Comprehensive genomic profiling of solid tumors for key ... · (CRL-2577). Number of replicates and...
Transcript of Comprehensive genomic profiling of solid tumors for key ... · (CRL-2577). Number of replicates and...
For Research Use Only. Not for use in diagnostic procedures. Thermo Fisher Scientific • 110 Miller Avenue Floor 2 • Ann Arbor, MI 48104 • thermofisher.com
TM
B (
muta
tions /
MB
)
A.
Figure 6: OCA Plus MSI test performance and reproducibility
INTRODUCTION
There is a growing trend in precision oncology research towards pan-cancer, tumor-agnostic approaches.
Comprehensive genomic profiling (CGP) of FFPE tumor samples by next-generation sequencing (NGS) is utilized
in both clinical and translational oncology research, and includes detection and analysis of single-gene and multi-
gene variants to identify biomarkers with the potential for future use in cancer diagnosis, prognosis and potential
therapeutic response[1]. Single-gene biomarkers are characterized by mutations in a single gene and include the
following types of genetic alterations: small variants including single nucleotide variants (SNVs) and indels, focal
copy number aberrations, and gene fusions. Multi-gene biomarkers, characterized by mutations in multiple
genes, include tumor mutation burden (TMB), microsatellite instability (MSI), loss of heterozygosity (LOH), and
often require sophisticated bioinformatics support for analysis and assessment.
The recent emergence of checkpoint inhibitors has stimulated research into multi-gene biomarkers that may be
associated with immunotherapy response. As the increased demand for single-gene and multi-gene biomarkers
drives the expansion of gene content accessible in a single panel, trade-offs are often needed in order to maintain
assay performance from FFPE samples and comes at the expense of input requirements, gene content, turn-
around time and bioinformatics capabilities. Furthermore, when faced with limiting FFPE samples, it is becoming
increasingly valuable to maximize the data that can be extracted from each sample.
To overcome these challenges, we developed a large comprehensive NGS assay that covers over 500 genes
across multiple tumor types and detects a range of important DNA and RNA variants associated with response to
targeted therapies and immune checkpoint inhibitors [2-6]. OncomineTM Comprehensive Assay Plus (OCA Plus)
requires low amounts of input FFPE material and enables detection of single gene biomarkers and multi-gene
biomarkers including TMB, MSI and mutational signatures such as DNA mismatch repair (dMMR). OCA Plus is
part of an automated sample-to-report workflow that offers rapid turn-around time and enables streamlined
comprehensive genomic profiling for oncology research.
MATERIALS AND METHODS
Gene content was prioritized based on potential clinical relevance and variant prevalence in solid tumors. Over
500 genes were selected including genes in the indication statements of approved drug labels, clinical guidelines,
and in the enrollment criteria of clinical trials. In addition, driver genes were selected in key pathways including
DNA repair and immune checkpoint response. Amplicon design strategies were optimized accordingly for key
hotspots, full coding sequences or copy number variation (CNV). The assay used Ion AmpliSeq™ technology with
manual library preparation or automated templating on the Ion Chef™ System and sequencing on the Ion
GeneStudio™ S5 platform. Twenty nanogram of purified DNA was routinely used as input. An automated tumor-
only workflow for variant calling and sample quality reporting was provided within Ion Reporter™ Software.
Streamlined access to reporting of variant relevance was enabled by Oncomine™ Reporter[7].
RESULTS
Over 500 genes with DNA based alterations and 49 RNA fusion drivers are included in the assay. Of the genes
measured for DNA alterations, 169 cover important cancer hotspots, 333 cover copy number variants (CNV), and
227 genes have full coding sequence (CDS) coverage for detection of truncating variants. The total genomic
coverage of the assay is 1.50MB with 1.05MB of exonic sequence, to support high-confidence TMB estimation. A
diverse set of microsatellite markers targeting MSI locations comprised of mono- and di-nucleotide repeats
ranging from 7 to 34 bp are included for MSI status assessment. RNA content for the assay is more fully
described in poster # 177/11 “RNA sequencing-based gene fusion detection with Oncomine Comprehensive
Assay Plus”.
