1. No category

Phase 3 SYNERGY Trial: Docetaxel +/- Custirsen and Overall Survival
in Patients with Metastatic Castration-Resistant Prostate Cancer and Poor Prognosis
5009
Kim N.
1
Chi,
Celestia
2
Higano,
Brent
3
Blumenstein,
James
4
Reeves,
Susan
5
Feyerabend,
Gwenaelle
6
Gravis,
Jean-Marc
7
Ferrero,
Cindy
8
Jacobs,
and Johann de
9
Bono
1BC
Cancer Agency, Vancouver, BC, Canada; 2University of Washington/Fred Hutchinson Cancer Research Center, Seattle, WA; 3Trial Architecture Consulting, Washington, DC; 4Florida Cancer Specialists, Fort Myers, FL; 5Urologic University Clinic, Reutlingen, Germany; 6Medical Oncology,
Institut Paoli Calmette, Hôpital de Jour, Marseille, France; 7Department d'Oncologie Medicale, Centre Antoine Lacassagne, Nice, France; 8OncoGenex Pharmaceuticals, Inc., Bothell, WA; 9The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, Sutton, United Kingdom
RESULTS
SAFETY
Survival by Statistical Model Score and Arm (N=984)
Incidence of ≥Grade 3 Adverse Events in ≥3% of Patients
in Good and Poor Prognosis Subset
GOALS OF THIS ANALYSIS
CLUSTERIN
 Clusterin (CLU) is a cytoprotective protein that functions similarly to ATP-independent heat
shock proteins.
 CLU is more highly expressed in aggressive cancers and upregulated in response to
treatment such as chemotherapy, hormone ablation, and radiation therapy.
 CLU is expressed in many malignancies including castration resistant prostate cancer, and
is implicated in treatment resistance. Thus, inhibition of CLU may be important in disabling
cancer cells’ adaptive defenses, especially in more aggressive or treatment-resistant
cancers.
Secretory
CLU
To explore whether inhibition of CLU production is more relevant
in patients with poor prognosis
 Identify well-established baseline prognostic factors relevant to SYNERGY from Control Arm
patients (no influence of custirsen)
 Develop Control Arm multi-variable statistical model relating putative prognostic variables
and survival outcome
 Score patients from Both Arms according to Control Arm statistical model
 Patients dichotomized into good and poor prognosis subsets
 Develop simple score to be used to identify poor prognosis patients for future study
 Null hypothesis: Custirsen effect (between-arm effect) is homogeneous across prognosis
scoring classifications
Good Prognosis Score Patients
Good prognosis:
Custirsen treatment
Control arm
Survival Probability (%)
BACKGROUND
Custirsen +
Docetaxel
(n = 240)
Docetaxel
(n = 252)
32.6
30.4
Median OS,
mo
HR (95% CI)
Poor prognosis:
Custirsen treatment
Control arm
Adverse Event
Any
Neutropenia
Anemia
Fatigue
Febrile neutropenia
Leukopenia
Asthenia
Diarrhea
Back pain
Pulmonary embolus
Pneumonia
1.02 (0.76 to 1.37), P = 0.901
Poor Prognosis Score Patients
Custirsen +
Docetaxel
(n = 250)
Docetaxel
(n = 242)
17.0
14.0
Median OS,
mo
HR (95% CI)
0.73 (0.59 to 0.90), P = 0.004
Time (Months)
METHODS
CUSTIRSEN
SYNERGY: STUDY DESIGN AND RESULTS
Randomized, open-label, multinational phase 3 study conducted in 140 centers in 12 countries
• mCRPC
• Candidates for
1st-line
chemotherapy
• Stratified by
opioid use and
radiographic
progression
N = 1022
R
A
N
D
O
M
I
Z
E
D
1:1
• Docetaxel 75
IV q3w
• Prednisone 10 mg qd
mg/m2
n = 512
• Docetaxel 75
IV q3w
• Prednisone 10 mg qd
• Custirsen 640 mg IV x
3 loading doses then q1w
mg/m2
n = 510
Endpoints
Primary
• OS
Secondary
• Safety
• Proportion of patients alive
and without radiographic
