Abstract
PURPOSE
Although primary germ cell tumors (GCTs) have been extensively characterized, molecular analysis of metastatic sites has been limited. We performed whole-exome sequencing and targeted next-generation sequencing on paired primary and metastatic GCT samples in a patient cohort enriched for cisplatin-resistant disease.
PATIENTS AND METHODS
Tissue sequencing was performed on 100 tumor specimens from 50 patients with metastatic GCT, and sequencing of plasma cell-free DNA was performed for a subset of patients.
RESULTS
The mutational landscape of primary and metastatic pairs from GCT patients was highly discordant (68% of all somatic mutations were discordant). Whereas genome duplication was common and highly concordant between primary and metastatic samples, only 25% of primary-metastasis pairs had ≥ 50% concordance at the level of DNA copy number alterations (CNAs). Evolutionary-based analyses revealed that most mutations arose after CNAs at the respective loci in both primary and metastatic samples, with oncogenic mutations enriched in the set of early-occurring mutations versus variants of unknown significance (VUSs). TP53 pathway alterations were identified in nine cisplatin-resistant patients and had the highest degree of concordance in primary and metastatic specimens, consistent with their association with this treatment-resistant phenotype.
CONCLUSION
Analysis of paired primary and metastatic GCT specimens revealed significant molecular heterogeneity for both CNAs and somatic mutations. Among loci demonstrating serial genetic evolution, most somatic mutations arose after CNAs, but oncogenic mutations were enriched in the set of early-occurring mutations as compared with VUSs. Alterations in TP53 were clonal when present and shared among primary-metastasis pairs.
CONTEXT
Key Objective
The genomic underpinnings of metastasis and chemotherapy resistance in germ cell tumors (GCTs) have not been fully characterized. We performed whole-exome sequencing and targeted next-generation sequencing on paired primary and metastatic GCT specimens from 50 patients, with analysis of plasma cell-free DNA (cfDNA) in a subset.
Knowledge Generated
We identified striking genomic heterogeneity between primary tumors and metastases. Phylogenetic reconstruction found that most mutations arose after copy number events involving their respective genomic loci. However, oncogenic mutations were more likely to be concordant between primary-metastasis pairs and to occur earlier in tumor evolution. TP53 mutations represented early clonal events that were concordant between primary-metastasis pairs and could be accurately characterized using plasma cfDNA analysis.
Relevance
Additional insight into the molecular basis of GCT progression will require broader analyses across multiple disease states with emphasis on metastatic sites. Plasma cfDNA analysis may enable noninvasive determination of TP53 mutation status.
INTRODUCTION
Testicular cancer is the most common solid tumor in young men (15-40 years of age), with 9,610 new cancers and 440 deaths estimated to occur in the United States in 2020.1 Although 95% of germ cell tumors (GCTs) arise in the testis, GCTs can also arise at extragonadal sites. GCTs are distinguished by exceptional sensitivity to cisplatin- and etoposide-based chemotherapy regimens, with cure expected to be achieved in nearly 80% of patients with metastatic disease.2 However, the emergence of chemotherapy resistance in a subset of patients with GCT represents a critical challenge, with disease-related mortality expected in approximately half of this group.
Risk stratification based on clinical factors is currently used to guide primary treatment and postchemotherapy management.3 A lack of understanding of the molecular alterations that drive GCT development, progression, and chemotherapy resistance has, however, limited the use of molecular data to assess GCT prognosis and guide therapy selection. Previous analysis by our group of a data set of 180 patients with testicular and mediastinal GCT (including 104 chemotherapy-resistant patients) using whole-exome sequencing (WES) or targeted next-generation sequencing (NGS) identified alterations in TP53 and MDM2 as independently associated with cisplatin resistance and inferior outcomes.4 Another WES analysis of 59 tumors from 51 patients described a pattern of widespread reciprocal loss of heterozygosity (RLOH) in GCTs.5 Phylogenetic analysis of five patients from this series in which primary and metastatic disease sites were concurrently studied suggested that chemotherapy resistance may be associated with increased RLOH events and loss of expression in the pluripotency and apoptosis regulators NANOG and POU5F1.5
With the goal of characterizing the biology of tumor evolution in the context of metastatic progression and chemotherapy resistance in GCT, we performed WES and/or targeted NGS6,7 on paired primary and metastatic tumor samples from 50 patients. We also performed targeted NGS on matched plasma cell-free DNA (cfDNA) samples collected from a subset of 11 patients with available material.
