase the risk of pancreatitis. It is well
known how high doses of smoking and alcohol intake are clearly associated with CP. This study evidences low
amount of alcohol intake and smoking habits remains two of the most important risk factors in the pathogenesis of
the disease.
IBS group(n=103)
CP group(n=145)
p
Male gender
38(36,89%)
106(73,10 %)
<0,001
Mean age±SD(years)
47,19±14,37
52,36±15,11
0,006
Smoking habits
37(35,92%)
107(73,79%)
<0,001
Mean number cigarettes±SD(cig/day)a
12,27±10,54
21,87±11,15
<0,001
Mean duration of smoking habits±SD(cig/day)a
19,97±11,80
30,00±12,60
<0,001
Year from end of smokingb
7,00±4,87
10,88±8,25
0,198
Alcohol habits
22(21,36%)
95(65,52%)
<0,001
Mean massimum alcohol intake±SD (g/day)c
26,50±25,02
86,24±71,39
<0,001
Duration of alcohol intakec
25,27±13,18
30,46±14,79
0,23
Duration of massimum alcohol intakec
26,0±14,28
30,63±15,62
0,24
Mean actual alcohol intake±SD(g/day)d
21,50±19,73
67,79±60.40
<0.001
Year from end of alcohol intakee
4,26±3,40
7,85±8,43
0,438
Pancreas Divisum
16(15,53%)
34(23,45%)
0,126
aCurrent or previous smoker
bPrevious smoker
c Current or previous drinker
dCurrent drinker
ePrevious drinker
P162 CYANOACRYLATE GLUE INJECTED ENDOSCOPICALLY TO CONTROL BLEEDING FROM A
PANCREATITIS INDUCED SPLENIC ARTERY PSEUDOANEURYM Fiona Ross1, Nigel Jamieson1, Sivanathan
Chandramohan2, Colin J McKay1; 1Department of Pancreatic Surgery, Glasgow Royal Infirmary, 2Department of
Radiology, Glasgow Royal Infirmary, Glasgow, GB
Introduction
Pseudoaneurysm is a leak of blood from an artery, contained by surrounding tissue, with ongoing communication
between the artery and the resultant haematoma. Peripancreatic pseudoaneurysms are located around the pancreas.
They arise secondary to pancreatitis and the resultant leak of enzymatic, proteolytic rich fluid causing autodigestion
thereby weakening vessel walls. Pseudoaneurysm formation following pancreatitis is a rare complication but with
potentially life threatening consequences of rupture and haemorrhage. Treatment options traditionally involve surgical
or endovascular repair. More recently use of image guided percutaneous injection of thrombin has had a role.
We report a case of a splenic artery pseudoaneurysm secondary to complex severe acute pancreatitis that was
managed by the injection of cyanoacrylate glue via endoscopic ultrasound (EUS) guidance and discuss the role of
the novel technique.
Case
A 49 year old man with severe acute gallstone pancreatitis presented initially with abdominal pain complicated by
organ failure. Initial management included percutaneous necrosectomies followed by EUS guided cystgastrostomy
and stent insertion. Unfortunately his management was complicated by significant gastric and retroperitoneal
haemorrhage. Angiography and laparotomy were required to control haemorrhage. Unfortunately angiography resulted
in splenic artery dissection with pseudoaneurysm formation arising from the coeliac axis.
Three months following discharge from his index presentation the patient represented with haematemesis and
melena. CT angiography revealed resolution of the previous coeliac axis pseudoaneurysm however a new splenic
artery pseudoaneurysm had developed, with unfortunately no endovascular access. Therefore EUS guided thrombin
injection was used to control bleeding. Post procedure imaging showed that both the splenic artery pseudoaneurysm
and the splenic artery had been thrombosed. However two months later follow up CT revealed that the splenic artery
pseudoaneurysm had recurred.
Following on a further two months the patient represented with a further presentation of upper GI bleed. On this
occasion the bleeding was again thought to be secondary to the pseudoaneurysm. 2ml of cyanoacrylate glue was
injected into the pseudoaneurysm under EUS guidance. This successfully controlled the bleeding and left the splenic
artery patent. CT angiography performed four days later showed that the pseudoaneurysm is completely thrombosed
with no distal embolisation into splenic vessels.
