Concordance of DNA Repair Gene Mutations in Paired Primary Prostate Cancer Samples and Metastatic Tissue or Cell-Free DNA

Key Points

Question  Can the mutational status of DNA repair genes in advanced prostate cancer be accurately assessed by sequencing primary prostate specimens?

Findings  In this genetic association study, which included primary samples with paired cell-free circulating tumor DNA and/or metastatic tissue from 51 men from 3 cohorts, gene alterations in DNA repair genes detected in cell-free circulating tumor DNA or metastatic tissue were concordant with primary prostate cancer when clonal hematopoiesis was excluded.

Meaning  Sequencing archival, primary prostate cancer tissue may be an accurate means to assess the status of actionable DNA repair gene alterations in men with metastatic disease.

Importance  DNA damage repair (DDR) gene mutations represent actionable alterations that can guide precision medicine strategies for advanced prostate cancer. However, acquisition of contemporary tissue samples for molecular testing can be a barrier to deploying precision medicine approaches. We hypothesized that most DDR alterations represent truncal events in prostate cancer and that primary tissue would faithfully reflect mutations found in cell-free circulating tumor DNA (ctDNA) and/or metastatic tissue.

Objective  To assess concordance in DDR gene alterations between primary prostate cancer and metastases or ctDNA specimens.

Design, Setting, and Participants  Patients were included if a DDR pathway mutation was detected in metastatic tissue or ctDNA and primary tissue sequencing was available for comparison. Sequencing data from 3 cohorts were analyzed: (1) FoundationOne, (2) University of Washington clinical cases (University of Washington–OncoPlex or Stand Up to Cancer–Prostate Cancer Foundation International Dream Team sequencing pipelines), and (3) University of Washington rapid autopsy series. Only pathogenic somatic mutations were included, and more than 30 days between primary tumor tissue and ctDNA and/or metastatic tissue acquisition was required. Clonal hematopoiesis of indeterminate potential (CHIP) and germline events were adjudicated by an expert molecular pathologist and excluded.

Main Outcomes and Measures  The DDR gene alterations detected in primary prostate tissue matched with metastatic tissue and/or ctDNA findings.

Results  A total of 72 men with known DDR alterations were included in the analysis, and primary samples with paired ctDNA and/or metastatic tissue were sequenced. After excluding patients with ctDNA where only CHIP and/or germline events (n = 21) were observed, 51 patients remained and were included in the final analysis. The median (range) time from acquisition of primary tissue to acquisition of ctDNA or tumor tissue was 55 (5-193) months. Concordance in DDR gene mutation status across samples was 84% (95% CI, 71%-92%). Rates of concordance between metastatic-primary and ctDNA-primary pairs were similar when patients with CHIP events were excluded. Multiclonal BRCA2 reversion mutations associated with resistance to PARP inhibitors and platinum chemotherapy were detected in ctDNA from 2 patients.

Conclusions and Relevance  In this genetic association study of 3 patient cohorts, primary prostate tissue accurately reflected the mutational status of actionable DDR genes in metastatic tissue, consistent with DDR alterations being truncal in most patients. After excluding likely CHIP events, ctDNA profiling accurately captured these DDR mutations while also detecting reversion alterations that may suggest resistance mechanisms.

Inactivating alterations in genes involved in DNA damage repair (DDR) occur in approximately 25% of patients with metastatic castration-resistant prostate cancer (mCRPC).¹ Many of these alterations may be associated with response to DNA-damaging therapies or immune checkpoint inhibitors and can be used to guide precision medicine strategies in advanced prostate cancer.^2-5 DNA-damaging therapeutics and drugs that impair aspects of DNA repair (eg, PARP inhibitors) are active in many patients with mCRPC with alterations in homologous recombination repair genes.^4-7, Likewise, immune checkpoint inhibitors are active in tumors with mismatch repair deficiency or microsatellite instability.^2,3

Acquisition of contemporary tissue for advanced molecular testing can be a barrier to deploying precision medicine in men with mCRPC, and archival primary tissue represents a convenient biospecimen for next-generation sequencing. Metastatic biopsies are morbid and costly, and low tumor content from bone biopsies often leads to indeterminate studies.⁸ Cell-free circulating tumor DNA (ctDNA) is an alternative to metastatic biopsy; however, technical issues can lead to missing focal somatic copy number aberrations, and sequencing is often unsuccessful except during progression. Furthermore, ctDNA sequencing may detect misleading clonal hematopoiesis of indeterminate potential (CHIP) events—somatic alterations in hematopoietic clones unrelated to prostate cancer.^1,9 We hypothesized that DDR gene alterations represent early truncal events in prostate cancer and that archival primary tissue would faithfully reflect the key DDR mutations present in metastases.