PATIENTS AND METHODS
Patient Eligibility
Institutional review board approval was obtained, and all samples were collected from patients at Memorial Sloan Kettering Cancer Center who had provided informed consent for tumor molecular characterization. Patients were eligible if they had tumor tissue containing viable GCT from both primary (testicular, ovarian, or mediastinal) and metastatic sites and provided a blood specimen for matched normal DNA.
NGS
Primary-metastasis pairs from a discovery cohort of 10 patients were selected for WES analysis. An additional 40 primary-metastasis pairs were used as an expansion cohort. For all 50 patients, targeted NGS was performed using the Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT) platform, a hybrid capture-based assay targeting 341, 410, or 468 cancer-associated genes at an average coverage of 631×. Methods regarding sample assessment, DNA extraction, library preparation, sequencing, and analysis have been previously described.6,7 Plasma cfDNA from 11 patients with available blood samples was also analyzed using MSK-IMPACT. Mutations were called as previously described6,8 with common variants filtered out using ExAC (Broad Institute, Cambridge, MA). Variants overlapping human Encyclopedia of DNA Elements genomic regions blacklisted for functional genomics analysis and repeat regions were also filtered. For the cfDNA data, mutations identified in the corresponding tumor tissue samples from the same patient were considered present if supported by one or more reads.
Total, allelic, and integer copy number alterations (CNAs) were estimated using the FACETS algorithm.9 Cancer cell fractions (CCFs) were calculated as previously described10 using local copy number determined by FACETS. Mutations were considered clonal if the CCF binomial upper CI was ≥ 0.8. Tumor samples were considered to have undergone whole-genome duplication (WGD) if the fraction of major allele > 1 was > 50%.11 RLOH was defined as gains of one parental allele with simultaneous loss of the other parental allele, resulting in loss of heterozygosity, either copy neutral or with amplification of the remaining parental allele.5
The timing of somatic mutations relative to observed DNA CNAs at the same locus was estimated for all mutated loci for which unambiguous timing was possible. Relative timing was determined using the most parsimonious explanation of observed copy number state.11 For concordant mutations (ie, those found in both primary and metastatic samples), only the primary sample mutation was analyzed.
Data Availability
The assembled prospective somatic mutational data for the entire cohort have been deposited for visualization and download in the cBioPortal for Cancer Genomics. WES data can be obtained from dbGaP (accession number: phs002229.v1.p1).
RESULTS
Patient and Tumor Characteristics
Of the primary tumor samples, the majority were from the testis (90%), with the balance of samples of ovarian or mediastinal origin (Table 1). Sites of metastatic disease varied and included retroperitoneal lymph nodes, liver, lung, and brain. Primary tumor and metastatic samples were predominantly nonseminomatous GCT (NSGCT; 72% for both) and were largely histologically concordant. One patient had seminoma only in the primary tumor but choriocarcinoma in metastatic retroperitoneal lymph nodes. The metastatic site in one patient with primary NSGCT reflected a secondary somatic malignancy with adenocarcinoma. In total, 80% of all patients had cisplatin-resistant disease, and the majority of metastatic samples were collected after chemotherapy (66%). Median follow-up was 33 months.
TABLE 1.
Patient and Sample Characteristics
Low Mutational Concordance Across Primary-Metastasis Matched Pairs
Consistent with prior studies, WES of the primary GCT site revealed a low tumor mutation burden (median, 0.42 mutations per megabase [mut/Mb]; range, 0.1-1.52 mut/Mb), whereas the paired metastasis samples had a significantly higher tumor mutational burden (median, 1.55 mut/Mb; range, 0.16-4.79 mut/Mb; Wilcoxon signed rank, P = .03; Fig 1A). Differences in sample purities did not account for the observed difference in tumor mutational burden between primary and metastatic disease sites (Wilcoxon signed rank, P = .4; Appendix Table A1). This set of 10 matched paired samples was also sequenced from the same sequencing libraries using the MSK-IMPACT targeted gene panel to test consistency across assays. The MSK-IMPACT assay sequences samples to greater depth than WES6 and thus can detect subclonal mutations with greater sensitivity. Notably, the variant allele frequency of mutations identified in genes covered by MSK-IMPACT was highly correlated with the variant allele frequencies from WES analysis (Spearman rank correlation, ⍴ = 0.91; P < 2.2 × 10−16; Appendix Fig A1). Furthermore, despite the greater depth of sequencing of MSK-IMPACT, there were no mutations detected by MSK-IMPACT that were not detected by WES in those genomic regions covered by both assays.