FFPE tumor samples from a variety of tissue types, and commercially available controls were sequenced using
the assay. The assay displays high (>95%) uniformity and consistent read depth (>2200x) to support robust
variant calling at low allele frequency. Excellent variant calling (SNV/INDEL) performance was observed, with
sensitivity ranging from 98%-100% and PPV ranging from 92%-100%, depending upon the variant type. CNV
detection sensitivity and specificity were 92%-99% and 100%, respectively. In-silico TMB assessment using
publicly available whole-exome cancer sequencing data resulted in a correlation of R2 > 0.90 (0-40 mut/mb) in
pan-cancer and specific cancer types including lung, colorectal and melanoma. TMB correlation (R2) was >0.96
with whole-exome sequencing (WES) in FFPE solid-tumor samples. MSI test performance was assessed using
172 FFPE samples with known MSI status. An overall performance of 96% sensitivity and 100% specificity was
observed. One-hundred percent sensitivity and specificity was observed in targeted gene-fusion using
commercially available reference controls.
CONCLUSIONS A targeted NGS assay was developed to support comprehensive genomic profiling in translational and clinical
oncology research. The assay includes more than 500 relevant genes across multiple cancer types and can detect
single and multi-gene biomarkers. All gene variant types can be analyzed (SNVs, indels, CNVs, gene fusions) along
with key immuno-oncology biomarkers including TMB and MSI and other mutational signatures. The OCA Plus
workflow ensures optimized characterization of all gene content from just 20 ng FFPE in 4 days for comprehensive
genomic profiling without trade-offs.
REFERENCES1. Hovelson DH, McDaniel AS, Cani AK, Johnson B, Rhodes K, Williams PD, Bandla S, Bien G, Choppa P, Hyland F, Gottimukkala R, Liu G, Manivannan M, Schageman J,
Ballesteros-Villagrana E, Grasso CS, Quist MJ, Yadati V, Amin A, Siddiqui J, Betz BL, Knudsen KE, Cooney KA, Feng FY, Roh MH, Nelson PS, Liu CJ, Beer DG, Wyngaard
P, Chinnaiyan AM, Sadis S, Rhodes DR, Tomlins SA: Development and Validation of a Scalable Next-Generation Sequencing System for Assessing Relevant Somatic
Variants in Solid Tumors. Oncogene 2015, 17:385-399.
2. Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, Brannigan BW, Harris PL, Haserlat SM, Supko JG, Haluska FG, Lous DN, Christiani DC, Settleman J,
Haber DA: Activating Mutations in the Epidermal Growth Factor Receptor Underlying Responsiveness of Non-Small-Cell Lung Cancer to Gefitinib. NEJM 2004: 350:2129-
2139.
3. Van Cutsem E, Köhne CH, Hitre E, Zaluski J, Chien CRC, Makhson A, D’Haens G, Pintér T, Lim R, Bodoky G, Rho JK, Folprecht G, Ruff P, Stroh C, Tejpar S, Schlichting M,
Nippgen J, Rougier P: Cetuximab and Chemotherapy as Initial Treatment for Metastatic Colorectal Cancer. NEJM 2009, 360:1408-1417
4. Camidge DR, Bang YJ, Kwak EL, Iafrate JA, Varella-Garcia M, Fox SB, Riely GJ, Solomon B, Ou SHI, Kim DW, Salgia R, Fidias P, Engelman JA, Gandhi L, Jänne PA,
Costa DB, Shaprio GI, LoRusso P, Ruffner K, Stephenson P, Tang Y, Wilner K, Clark JW, Shaw AT: Activity and Safety of Crizotinib in Patients with ALK-Positive Non-Small-
Cell Lung Cancer: Updated Results from a Phase 1 Study. Lancet Oncol. 2012 13:1011-1019
