Quantitative Mutational Assessment of Circulating Tumor DNA
Using Massively Parallel Deep Sequencing in Plasma and Urine
from Advanced Colorectal Cancer Patients
3
0
20
Theoretical
20
0
17
3
18
2
14
6
6
14
2
18
1
SAMPLE QUALITY
o Total of 190 plasma and urine samples archived for 3-5 years were tested.
o 101 of 101 plasma samples (100%) and 79 of 92 urine samples (86%) had
sufficient DNA concentration and were deemed evaluable. The degree of
DNA degradation in these archival urine samples is significantly elevated
compared to typical clinical samples.
19
Actual
(G12A)
• Poisson distribution model: The theoretical model shows the number of positive and
negative samples that would be present in a perfect Poisson distribution of our test
using 20 replicates at each spike-in level. The actual positive/negative hit distribution
follows the theoretical model for a Poisson distribution very closely, demonstrating the
single copy sensitivity of the assay.
Monitoring during treatment
(total n=193 plasma and urine samples)
25
Plasma
20
15
Palliative RT
10
Death
0
6/21/2010
5/2/2010
3/13/2010
1/22/2010
12/3/2009
10/14/2009
8/25/2009
5
Fold Change KRAS Copies/100K GE
(PLASMA)
Urine
Collection Date
Analytical Specificity
100%
1
2.0
1.5
2
Day 1 post
1.0
1
0.5
0
0.0
Collection Date
11/3/2009
Analytical Sensitivity
Single copy detection
LOD = 0.0067% mutant DNA in a background of wildtype DNA
2.5
3
10/29/2009
2 – 480 ng, recommended input 60 ng
Accurate quantitation of mutant alleles with any DNA input
3.0
Day 5 post
10/24/2009
Input DNA
During
10/19/2009
Urine (proprietary collection kit and DNA isolation technology)
Plasma (BD Vacutainer EDTA, Streck Cell-Free DNA BCT tube)
4
10/14/2009
Tissue Source
5
10/9/2009
Performance
Week of Surgery
Day 1 post
Day 5 post
10/4/2009
Performance
Characteristics
Patient 4 (KRASG12D)
Fold Change KRAS Copies/100K GE
(PLASMA)
• Boxplots show 25 and 75% quartiles and bars indicate 1.5x of the interquartile ranges.
* Assay characteristics were developed on a separate dataset using DNA blends from cell lines
Fold Change KRAS Copies/100K GE
(PLASMA)
1/22/2010
12/3/2009
10/14/2009
Patient 12 KRASG13D
2.0
Urine
Fold Change KRAS Copies/100K GE
(PLASMA)
1.2
Plasma
1.0
1.5
0.8
1.0
0.6
Week of
Surgery
Last Sample
6/21/2011
0.5
0.0
8
7
6
5
4
3
2
1
0
0.4
0.2
0.0
Collection Date
Patient 3 KRASG12V
7/6/2009
• Standard curves were developed for each mutation using 288 independent enrichments
reactions/curve with different amounts of spiked DNA input from 0-500 copies. Examples
of master standard curves for G12V/A/R/S are shown below.
350
300
250
200
150
100
50
0
8/25/2009
7/6/2009
Collection Date
Representative cases of 20 patients shown
Standard Curves for Quantitation*
5/2/2010
17
0
Fold Change KRAS Copies/100K GE
(PLASMA)
13
2/26/2011
7
10/18/2011
8
11/18/2010
12
4/14/2011
4
0
8/10/2010
16
1
6/27/2010
2
Death 4/30/2011
4/8/2010
18
Palliative CHT
10/30/2009
0
2
1/22/2010
20
URINE
o KRAS mutation concordant with tissue detected in 11 of 12 evaluable
baseline urine samples (92%).
Plasma
3
8/4/2009
WT M
Urine
10/14/2009
60/20
Patient 10 KRASG13D
7/6/2009
40/20
5/17/2009
o We developed a highly sensitive mutation enrichment assay for the
detection of KRAS codon 12/13 mutations in highly fragmented urinary
and plasma ctDNA.
o To achieve greater sensitivity in fragmented cfDNA, the assay utilizes a
31bp footprint and contains a selective enrichment step for mutated DNA
fragments while suppressing wild-type (WT) sequence amplification with
a blocker.
o Barcoded adaptor primers are added for compatibility with massively
parallel deep sequencing.
o Following sequencing , our proprietary analysis algorithm allows accurate
quantitation of input level of mutant DNA.