FIG 1.
Mutational heterogeneity of paired primary and metastatic germ cell tumor (GCT) specimens analyzed by whole-exome sequencing. (A) Primary GCT samples had a low mutation rate (median, 0.42 mutations per megabase [mut/Mb]; range, 0.1-1.52 mut/Mb), whereas paired metastasis samples from the same patients demonstrated a significantly higher tumor mutational burden (median, 1.55 mut/Mb; range, 0.16-4.79 mut/Mb; Wilcoxon signed rank, P = .03). Individual mutations were designated as occurring in the primary sample (orange) and/or the metastasis sample (green) and identified as oncogenic (red) or variants of unknown significance (VUSs; blue). (B) The median percentage of concordant mutations was 32.4% (range, 0%-70.3%), with six of 10 patients having a majority of discordant mutations. Oncogenic mutations were more likely to be concordant across primary-metastasis pairs (72.7%; eight of 11 concordant) compared with variants of unknown significance (19.1%; 111 of 581 concordant; Fisher’s exact test, P = .0002).
Mutations were considered concordant across primary and metastasis specimen pairs if there was any evidence for the mutation in both samples, no matter their clonality in the affected sample. The median percentage of concordant mutations was 32.4% (range, 0%-70.3%), with six of 10 patients having a majority of discordant mutations (Fig 1B). By WES, we identified 11 oncogenic mutations in seven of these 10 patients (median, one mutation; range, zero to four mutations), defined as mutations known or predicted to be oncogenic based on the OncoKB precision oncology knowledge base.12 Notably, oncogenic mutations were more likely than variants of unknown significance (VUSs) to be concordant between specimens from the same patient (73% v 19% concordant, respectively; Fisher’s exact test, P = .0002; Fig 1A). Oncogenic mutations present in the primary tumors but absent in their corresponding metastatic samples included a clonal PIK3CD C416R mutation and a subclonal BCL2L11 G156R mutation. In another patient, a clonal SETD2 truncating mutation detected in a metastatic sample was absent in the corresponding primary tumor specimen.
To determine whether any oncogenic mutations were recurrently discordant between primary and metastatic tumor sites in patients with GCT, we performed targeted sequencing analysis of an independent cohort of 40 additional patients with GCT using the MSK-IMPACT panel. From this cohort, 34 of 40 primary-metastasis pairs had sufficient tumor content (purity) to call mutations in both samples. Four of these 34 matched pairs had no detectable mutations via MSK-IMPACT (Appendix Table A1). For the remaining 30 pairs, metastasis samples also had a higher number of mutations (median, 1.76 mut/Mb; range, 0-6.69 mut/Mb) than primary samples (median, 0.98 mut/Mb; range, 0-3.94 mut/Mb; Wilcoxon signed rank, P = .001). Across all 30 patients, 29 (34.5%) of 84 mutations were concordant. Only eight (26.7%) of 30 patients had a majority of concordant mutations (Appendix Fig A2). As in the WES data set, oncogenic mutations detected by MSK-IMPACT were more likely to be concordant, with 13 (46.4%) of 28 oncogenic mutations detected in both the primary and metastatic tumors versus 16 (28.6%) of 56 VUSs. However, given the low mutational rate of GCTs, this result was not statistically significant (Fisher’s exact test, P = .14).