Discussion
Both surgical and angiographic repair of pseudoaneurysm are associated with a high risk of morbidity and mortality.
Endoscopic management is another option. In comparison to both surgery and angiographic repairs, it is associated
with lower risk of complications and mortality. It is particularly useful, as in our case, when there is no suitable
endovascular access.
Cyanoacrylate glue is synthetic glue that rapidly solidifies when in contact with blood. It is used endoscopically for
control of variceal haemorrhage due to its rapid solidification properties and therefore rapid control of bleeding. Use in
endoscopic pseudoaneurysm management would appear to be another potential use for synthetic glues.
It would appear that very little literature exists around endoscopic management of pseudoaneurysm with synthetic
glue. Given the success of this case further investigation would be necessary to determine its effectiveness, risk
and long term results.
P163 SYSTEMIC LEVELS OF SOLUBLE UROKINASE­TYPE PLASMINOGEN ACTIVATOR RECEPTOR
(SUPAR) PREDICT THE SEVERITY OF ACUTE ALCOHOL PANCREATITIS Anssi Nikkola, BM1, Janne
Aittoniemi, PhD, MD2, Reetta Huttunen, PhD, MD3, Juhani Sand, PhD, MD1, Johanna Laukkarinen, PhD, MD1;
1Dept of Gastroenterology and Alimentary Tract Surgery, Tampere University Hospital, Tampere, Finland, 2Fimlab
Laboratories, Tampere, Finland, 3Dept of Internal Medicine, Tampere University Hospital, Tampere, Finland,
Tampere, FI
Background. Urokinase­type plasminogen activator receptor (uPAR) is a membrane protein expressed in various
immune cells, endothelial cells and cancer cells. Systemic levels of its soluble form (suPAR) are associated with
activation of the immune system. SuPAR is a documented prognostic marker of sepsis, SIRS and various cancers.
Our aim was to study the activation and prognostic value of plasma suPAR (P­suPAR) in patients with first acute
alcohol pancreatitis (AAP).
Methods. P­suPAR concentrations were measured during hospitalization in 104 patients with first AAP and in 77
patients 6­24 months after discharge using a commercial ELISA from ViroGates A∕S (Birkerød, Denmark). The
severity of AAP was classified according to the revised Atlanta criteria.
Results. In 104 patients with first AAP, pancreatitis was mild in 69 (66%), moderately severe in 29 (28%) and
severe in 6 (6%). For further analysis, the moderately severe and severe AAP were combined (non­mild AAP; n=35,
34%). P–suPAR levels were significantly higher in non­mild AAP (median 6.2 ng/mL; range 1.9­39) compared to mild
AAP (4.2 ng/mL; 1.6­16.4, p<0.001). P­suPAR levels were low as expected after the recovery from AAP (3.1 ng/mL;
1.6­6.9). There was a good predicting value of P­suPAR for AAP severity when the analysis was made within 1­4
days after admission (n=68): area under the receiver­operating curve was 0.81 (95% CI, 0.70­0.92, p<0.001). At a
cut­off level of 5.0 ng/mL, sensitivity and specificity to predict a non­mild AAP were 79% and 78% respectively. At
admission suPAR was found to be a better prognostic marker for the severity of the disease than C­reactive protein,
hematocrit or creatinine.
Conclusion. P­suPAR concentrations are elevated in AAP. Plasma suPAR over 5.0 ng/mL predicts the
development of a moderately severe or severe AAP. These results suggest that P­suPAR may have a potential to
serve as a novel prognostic marker for AAP severity at admission.
P164 THE EFFECT OF PERIPANCREATIC VASCULAR DISORDERS ON SURGERY FOR CHRONIC
PANCREATITIS Moritz F Pross1, T Keck1, F Makowiec2, D Bausch1, U F Wellner1, U Hopt2, K C Honselmann1,
D Tittelbach­Helmrich1; 1UKSH Luebeck, 2University clinic Freiburg, Luebeck, DE
Background: The aim was to investigate the effect of peripancreatic vascular disorders on the outcome on surgical
therapy of chronic pancreatitis.