Patients were included in the analysis if a DDR pathway gene mutation was detected in metastatic tissue or ctDNA and primary tissue sequencing was available for comparison. Sequencing data from 3 cohorts were analyzed: (1) FoundationOne (FoundationOne CDx for tissue and FoundationOne Liquid CDx [Foundation Medicine] were used), (2) University of Washington (UW) clinical cases, and (3) UW rapid autopsy series. Rapid autopsy cases were sequenced as previously described,¹⁰ and patients with an intact primary were included. The UW clinical cases were sequenced on UW-OncoPlex, a clinical-grade targeted sequencing platform, or as part of the Stand Up to Cancer–Prostate Cancer Foundation International Dream Team pipeline as previously described.^1,11 FoundationOne samples were sequenced as part of their clinical pipeline. This study was approved by the UW–Fred Hutchinson Cancer Research Center Institutional Review Board or the Western Institutional Review Board. This research was exempt from patient informed consent requirements because it was deemed a minimal risk to participants.

Cases were included if 1 or more pathogenic somatic DDR gene alteration was detected. We also required more than 30 days between primary tumor tissue and ctDNA or metastatic tissue acquisition. For cases without matched nontumor DNA, alterations were assessed as likely germline by cross-referencing against the ClinVar database and accounting for variant allele fraction in the context of tumor content, ploidy, and loss of heterozygosity. Given that CHIP events are commonly detected in plasma from patients with advanced prostate cancer, we excluded alterations felt to have a high probability of representing CHIP.⁹ Variants detected only in plasma were considered likely to be CHIP or low subclones if the variant allele fraction was less than 1% and/or more than 5-fold less than the estimated tumor content in plasma. Cases were considered concordant if they shared at least 1 somatic alteration in the same gene. Alterations that commonly arise owing to selective pressure (eg, BRCA2 reversion mutations) were considered concordant if a somatic alteration affecting the same gene was detected in primary prostate tissue. Stata, version 15.1 (StataCorp), was used for analysis, and a 2-sided P < .05 was considered statistically significant.

A total of 72 patients with DDR gene alterations had paired primary and metastatic tissue or ctDNA sequencing data available. After excluding patients with ctDNA where only CHIP (n = 12), germline (n = 8) or germline/CHIP (n = 1) events were identified (Figure 1), 51 patients remained and were included in the final analysis. Two patients had primary, metastatic, and ctDNA samples sequenced, which allowed for comparison of concordance between 53 sample pairs in total. Across all samples sequenced, a total of 61 DDR gene alterations were identified.

Excluding rapid autopsy cases, the median (range) time from acquisition of prostate tissue to ctDNA or metastatic tissue collection was 55 (5-193) months. Of the 53 paired samples, at least partial concordance in DDR genes was identified in 43 cases (84%; 95% CI, 71%-92%) (Figure 2 and eTable in the Supplement). Concordance was numerically higher between ctDNA primary pairs compared with metastatic primary pairs; however, this difference was not statistically significant (92% vs 79%; Fisher exact P = .27). Concordance on a gene-by-gene basis was 79% (95% CI, 66%-87%) (Table). The majority of discordant mutations were present exclusively in metastatic tissue or ctDNA; however, there were 2 patients (FM15 and FM22) who had monoallelic DDR gene alterations only found in primary tissue.

There was evidence of clonal evolution in both patients (UW8 and UW9) with primary, metastatic, and ctDNA sequencing performed, and both received carboplatin-based chemotherapy. Patient FM12 had a PALB2 reversion, and patients FM27 and FM32 had BRCA2 reversions detected in downstream samples that were not present in primary tissue. Patient FM27 received olaparib prior to acquisition of metastatic tissue. Clinical details on the other 2 patients were not available.

These data provide evidence that primary prostate tissue accurately reflects the key targetable alterations in DDR genes found in men with metastatic prostate cancer, supporting the hypothesis that DDR alterations are early truncal events. While there were a few notable differences between primary and metastatic tissue or ctDNA, nearly all patients included in this study would have been deemed eligible for precision medicine strategies had only primary tissue been sequenced. It is notable that these findings are consistent with an earlier report showing primary metastatic tissue concordance in DDR gene alterations across 9 patients with paired tissue sequencing.¹²

As expected, most discordant cases developed mutations in downstream metastatic or ctDNA samples. When examining these cases more closely, we observed that acquired mutations were predominately in genes with an unclear role in predicting response to DNA damaging therapy (eg, RAD51 and ATR), with only 1 patient developing a BRCA2 alteration at a later time point that did not have other identified DDR alterations.^5,13,14 While the obvious explanation for these discordant findings is that they evolved as a consequence of selective therapeutic pressure, it is worthwhile to consider other causes. One possibility is that these findings may relate to the genomic heterogeneity between tumor foci within the prostate. If the prostate tumor selected for sequencing was not representative of the clone that ultimately metastasized, we would anticipate discordance with the mutational profile observed at the time of metastasis. Indeed, the clonal origin of lethal prostate cancer has been documented to arise from seemingly low-grade primary tumors.¹⁵

Surprisingly, we identified 2 cases where a BRCA1/2 mutation was present in the primary sample but absent in downstream samples. It is possible that these cases also relate to issues of intraprostatic genomic heterogeneity; however, it is also plausible that DNA damaging therapies may have eradicated clones that were sensitive to these therapies. Unfortunately, clinical details on these cases are unknown, and the context in which these mutations were lost is unknown. We did observe a shifting molecular landscape in several cases with clear evidence of reversion mutations predicted to restore BRCA2 or PALB2 function. These cases highlight the power of serial ctDNA sequencing to gain insights into mechanisms of resistance and clonal evolution.