Concordance of Early Copy Number Events Is Evident Across Primary-Metastasis Pairs
As a result of the design of the MSK-IMPACT assay with uniformly distributed tiling probes across the genome, we were able to perform copy number analysis as a unified WES and MSK-IMPACT cohort. Of the 50 primary-metastasis pairs, 40 pairs (WES, n = 10; IMPACT, n = 30) were of sufficient purity in both samples to compare genome-wide copy number profiles (Appendix Table A1). Because GCTs have one of the highest rates of WGD across cancer types,11 we analyzed each sample for evidence of WGD. Of the 40 primary tumor samples, 35 (88%) displayed evidence of WGD, in contrast to 32 (80%) of 40 metastatic samples, with a concordance rate of 87.5% (35 of 40 samples) across the primary-metastasis pairs (primary-only WGD, n = 4; metastasis-only WGD, n = 1). We also observed a previously described high rate of reciprocal deletion on chromosomal arms with amplifications (ie, RLOH).5 In primary samples, 39% of arm-level amplifications had an accompanying RLOH, whereas in metastatic samples, 51% of arm-level amplifications displayed RLOH. Additional losses and gains in individual samples from each pair were reflected in the total percentage of the genome with the same major and minor integer copy number (as measured in base pairs sequenced in both samples), with only 10 (25%) of 40 pairs demonstrating ≥ 50% identical integer copy number (median, 39%; range, 16.5%-79%; Fig 2A).
FIG 2.
Copy number analysis and timing of mutations during tumor evolution. (A) Continuous losses and gains, as partly reflected in the discordance of whole-genome duplication and reciprocal loss of heterozygosity rates across 40 evaluable primary-metastasis pairs (sequenced by whole-exome sequencing [WES] and/or Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets—MSK-IMPACT—were reflected in the total percentage of the genome with the same major and minor integer copy number (as measured in base pairs sequenced in both samples). Ten (25%) of 40 pairs demonstrated ≥ 50% identical integer copy number (median, 38.9%; range, 16.5%-78.9%). (B) For 10 primary-metastasis pairs analyzed by WES (each represented by a different colored dot), the majority of mutations arose after the observed copy number alteration at the respective locus: (1) before: median, 2.5 mutations, range (1-10); (2) after: median, 25.5 mutations, range (8-66); Wilcoxon signed rank test, P = .006). (C) For primary-metastasis pairs analyzed by WES, oncogenic mutations were enriched in the set of early-occurring mutations compared with variants of unknown significance (VUSs). Five (62.5%) of eight oncogenic mutations occurred early compared with 30 (10.5%) of 287 VUSs in the WES samples (Fisher’s exact test, P = .001). (**) Statistical significance for the illustrated comparisons.
Most Mutations Arise Late During Tumor Evolution
To define the timing at which mutations arose in the primary and metastasis pairs (n = 10) in relation to the allelic copy number state present at the respective loci, mutations were split into two groups, those found in the primary sample (both concordant and discordant) and those found exclusively in the metastatic sample. For those mutations that could be unambiguously timed (median, 64%; range, 34%-83% across samples), we determined whether the mutation occurred before or after the observed copy number event at the same locus by comparing the allelic frequencies of the mutant variant and the allelic CNA observed at the locus. The majority of mutations arose after the observed CNA at the respective locus (Wilcoxon signed rank, P = .006; Fig 2B). A similar pattern was observed in the MSK-IMPACT cohort with significantly more mutations occurring after the observed CNAs (Wilcoxon signed rank, P = 6.297 × 10−5; Appendix Fig A3A). Furthermore, in both cohorts, oncogenic mutations were enriched in the set of early-occurring mutations compared with VUSs. In the WES samples, five (62.5%) of eight oncogenic mutations occurred early (ie, before the observed copy number event at the locus) compared with 30 (10.5%) of 287 VUSs (Fisher’s exact test, P = .001; Fig 2C). Similarly, in the MSK-IMPACT sequenced tumor pairs, seven (41.2%) of 17 oncogenic mutations arose early, compared with four (11.1%) of 36 VUSs (Fisher’s exact test, P = .03; Appendix Fig A3B).
TP53 Pathway Alterations in Cisplatin-Resistant GCTs
Prior work by our group suggests that TP53 pathway alterations including TP53 mutations and MDM2 amplification are associated with chemotherapy resistance in patients with GCT. Nine patients had TP53 pathway alterations, including three patients with TP53 truncating mutations, all of which were concordant in the primary and metastatic pairs. Another six patients had MDM2 amplifications, four of which were concordant across their respective primary-metastasis pairs. For the two MDM2 discordant patients, MDM2 amplifications were only present in the metastatic samples collected after chemotherapy (Fig 3). Because the detection of TP53 pathway alterations may be useful in guiding treatment selection in patients with GCT, we assessed whether analysis of cfDNA could accurately characterize TP53 status. Thus, plasma samples from 11 patients were analyzed using the MSK-IMPACT assay to provide primary-metastasis-cfDNA trios for comparison. TP53 status was concordant in all patients analyzed, including two patients whose tumors were TP53 mutant (Fig 4A) and nine patients whose tumors were TP53 wild type. Across all 11 trios, a total of 19 mutations (including nine oncogenic mutations) were detected (median, two mutations; range, zero to five mutations) in cfDNA. Analysis of cfDNA did not reveal any additional mutations not detected in tissue. In contrast, 19 mutations identified in the tissue samples were absent in the matched cfDNA sample, including seven oncogenic mutations (Fig 4B). Notably, the majority of mutations detectable in cfDNA (15 [79%] of 19 mutations) were concordant in the primary and metastatic tissue samples. Of the four mutations identified by cfDNA that were not concordant in the primary-metastasis pairs, two each were identified in primary-cfDNA and metastasis-cfDNA pairs, respectively (Fig 4B).