5. Le DT, Uram JN, Wang H, Bartlett BR, Kemberling H, Eyring AD, Skora AD, Luber BS, Azad NS, Laheru D, Biedrzycki B, Donehower RC, Zaheer A, Fisher GA, Crocenzi TS,
Lee JJ, Duffy SM, Goldberg RM, de la Chapelle A, Koshiji M, Bhaijee F, Huebner T, Hruban RH, Wood LD, Cuka N, Pardoll DM, Papadopoulos N, Kinzler KW, Zhou S,
Cornish TC, Taube JM, Anders RA, Eshleman JR: PD-1 Blockade in Tumors with Mismatch-Repair Deficiency. NEJM 2015, 372:2509-2520
6. Snyder A, Makarov V, Merghoub T, Yuan J, Zaretsky JM, Desrichard A, Walsh L, Postow MA, Wong P, Ho TS, Hollmann TJ, Bruggeman C, Kannan K, Li Y, Elipenahli C, Liu
C, Harbison CT, Wang L, Ribas A, Wolchok JD, Chan TA:Genetic Basis for Clinical Response to CTLA-4 Blockade in Melanoma. NEJM 2014, 371:2189-2199
7. Oncomine Reporter: https://www.thermofisher.com/order/catalog/product/A34298#/A34298
8. AcroMetrix Oncology Hotspot Control: https://www.thermofisher.com/order/catalog/product/969056
9. Ellrott K, Bailey MH, Saksena G, Covington KR, Kandoth C, Stewart C, Hess J, Ma S, Chiotti KE, McLellan M, Sofia HJ, Hutter C, Getz G, Wheeler D, Ding L, MC3 Working
Group: Scalable Open Science Approach for Mutation Calling of Tumor Exomes Using Multiple Genomic Pipelines. Cell Systems 2018, 6:271-281.e7
10.https://www.focr.org/tmb
11.https://www.thermofisher.com/us/en/home/clinical/clinical-genomics/molecular-oncology-solutions/microsatellite-instability-assay.html
Vinay Mittal, Jennifer Kilzer, Gary Bee, Santhoshi Bandla, Dinesh Cyanam, Sameh El-Difrawy, Aren Ewing, Anelia Kraltcheva, Denis Kaznadzey, Cristina Van Loy, Scott Myrand, Warren Tom, Yu-Ting Tseng, James Veitch, Paul Williams, Chenchen Yang, Janice Au-Young, Zheng Zhang, Fiona Hyland, Elaine Wong-Ho, Seth SadisThermo Fisher Scientific, Ann Arbor, MI, USA; South San Francisco, CA; Austin, TX, Carlsbad, CA.
Comprehensive genomic profiling of solid tumors for key targeted and immuno-oncology biomarkers research using Ion Torrent NGS technology on the Oncomine
Comprehensive Assay Plus
Figure 3. In-Silico comparison of OCA Plus TMB with Whole Exome Sequencing (WES)
Figure 1. Oncomine Comprehensive Assay Plus Cancer Gene Targets
Table 1. Representative variant calling performance using controls and FFPEs
Table 2. Representative copy-number variant calling performance
The OCA Plus workflow enables detection of copy number gain in oncogenes and copy number loss in tumor suppressor
genes. CNV detection performance was assessed using known copy number status from internally characterized solid
tumor FFPE samples (n=9), commercially available ATCCTM cell lines (n=3), and BioChainTM CancerSeqPlus Paraffin
Tissue Curl. For the FFPE samples, known truth was determined using FISH and orthogonal Oncomine NGS assays; for
ATCCTM cell lines, known truth was derived from the literature and publicly available genotyping expression arrays; for
BioChainTM samples, known truth was obtained from the product insert guide. Two replicates of each sample was used.
Gene level copy gain is called when the lower confidence interval (CI) of the copy number estimate is ≥4; copy loss is
called when the CI upper bound <2 and fold-change relative to normal baseline is <0.85. For specificity calculations, wild-
type copy-number estimates (autosomal copy number) were considered as true-negatives for selected genes while CN
estimates outside the normal CN range were considered as false-positive.