progression on day 140
Exploratory
• Serum clusterin and OS
• PSA trends
• Pain and analgesic use
 Final analysis following a target 509 deaths to assure 90% power for hypothesized HR 0.75,
with one-sided type I error of 0.025 and type II error of 0.1
Survival Probability
Overall Survival for SYNERGY (N=1022)
Median OS, mo*
HR (95% CI)
P value (2-sided)
Custirsen treatment
Control arm
Docetaxel +
Docetaxel
Custirsen
(n = 512)
(n = 510)
23.4
22.2
0.93, 0.78-1.11
0.42
*Survival compared using a stratified log-rank test and included
all randomized patients
STATISTICAL MODEL SCORE
 Prognostic statistical modeling based on Control Arm only data
• Survival as a function of multiple baseline covariates using proportional hazards
regression
• Up to third degree interactions available for inclusion with term selection based on
hierarchical step-down
• Continuous variables also evaluated as overall median-dichotomized
 Selected Control Arm statistical model included following dichotomous baseline covariates
(and some interactions):
• Karnofsky Performance Status ≤ 80
• Liver metastasis present
• Opioid use for prostate cancer pain
• Hemoglobin < 124 g/L
• LDH ≥ 331 IU/L
• PSA ≥ 59 ng/mL
• Alkaline phosphatase ≥ 92 U/L
 Coefficients used to compute “statistical model score” of patients in both arms
 Statistical model score evaluated by dichotomizing at overall median of both arms,
classifying patients as “poor” versus “good” prognosis
Good prognosis:
Custirsen treatment
Control arm
Docetaxel
(n = 295)
29.4
27.7
Median OS,
mo
HR (95% CI)
Custirsen +
Docetaxel
(n = 208)
Docetaxel
(n = 201)
16.0
13.7
Median OS,
mo
HR (95% CI)
Time (Months)
0.73 (0.58-0.92), P = 0.008
*Negative Prognostic Factors: Karnofsky Performance Status ≤ 80%;
presence of liver metastasis; hemoglobin < 120 g/L;
LDH ≥ 360 IU/L; PSA ≥ 150 ng/mL
Good Prognosis
Custirsen +
Docetaxel
(n = 240)
Any
Time to treatment (median, mo)
Abiraterone
Time to treatment (median, mo)
Enzalutamide
Time to treatment (median, mo)
Cabazitaxel
Time to treatment (median, mo)
Poor Prognosis
Docetaxel
(n = 252)
Custirsen +
Docetaxel
(n = 250)
Docetaxel
(n = 242)
79%
81%
74%
74%
10.7
10.6
8.4
8.0
55%
60%
55%
53%
13.4
12.8
9.6
9.4
28%
26%
16%
17%
22.1
22.3
15.4
12.8
20%
25%
16%
14%
25.7
24.9
14.9
13.3
Time to PSA Progression For Poor Prognosis Subgroup
By Statistical Model Score and PCWG Criteria
PCWG1 Criteria
Custirsen + Docetaxel (n=247)
%
Docetaxel (n=235)
%
80
33
17
12
10
9
8
6
4
4
4
66
20
9
11
5
6
3
3
5
4
3
CONCLUSIONS
Use Of Post-Protocol Therapies
 In this exploratory analysis, mCRPC patients with poor prognosis treated with
custirsen/docetaxel/prednisone had improved survival as compared to
docetaxel/prednisone when used as first-line therapy.
 This observation is consistent with custirsen’s mechanisms of action to alter
treatment resistance; in patients with poor prognostic features and increased
treatment resistance, custirsen may have greater benefit by preventing or
reducing CLU production within tumor cells.
 Use and timing of post-protocol therapy was similar between treatment arms.
 Although custirsen/docetaxel/prednisone is associated with some increased
hematologic and gastrointestinal toxicity, the overall adverse event profile
appeared similar between patients in the good and poor prognosis subgroups.
 Further evaluation of custirsen in patients with poor prognosis disease is
warranted and is on-going in the AFFINITY trial (NCT01578655).