FIG 3.
Discordance of MDM2 amplification across primary-metastasis pairs in two patients. (A) In the first patient, the MDM2 amplification was identified only in the metastatic sample collected after chemotherapy. (B) The same finding is demonstrated in a second patient.
FIG 4.
Analysis of plasma cell-free DNA (cfDNA) samples from 11 patients using Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets—MSK-IMPACT, which enabled comparison of primary-metastasis-cfDNA trios. (A) TP53 status was concordant across primary, metastasis, and plasma cfDNA specimens in two patients whose tumors were TP53 mutant. Additional mutations in the trio samples from these two patients are also shown. (B) Across all 11 trios, a total of 19 mutations were detected (median, two mutations, range (0-5) in cfDNA, including nine oncogenic mutations. Plasma cfDNA analysis did not reveal any additional mutations undetected in tumor tissue. Nineteen mutations identified in the tumor samples were absent in the matched cfDNA sample, including seven oncogenic mutations. The majority of mutations (15 of 19) detected in cfDNA were concordant in the primary-metastasis pair. Four mutations identified by cfDNA were not concordant in the primary-metastasis pairs, with two each identified in primary-cfDNA and metastasis-cfDNA pairs, respectively. VUS, variant of unknown significance.
DISCUSSION
To our knowledge, we report the largest genomic analysis to date of paired primary and metastatic GCT specimens. This cohort was highly enriched for patients with cisplatin-resistant disease, with most metastatic samples collected after treatment with chemotherapy. Our results highlight a high degree of heterogeneity in somatic mutations and CNAs between primary and metastatic sites in patients with GCT. Notably, oncogenic mutations were more likely to be concordant between primary-metastasis pairs, and mutations that arose early (before the copy number event at the respective locus) were enriched for those known or predicted to be oncogenic. Importantly, TP53 mutations, which have previously been associated with cisplatin-resistant disease,4 were always clonal and, when present, were identified in both the primary and metastatic sites. In the current cohort, 8% of cisplatin-resistant patients harbored TP53 mutations, and such alterations were not found in cisplatin-sensitive patients, consistent with prior observations. Two of these patients were distinct from the previously reported cohort.4 These results further support a role for TP53 dysregulation in the biology of a subset of patients with cisplatin-resistant disease, perhaps through the abrogation of mitochondrial priming.5 TP53 mutations could be reliably identified using plasma cfDNA NGS, which may offer a potential noninvasive alternative to tissue biopsies for the identification of patients whose tumors harbor this predictive and prognostic biomarker.
Evolutionary-based analyses revealed a frequently observed sequence of genomic events in GCT—early WGD followed by genome instability with copy number losses and gains occurring both before and after metastasis. Although the majority of somatic mutations arose after CNAs at serially evolving loci genome wide, a small number of mutations were identified in metastatic samples that arose before the respective CNA, suggesting a highly dynamic and continually evolving landscape in GCTs through metastasis and therapy. These findings complement a recent comprehensive genomic landscape analysis that included comparison of longitudinal samples from 17 patients with platinum-resistant GCT, identifying substantial mutational heterogeneity between matched primary and metastatic tumors as well as accumulation of copy number events over time.13 An intriguing consideration is whether this genomic evolution under the selective pressure of therapy would expose a targetable vulnerability in a more advanced disease state. With the goal of identifying alterations relevant to the cisplatin-resistant disease state, we propose that future genomic analyses should prioritize the analysis of postchemotherapy metastatic sites or plasma cfDNA in addition to primary or prechemotherapy tissue.