OCA Plus determines MSI status in tumor samples. Using the primary signal from semi-conductor sequencing, the MSI
algorithm infers the amount of change in the homopolymer length of MSI target markers present on the panel relative to
the known control for each target microsatellite. An overall MSI score is calculated and MSI status is determined using the
upper and lower bound thresholds of the MSI score. A. 171 solid tumor samples (pan-cancer) with % tumor cellularity of
≥15%. Samples were previously typed for MSI using on-market tests which include BAT-25, BAT-26, BAT-40, CAT-25, NR-
21, NR-22, NR-24, NR-27 in addition to proprietary MSI markers[11]. Samples flagged for QC by the analysis workflow
were excluded. An overall sensitivity of ~96% and specificity of 100% was observed. B. Assessment of reproducibility in
MSI using OCA Plus in selected samples (tumor FFPE, FFPE control, normal cell-line (NA12878) and ATCCTM Cell-line
(CRL-2577). Number of replicates and associated CV (coefficient of variation; standard -deviation/mean) is indicated.
OCA Plus ResultsReference Test Results
Positive Negative
Positive (MSI-
High)
TP=44 FP=0
Negative (MSS) FN=2 TN=125
Sensitivity 95.6%
Specificity 100%
To assess small variant calling performance, we analyzed reference controls and clinical FFPE samples. AcroMetrix
Oncology Hotspot Control (AOHC, n=4) contains hundreds of SNVs and tens of indels, and Seraseq™ Tri-Level Tumor
Mutation DNA Mix v2 (n=2) contains tens of SNV/indels. Ten FFPE solid tumor samples (analyzed in duplicate) were also
sequenced to characterize performance with real-world samples. SNV hotspot and de novo performance was assessed at
variant allele frequency LOD of ≥ 5%, indel hotspots at LOD of 10% and de novo indel at LOD of 20%.
Synthetic and FFPE controls Tissue FFPEs
High TMB is associated with microsatellite
instability in certain tissue types. For this study, 20
colorectal FFPE tumor samples previously typed
for microsatellite instability (MSI) were sequenced
using OCA Plus and TMB was determined. The
TMB scores obtained with OCA Plus were
significantly higher in the 12 MSI samples relative
to the 8 MSS samples (Student’s t-test p-value =
1.13E-06).
Figure 4. OCA Plus TMB correlation with WES
Figure 5. Correlation of TMB estimates with MSI status in FFPE samples
TMB performance was evaluated through in-silico comparison with WES. WES data was obtained from TCGA MC3[9]. The
OCA Plus mutation rate was computed by limiting the mutations to the targeted genomic regions. TMB was estimated as
mutation counts per megabase of the targeted genomic region. The scatter plot in Fig 3A, shows the correlation between
OCA Plus (y-axis) and WES (x-axis) TMB estimates for pan-cancer. The results show high correlation in TMB values with
R2 = 0.902. Figure 3B displays similar analyses for select cancer types (BLCA, Bladder Urothelial Carcinoma; COAD, Colon
Adenocarcinoma; HNSC, Head and Neck squamous cell carcinoma; LUAD, Lung Adenocarcinoma; LUSC, Lung Squamous cell carcinoma; SKCM, Skin
Cutaneous Melanoma; STAD, Stomach adenocarcinoma; UCEC, Uterine Corpus Endometrial Carcinoma).
Correlation of TMB assessment with WES. In Figure 4A, ten cancer cell-line samples, provided by Friends of Cancer
Research (FOCR)[10], were evaluated with OCA Plus, Oncomine Tumor Mutation Load Assay (OTMLA), and WES. While
OCA Plus and OTMLA TMB scores were from tumor-only workflow, WES analysis was performed via paired-tumor
normal (matched) analysis. OCA Plus and OTMLA correlated well with WES analysis of FOCR cell lines. In Figure 4B,
21 solid tumor FFPE samples were evaluated with OCA Plus and WES. High correlation was observed with R2 = 0.962.
MSSMSI
OC
A
Plu
s T
MB
(M
uta
tions /
MB
)
OC
A P
lus (
mu
tatio
ns / M
B)
WES (mutations/MB)
Cancer Type
R2 = 0.902
WES (mutations/MB)
B.A.
B.