Methods: We performed a retrospective analysis of 324 patients with chronic pancreatitis who underwent surgery
between 1999 and 2009. The outcome of 108 patients with vascular disorders was compared to 216 patients without
vascular findings for peri­ and postoperative parameters. Patients were divided into groups related to following
vascular disorders: splenic vein thrombosis (SVT), vascular compression (VCO), left­sided portal hypertension
(LPH), portal vein thrombosis (PVT), generalized portal hypertension (GPH) and arterial stenosis (AST).
Results: 34% of the 324 patients showed vascular disorders. Angiography and MRI proved to have the highest
sensitivity in detecting vascular involvement. All analysed vascular findings lead to a significant higher perioperative
rate of blood transfusion with a significant higher amount of transfused blood volume and prolonged operation time.
No increase of postoperative pancreatic fistula (POPF), post pancreatectomy hemorrhage (PPH) or delayed gastric
emptying (DGE) was found in patients with vascular disorders. Furthermore no effect was detected on hospital stay,
re­operation rate or overall mortality. Beger’s procedure was identified as the surgical approach with the lowest
perioperative risk profile in patients with vascular disorders.
Conclusion: Surgical therapy in patients suffering from chronic pancreatitis with vascular involvement is save but
bears higher risk for perioperative blood loss and increased operation time. A broad preoperative stratification of
vascular involvement via MRI seems crucial to identify patients at risk. Where possible Beger’s procedure seems to
be a save option for patients suffering from chronic pancreatitis complicated by vascular alterations.
P165 U­TUBE DRAINAGE FOR NECROTIZING PANCREATITIS: RESULTS OF A NOVEL INTERVENTION
AT A HIGH VOLUME PANCREATIC DISEASE CENTER Cc Stahl, Js Moulton, D Vu, Rl Ristagno, Jj Sussman,
Sa Shah, Sa Ahmad, De Abbott; University of Cincinnati, Cincinnati, US
Introduction/Background: U­tube drainage (UTD) is a described but not yet­popularized intervention for necrotizing
pancreatitis (NP). Theoretic benefits include more effective flushing (larger bore catheter, multiple large side holes),
greater interface with large fluid collections (more rapid resolution of retroperitoneal necrosis), less risk of
dislodgement/fewer catheter exchanges and creation of a large­diameter fistula tract for potential
fistulojejunostomy. This approach, however, has not been reported in any large series; here we report the largest
clinical experience with UTD.
Methods: From 2011­2014, a period in which UTD became the primary intervention for NP our institution, all
patients requiring intervention were identified from departmental records. Clinical variables including mortality,
complications, length of stay (LOS), number of CT scans, number of interventional radiology (IR) procedures,
number of surgical interventions, and long­term outcomes populated our dataset.
Results: 22 patients had U­tube placement as primary intervention for NP, with a median follow up of 10.2 months.
The cohort was primarily male (n=18, 81.8%), and NP was most commonly due to gallstone disease (n=9, 40.9%),
followed by idiopathic disease (n=5, 22.7%) and alcohol abuse (n=4, 18.2%). During the course of UTD and
definitive surgical therapy (when required), patients had median hospital stays of 31 days, 6 IR procedures, and 6
CT scans. Disease­specific mortality was 9.1%. Surgical intervention was not necessary in nearly ½ of
patients (n=9; 40.9%). In the other 13 patients, 4 patients underwent distal pancreatectomy/splenectomy, 8 had a
fistulojejunostomy performed, and 1 underwent both procedures. One patient developed a recurrent leak following
their distal pancreatectomy/splenectomy; the other 12 patients had permanent resolution of their disrupted duct.
Discussion/Conclusion: UTD for NP patients requiring intervention is associated with low morbidity, hospital
resource utilization and frequent non­operative resolution, all comparing favorably to historical NP
interventions. With skilled interventional radiologists (IR) and coordination between IR and surgeons, this technique
is a valuable means of minimizing morbidity for this difficult pathology. Prospective evaluation of UTD versus other
percutaneous approaches for NP is required.