Given the low mutation rate and high degree of patient-to-patient genomic heterogeneity of GCTs observed in this and other studies, this cohort, although the largest reported to date, was not sufficiently powered to identify recurrent mutations found only in metastatic specimens, with a truncating mutation in SETD214 representing the only metastasis-exclusive oncogenic mutation identified. Because the genes most commonly mutated in human cancer are infrequently altered in GCTs, whole-genome sequencing, transcriptome sequencing, and epigenetic profiling platforms should be considered for future studies of GCT biology. We acknowledge several other limitations to the analysis, as follows: lack of uniformity with regard to pre- or postchemotherapy metastatic tissue; a wide range of sample collection intervals; limited follow-up for some patients; the presence of teratoma in a subset of patients, which could have inadvertently been sequenced; and variable plasma cfDNA collection timing relative to metastatic tissue collection. We also acknowledge that spatial heterogeneity in the pretreatment samples may confound interpretation of primary versus metastasis sequencing. Nevertheless, to our knowledge, this is the first large series to evaluate the genomic concordance between paired primary and metastatic GCTs and lends insight into the understanding of the biologic evolution of these tumors.
In summary, our analysis of pairs of matched primary and metastatic tumor GCT revealed significant lesion-to-lesion genomic heterogeneity, especially for mutations but also for CNAs. TP53 mutations were a notable exception in that they were early clonal events when present and restricted to cisplatin-resistant tumors. Because TP53 mutations could be detected noninvasively through analysis of plasma cfDNA, this may represent a promising approach to determine TP53 mutation status and allow for early identification of patients for whom standard chemotherapy is unlikely to be effective.
Appendix
FIG A1.
Correlation of variant allele frequency between whole-exome sequencing (WES) and Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets—MSK-IMPACT—in the WES cohort. The variant allele frequency of mutations identified in genes covered by the MSK-IMPACT was highly correlated between WES and MSK-IMPACT (Spearman rank correlation, ⍴ = 0.93; P < 2.2 × 10−16).
FIG A2.
Mutational heterogeneity in the Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT) cohort. Across 30 evaluable patients who were sequenced only with MSK-IMPACT, 29 (34.5%) of 84 mutations were concordant. Only eight (26.7%) of 30 patients had a majority of concordant mutations.
FIG A3.
Timing of mutations during tumor evolution in the Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets—MSK-IMPACT—cohort. (A) Significantly more mutations occurred after the observed copy number alterations (CNAs; before: median, zero mutations, range (0-3); after: median, two mutations, range (0-5); Wilcoxon signed rank test, P = 6.297 × 10−5). (B) Seven (41.2%) of 17 oncogenic mutations arose early versus four (11.1%) of 36 variants of unknown significance (VUSs; Fisher’s exact test, P = .03). (**) Statistical significance for the illustrated comparisons.
TABLE A1.
Sample Comparisons
PRIOR PRESENTATION
Presented, in part, at the 53rd Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, June 2-6, 2017.
SUPPORT
Supported by the National Institutes of Health/National Cancer Institute Cancer Center Support Grant No. P30 CA008748, the Marie-Josée and Henry R. Kravis Center for Molecular Oncology, the STARR Foundation, Cycle for Survival, and Movember GAP5.
EQUAL CONTRIBUTION
M.L.C. and M.T.A.D. contributed equally to this article.
Conflicts of Interest Statement:The authors have declared that no competing interests exist.
Conflicts of Interest Statement:Authors’ disclosures of potential conflicts of interest and contributions are found at the end of this article.