R2= 0.822
FFPE WES TMB (Mutations / MB)
Variant type TP FP FN % Sen. % PPV
Hotspot SNV 588 0 6 98.98 100
De-novo SNV 419 2 5 98.82 99.52
Hotspot INDEL 28 0 0 100 100
De-novo INDEL 12 1 0 100 92.30
Variant type TP FP FN % Sen. % PPV
Hotspot SNV 24 0 0 100 100
De-novo SNV 4 0 0 100 100
Hotspot INDEL 2 0 0 100 100
De-novo INDEL 8 0 0 100 100
R2 = 0.962
Figure 7: Targeted gene fusion detection using tumor gene fusion RNA reference controls
OC
A P
lus T
MB
(m
uta
tions /
MB
)
Cell
Lin
e-0
1
Cell
Lin
e-0
2
Cell
Lin
e-0
3
Cell
Lin
e-0
4
Cell
Lin
e-0
5
Cell
Lin
e-0
6
Cell
Lin
e-0
7
Cell
Lin
e-0
8
Cell
Lin
e-0
9
Cell
Lin
e-1
0
OC
A P
lus M
SI
Score
Tumor FFPE FFPE Control NA12878 CRL-2577
CV: 0.10; n=9
CV: 0.12; n=6
CV: 0.19; n=14
CV: 0.05; n=16
A. B.Sensitivity and Specificity
Reproducibility
Figure 2. Oncomine Comprehensive Assay Plus performance with clinical samples
Assay performance was evaluated for variant detection across different mutational types using a pool of commercially
sourced, orthogonally validated FFPE samples. All variants were successfully detected with 100% sensitivity.
OC
A P
lus (
mu
tatio
ns / M
B)
Post-sequencing
analysis
Ion Reporter™ Software
Oncomine™ Reporter
Tumor sample
DNA and RNA extracted from FFPE
Next-generation sequencing
Ion GeneStudio™ S5 System
Library and
template preparation
Targeted assay
Ion AmpliSeq™
chemistry
Figure 8. Schematic flow-diagram of the complete four-day workflow
The Oncomine Comprehensive Assay Plus workflow uses a recommended 20 ng of input FFPE material (10 ng per
amplicon pool). The assay can leverage manual or automated library preparation and templating on the Ion Chef. Up to
four samples can be multiplexed on the Ion 550 chip to achieve sufficient read depth. The analysis pipeline utilizes a
custom variant calling pipeline optimized for solid-tumor samples, quantification of somatic mutations for TMB, and a
novel algorithm that leverages the unique signal processing properties inherent in semi-conductor sequencing for MSI
detection. The analysis solution is optimized as a tumor sample only workflow for comprehensive genomic profiling of
clinical research samples.
www.thermofisher.com/2020AACR-posters
Fusion read count Fusion read count
Data is based on OncomineTM Reporter v4.6, March 2020
Fusion detection performance was evaluated using Seraseq(R) RNA controls (SeraCare v4 and SeraCare NTRK), with
expected fusions derived from the SeraCare product insert guides. Gene fusions in the control samples were
sequenced and characterized using the OCA Plus workflow and fusion pipeline. The threshold for a positive fusion call
was an absolute read count of ≥ 100. Fig. 7A: Detection of 16/16 fusion isoforms and 2/2 intragenic rearrangements
(MET exon-14 skip, EGFRvIII) in SeraCare v4 control. Fig. 7B: Detection of 15/15 NTRK fusion isoform in SeraCare
NTRK control. 100% detection sensitivity was maintained when controls were diluted down to 10% (in the background
of normal total RNA). No false positive call above the read threshold (100) was observed. The overall sensitivity and
specificity were 100%.
Post-sequencing
analysis
Ion Reporter™ Software
Oncomine™ ReporterTP FP FN TN Sensitivity Specificity
CNV Gain 28 0 0 16 100%
100%
CNV loss 30 0 2 18 93%
A. B. Non-diluted
Diluted to 10%
Non-diluted
Diluted to 10%
TRADEMARKS/LICENSINGAll trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified. Seraseq is a registered
trademark of SeraCare. ATCC is a registered trademark of the American Type Culture Collection. Study sponsored by Thermo Fisher.
Conflict of Interest: The authors are employees of Thermo Fisher.
ACKNOWLEDGEMENTSPatients and their families for donation of samples to scientific studies and for public access.
Contact Information: [email protected]