PCWG2 Criteria
ACKNOWLEDGEMENTS
This research was supported by OncoGenex Technologies Inc (Bothell, WA).
Clinicaltrials.gov identifier: NCT01188187.
Note: all are established prognostic factors in prostate cancer 4,5
Custirsen +
Docetaxel (n = 245)
Abbreviations: IV, intravenous; mCRPC, metastatic castration-resistant prostate cancer; mo, month; PCWG, Prostate Cancer
Working Group; PSA, prostate-specific antigen; q1w, weekly; q3w, every 3 weeks; qd, daily; OS, overall survival
Any
Neutropenia
Anemia
Fatigue
Febrile neutropenia
Leukopenia
Asthenia
Diarrhea
Back pain
Pulmonary embolus
Pneumonia
1.00 (0.78 to 1.29), P = 0.979
Median, mo
HR (95% CI)
Custirsen + Docetaxel
(n = 245)
Docetaxel
(n = 234)
7.9
6.4
0.75 (0.60-0.93)
REFERENCES
Poor Prognosis
Poor Prognosis
Time (Months)
Adverse Event
Poor Prognosis
(≥2 of 5 Negative Prognostic Factors*)
Poor prognosis:
Custirsen treatment
Control arm
SIMPLIFIED PROGNOSTIC INDEX
 Findings from statistical model score used to develop simplified 5-criteria prognostic index
 Criteria for prognostic index: (1) matches closely prognostic performance of the statistical
model score, (2) identifies sufficient proportion of poor prognosis patients, and
(3) easy to implement
 Index selected classifies patients as “poor” prognosis based on having 2 or more
of following 5 features:
• Karnofsky Performance Status ≤ 80%
• Presence of liver metastasis
• Hemoglobin < 120 g/L
• LDH ≥ 360 IU/L
• PSA ≥ 150 ng/mL
Custirsen +
Docetaxel
(n = 283)
Poor Prognosis (n=482)
%
73
26
13
11
7
7
5
4
4
4
3
Incidence of ≥Grade 3 Adverse Events in ≥3% of Patients
in Poor Prognosis Subset By Arm
Good Prognosis
(<2 of 5 Negative Prognostic Factors*).
Survival Probability (%)
 Custirsen is a second-generation antisense oligonucleotide (ASO) designed to bind to CLU
mRNA, resulting in inhibition of production of human CLU protein. Preclinical studies
demonstrate enhanced efficacy of chemotherapy when combined with custirsen and
clusterin inhibition, including reversing taxane resistance.
 A phase 1 pre-surgery study determined that clusterin inhibition by custirsen in prostate
cancer tissues was dose-dependent, with a biologically effective dose of 640 mg.1
 A randomized phase 2 study in patients with mCRPC showed that the addition of 640 mg
custirsen to docetaxel plus prednisone (DP) reduced serum clusterin by 26% and prolonged
overall survival (OS) vs. DP alone (23.8 vs. 16.9 mo; Cox regression hazard ratio [HR] 0.50;
95% confidence interval [CI] 0.29-0.87).2
 The SYNERGY study (OGX-011-11) was a randomized study of docetaxel/prednisone +/custirsen and enrolled 1022 patients. Following 509 deaths, the median OS was 23.4
months (m) vs. 22.2 m for the custirsen and control arms, respectively (HR 0.93; P = 0.42).3
Survival by Simplified Prognostic Index and Arm (N=984)
Good Prognosis (n=484)
%
65
25
3
8
11
6
5
5
2
4
3
Median, mo
HR (95% CI)
Docetaxel
(n = 234)
8.2
7.4
0.82 (0.64-1.05)
1Chi
KN, et al. J Natl Cancer Inst 2005;97(17):1287-96.
2Chi KN, et al. J Clin Oncol 2010;28(27):4247-54.
3Chi KN, et al. Ann Oncol 2014;25(suppl 4):iv256.
4Halabi
S, et al. J Natl Canc Inst 2013;105:1729-37.
5Halabi S, et al, J Clin Oncol 2014;32(7):671-677.
ASCO MAY 2015