AUTHOR CONTRIBUTIONS
Conception and design: Michael L. Cheng, François Audenet, Eugene J. Pietzak, Aditya Bagrodia, Maria E. Arcila, Joel Sheinfeld, Hikmat Al-Ahmadie, David B. Solit, Darren R. Feldman
Financial support: David B. Solit, Darren R. Feldman
Administrative support: Caitlin Bourque, Lilan Ling, S. Duygu Selcuklu, David B. Solit, Darren R. Feldman
Provision of study materials or patients: Victor E. Reuter, George J. Bosl, Joel Sheinfeld, David B. Solit, Darren R. Feldman
Collection and assembly of data: Michael L. Cheng, Mark T.A. Donoghue, François Audenet, Eugene J. Pietzak, Sumit Isharwal, Gopa Iyer, Samuel Funt, Aditya Bagrodia, Victor E. Reuter, Jana Eng, Gabriella Joseph, Caitlin Bourque, Maria Bromberg, Lilan Ling, S. Duygu Selcuklu, Maria E. Arcila, Dana W.Y. Tsui, Ahmet Zehir, Agnes Viale, Hikmat Al-Ahmadie, David B. Solit, Darren R. Feldman
Data analysis and interpretation: Michael L. Cheng, Mark T.A. Donoghue, François Audenet, Nathan C. Wong, Eugene J. Pietzak, Craig M. Bielski, Samuel Funt, Dean F. Bajorin, Victor E. Reuter, Gabriella Joseph, Maria E. Arcila, Dana W.Y. Tsui, Michael F. Berger, George J. Bosl, Eliezer Van Allen, Barry S. Taylor, Hikmat Al-Ahmadie, David B. Solit, Darren R. Feldman
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.
Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).
Michael L. Cheng
Honoraria: The Lynx Group (supported by Bristol Myers Squibb), WebMD (supported by AstraZeneca), PCME (supported by Merck and Lilly)
Consulting or Advisory Role: AstraZeneca, Inivata
Travel, Accommodations, Expenses: Daiichi Sankyo, Allergan, Sanofi, Natera, AstraZeneca, Guardant Health, WebMD, PCME
François Audenet
Honoraria: Nucleix
Travel, Accommodations, Expenses: Ferring, Ipsen
Eugene J. Pietzak
Honoraria: UpToDate
Consulting or Advisory Role: Merck, Chugai Pharma
Sumit Isharwal
Open Payments Link: https://openpaymentsdata.cms.gov/physician/1232544
Gopa Iyer
Consulting or Advisory Role: Bayer, Janssen, Mirati Therapeutics
Research Funding: Mirati Therapeutics (Inst), Novartis (Inst), Debiopharm Group (Inst), Bayer (Inst)
Samuel Funt
Stock and Other Ownership Interests: Kite Pharma, Urogen Pharma (I), Allogene Therapeutics, Neogene Therapeutics (I), Kronos Bio (I), Vida Ventures (I), Vaxigene
Consulting or Advisory Role: AstraZeneca/MedImmune, Merck, Immunai
Research Funding: Genentech (Inst), AstraZeneca (Inst), Decibel Therapeutics (Inst)
Travel, Accommodations, Expenses: Bristol Myers Squibb, AstraZeneca/MedImmune
Dean F. Bajorin
Honoraria: Merck Sharp & Dohme
Consulting or Advisory Role: Merck, Fidia Farmaceutici, Hoffman-La Roche, Dragonfly Therapeutics
Research Funding: Novartis (Inst), Genentech (Inst), Merck (Inst), Bristol Myers Squibb (Inst), AstraZeneca (Inst), Astellas Pharma (Inst), Seattle Genetics/Astellas (Inst)
Travel, Accommodations, Expenses: Genentech, Merck
Victor E. Reuter
Consulting or Advisory Role: Cepheid
Uncompensated Relationships: PaigeAI
Gabriella Joseph
Stock and Other Ownership Interests: Amgen
Maria E. Arcila
Honoraria: Invivoscribe, Biocartis
Consulting or Advisory Role: AstraZeneca
Travel, Accommodations, Expenses: AstraZeneca, Invivoscribe, Raindance Technologies
Dana W.Y. Tsui
Honoraria: Cowen, BofA Merrill Lynch
Patents, Royalties, Other Intellectual Property: I am a co-inventor on a provisional patent application filed by Memorial Sloan Kettering Cancer Center
Ahmet Zehir
Honoraria: Illumina
Michael F. Berger
Consulting or Advisory Role: Roche
Research Funding: Grail
Patents, Royalties, Other Intellectual Property: Provisional patent pending for “Systems and Methods for Detecting Cancer via cfDNA Screening”
Eliezer Van Allen
Stock and Other Ownership Interests: Syapse, Tango Therapeutics, Genome Medical, Microsoft, Ervaxx
Consulting or Advisory Role: Syapse, Roche, Third Rock Ventures, Takeda, Novartis, Genome Medical, InVitae, Illumina, Tango Therapeutics, Ervaxx, Janssen
Speakers' Bureau: Illumina
Research Funding: Bristol Myers Squibb, Novartis
Patents, Royalties, Other Intellectual Property: Patent on discovery of retained intron as source of cancer neoantigens (Inst); patent on discovery of chromatin regulators as biomarkers of response to cancer immunotherapy (Inst); patent on clinical interpretation algorithms using cancer molecular data (Inst)
Travel, Accommodations, Expenses: Genentech
Barry S. Taylor
Consulting or Advisory Role: Boehringer Ingelheim, Loxo Oncology at Lilly
Research Funding: Genentech
Hikmat Al-Ahmadie
Consulting or Advisory Role: Bristol Myers Squibb, EMD Serono, AstraZeneca/MedImmune, Janssen Biotech
David B. Solit
Stock and Other Ownership Interests: Loxo
Consulting or Advisory Role: Pfizer, Loxo, Illumina, Vivideon Therapeutics, Lilly Oncology, QED Therapeutics, BridgeBio Pharma
Darren R. Feldman
Research Funding: Novartis, Seattle Genetics, Decibel Therapeutics (Inst), Astellas Pharma
Other Relationship: UpToDate
No other potential conflicts of interest were reported.