20/20
9/29/2009
COPYRIGHT © 2013, TROVAGENE, Inc.
Assay Design
10/20
3/28/2009
o
5/20
9/24/2009
o
2.5/20
Fold Change KRAS Copies/100K GE
(URINE)
o
Samples were collected from 20 patients with Stage IV colorectal cancer
and KRAS positive primary tumor.
All patients were undergoing surgical treatment in combination with
various systemic medical therapies including neoadjuvant
radio/chemotherapy, adjuvant targeted therapy and adjuvant
chemotherapy.
Archived, matched longitudinal plasma and urine samples were collected
and evaluated prior to surgery (baseline) as well as time intervals
throughout the course of the disease.
In a blinded retrospective study, concordance between archived
tissue/plasma/urine samples was studied for baseline and urine and plasma
KRAS was monitored longitudinally.
0/20
4
Week of Surgery
7/20/2009
PLASMA
o KRAS mutation concordant with tumor tissue detected in 19 of 20
evaluable baseline plasma samples (95%). In one patient, a KRAS mutation
was detected consistently across all serial plasma and urine samples but
not in tissue biopsy.
Fold change KRAS copies/100K GE
(URINE)
o
Spiked DNA
(mutant copies
/total #
reactions)
50
Collection Date
WT DNA (ng)
Clinical study
Death
12/14/2009
92%
Serial monitoring during and post surgery
(n=193 plasma and urine samples)
Single Copy Detection*
100
7/20/2009
• 7 mutant DNA copies reliably detected in:
60ng WT DNA ~17,400 genomic copies (LOD = 0.04%)
360ng WT DNA ~104,400 genomic copies (LOD = 0.0067%)
100%
150
3/28/2009
95%
Urine vs. Tissue Biopsy
Concordant
KRAS mutation at baseline
11/12 evaluable
Surgery
7/21/2009
Urine vs. Plasma
Concordant
KRAS mutation at baseline
11/11 evaluable
200
12/18/2008
Plasma vs. Tissue Biopsy
Concordant
KRAS mutation at baseline
19/20 evaluable
Neo-CHT
250
Plasma
5/17/2009
Urine (20 Patients)
Plasma (20 Patients)
Urine
3/28/2009
Limit of Detection* = 0.0067%
18
16
14
12
10
8
6
4
2
0
Patient 9 KRASG12A
2/6/2009
o
Technologies enabling the assessment of circulating tumor DNA (ctDNA) in
urine and plasma expand the clinical utility to detect and monitor cancer
patient oncogenic mutations by minimally invasive and non-invasive liquid
biopsy methods.
Mutational tumor load quantification with high clinical sensitivity is vital for
robust individualized assessment of systemic therapeutic responsiveness
and resistance.
Concordance Study
Fold Change KRAS Copies/100K GE
(URINE)
o
Assay Performance
Fold Change KRAS Copies/100K GE
(URINE)
Background
Fold Change KRAS Copies/100K GE
(URINE)
Poster #: P143
Jason C. Poole1, Cecile Rose T. Vibat1, Lucie Benesova2, Barbora Belsanova2, Saege Hancock1, Errin Samuelsz1, Timothy T. Lu1, Vlada Melnikova1,
Mark G. Erlander1, Marek Minarik2
1Trovagene Inc. San Diego, USA, 2Center for Applied Genomics of Solid Tumors, Genomac Research Institute, Prague, Czech Republic
Conclusions
o A quantitative ctDNA assay using a massively parallel deep sequencing approach
was developed to determine patient ctDNA status of KRAS (Exon 2) mutations in
plasma and urinary ctDNA.
o We achieved an analytical sensitivity of 0.0067% in a background of WT DNA.
o In a blinded study of colorectal cancer patients with known KRAS mutational
status in tumor tissue, a correct KRAS mutation was identified in 95% of archival
plasma and 92% of archival urine specimens.
o Clear correlation and compatible fold change was observed between plasma and
urinary ctDNA KRAS levels in course of surgery and adjuvant therapy (chemo or
targeted).
o Analysis of association between the dynamics of ctDNA KRAS load and changes
in clinical status is ongoing.
o The first clinical correlation between urinary and plasma ctDNA.
For more information, please contact
Mark Erlander, PhD
11055 Flintkote Ave. San Diego, CA 92121
[email protected]
Partially supported by the
Czech Ministry of Health
grant no. NT 13660