Key limitations of this research include its retrospective nature and the lack of clinical details available for patients included. In addition, the genetic heterogeneity between prostate tumor foci was not accounted for when assessing mutational concordance between samples.

Prospective studies evaluating the mutational concordance between metastatic and primary prostate specimens are needed. However, data in this genetic association study support the truncal nature of most DDR alterations and indicate that archival primary prostate tissue can provide valuable insights regarding the mutational status of DDR genes.

Accepted for Publication: May 3, 2021.

Published Online: June 4, 2021. doi:10.1001/jamaoncol.2021.2350

Corresponding Authors: Michael T. Schweizer, MD, Division of Oncology, Department of Medicine, University of Washington, 1144 Eastlake Ave E, MS: LG-454, Seattle, WA 98109 (schweize@uw.edu); Bruce Montgomery, MD, Prostate Cancer Foundation Precision Oncology Program for Cancer of the Prostate, VA Puget Sound Health Care System, 1660 S Columbian Way (111ONC), Seattle, WA 98108 (rbmontgo@uw.edu).

Author Contributions: Dr Schweizer had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Schweizer, Konnick, Nelson, Pritchard, Montgomery.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Schweizer, Coleman, Konnick, Nelson, Pritchard, Montgomery.

Critical revision of the manuscript for important intellectual content: Schweizer, Sivakumar, Tukachinsky, De Sarkar, Yu, Konnick, Nelson, Pritchard, Montgomery.

Statistical analysis: Schweizer, Montgomery.

Obtained funding: Schweizer, Nelson, Pritchard, Montgomery.

Administrative, technical, or material support: Schweizer, Sivakumar, Yu, Konnick, Nelson, Pritchard, Montgomery.

Supervision: Schweizer, Pritchard, Montgomery.

Conflict of Interest Disclosures: Dr Schweizer reported personal fees from Janssen and Resverlogix, as well as grants from Zenith Epigenetics, Bristol Myers Squibb, Merck, Immunomedics, Janssen, AstraZeneca, Pfizer, Madison Vaccines, Tmunity, and Hoffmann-La Roche outside the submitted work. Dr Sivakumar reported personal fees from Foundation Medicine during the conduct of the study. Dr Tukachinsky reported personal fees from Foundation Medicine during the conduct of the study. Dr Yu reported personal fees from Janssen, Bayer, Merck, Clovis, Advanced Accelerator Applications, Exelixis, AbbVie, Sanofi, AstraZeneca, Amgen, Pharmacyclics, Dendreon, and Seattle Genetics, as well as grants from Daiichi Sankyo, Taiho, Dendreon, Merck, Seattle Genetics, Blue Earth, and Bayer outside the submitted work. Dr Konnick reported personal fees from Ventana Medical Systems, Roche, River West Meeting Associates, National Comprehensive Cancer Network, Medscape, and Clinical Care Options outside the submitted work. Dr Nelson reported grants from the National Institutes of Health, Congressionally Directed Medical Research Programs, and the Prostate Cancer Foundation during the conduct of the study, as well as serving on advisory boards for Janssen, Astellas, and Bristol Myers Squibb for work unrelated to the study. Dr Pritchard reported personal fees from AstraZeneca during the conduct of the study. No other disclosures were reported.

Funding/Support: This research was supported by the University of Washington–Fred Hutchinson Cancer Research Center’s Institute for Prostate Cancer Research, the Pacific Northwest Prostate Cancer Specialized Program of Research Excellence (CA097186), a Prostate Cancer Foundation’s Young Investigator Award (Drs Schweizer and De Sarkar), a Prostate Cancer Foundation Challenge Award (Drs Montgomery and Nelson); Congressionally Directed Medical Research Program Awards (W81XWH-16-1-0484 [Dr Schweizer], W81XWH-18-1-0355 [Drs Nelson, Pritchard, and Montgomery], W81XWH-17-1-0380 [Dr De Sarkar], W81XWH-18-1-0756 [PC170510, Dr Pritchard], W81XWH-18-1-0356 [PC170503P2 and PC200262P1, Dr Pritchard], and W81XWH-18-1-0406 [Dr Nelson]), the Brotman Baty Institute for Precision Medicine (Drs Pritchard and De Sarkar), and a Stand Up To Cancer–Prostate Cancer Foundation Prostate Dream Team Translational Cancer Research Grant (SU2C-AACR-DT0712).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Meeting Presentation: This study was presented at the virtual annual meeting of the American Society of Clinical Oncology; June 4, 2021.