REFERENCES
- 1.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin. 2020;70:7–30. doi: 10.3322/caac.21590. [DOI] [PubMed] [Google Scholar]
- 2.Hanna NH, Einhorn LH. Testicular cancer: Discoveries and updates. N Engl J Med. 2014;371:2005–2016. doi: 10.1056/NEJMra1407550. [DOI] [PubMed] [Google Scholar]
- 3.International Germ Cell Cancer Collaborative Group International Germ Cell Consensus Classification: A prognostic factor-based staging system for metastatic germ cell cancers. J Clin Oncol. 1997;15:594–603. doi: 10.1200/JCO.1997.15.2.594. [DOI] [PubMed] [Google Scholar]
- 4.Bagrodia A, Lee BH, Lee W, et al. Genetic determinants of cisplatin resistance in patients with advanced germ cell tumors. J Clin Oncol. 2016;34:4000–4007. doi: 10.1200/JCO.2016.68.7798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Taylor-Weiner A, Zack T, O’Donnell E, et al. Genomic evolution and chemoresistance in germ-cell tumours. Nature. 2016;540:114–118. doi: 10.1038/nature20596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zehir A, Benayed R, Shah RH, et al. Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nat Med. 2017;23:703–713. doi: 10.1038/nm.4333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cheng DT, Mitchell TN, Zehir A, et al. Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT): A hybridization capture-based next-generation sequencing clinical assay for solid tumor molecular oncology. J Mol Diagn. 2015;17:251–264. doi: 10.1016/j.jmoldx.2014.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Al-Ahmadie HA, Iyer G, Lee BH, et al. Frequent somatic CDH1 loss-of-function mutations in plasmacytoid variant bladder cancer. Nat Genet. 2016;48:356–358. doi: 10.1038/ng.3503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Shen R, Seshan VE. FACETS: Allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing. Nucleic Acids Res. 2016;44:e131–e131. doi: 10.1093/nar/gkw520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Dentro SC, Wedge DC, Van Loo P. Principles of reconstructing the subclonal architecture of cancers. Cold Spring Harb Perspect Med. 2017;7:a026625. doi: 10.1101/cshperspect.a026625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bielski CM, Zehir A, Penson AV, et al. Genome doubling shapes the evolution and prognosis of advanced cancers. Nat Genet. 2018;50:1189–1195. doi: 10.1038/s41588-018-0165-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chakravarty D, Gao J, Phillips SM, et al. OncoKB: A precision oncology knowledge base. JCO Precis Oncol. doi: 10.1200/PO.17.00011. 10.1200/PO.17.00011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Loveday C, Litchfield K, Proszek PZ, et al. Genomic landscape of platinum resistant and sensitive testicular cancers. Nat Commun. 2020;11:2189. doi: 10.1038/s41467-020-15768-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kanu N, Grönroos E, Martinez P, et al. SETD2 loss-of-function promotes renal cancer branched evolution through replication stress and impaired DNA repair. Oncogene. 2015;34:5699–5708. doi: 10.1038/onc.2015.24. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The assembled prospective somatic mutational data for the entire cohort have been deposited for visualization and download in the cBioPortal for Cancer Genomics. WES data can be obtained from dbGaP (accession number: phs002229.v1.p1).