Pathogenic Variants in Cancer Predisposition Genes and Prostate Cancer Risk in Men of African Ancestry

Abstract

PURPOSE

In studies of men of European ancestry, rare pathogenic variants in DNA repair pathway genes have been shown to be associated with risk of aggressive prostate cancer. The contribution of rare coding variation to prostate cancer risk in men of African ancestry has not been established.

METHODS

We sequenced a panel of 19 DNA repair and cancer predisposition genes in 2,453 African American and 1,151 Ugandan cases and controls with prostate cancer. Rare variants were classified as pathogenic or putatively functionally disruptive and examined in association with prostate cancer risk and disease aggressiveness in gene and pathway-level association analyses.

RESULTS

Pathogenic variants were found in 75 of 2,098 cases (3.6%) and 31 of 1,481 controls (2.1%; odds ratio [OR], 1.82; 95% CI, 1.19 to 2.79; P = .0044), with the association being stronger for more aggressive disease phenotypes (OR, 3.10; 95% CI, 1.54 to 6.23; P = .0022). The highest risks for aggressive disease were observed with pathogenic variants in the ATM, BRCA2, PALB2, and NBN genes, with ORs ranging from approximately 4 to 15 in the combined study sample of African American and Ugandan men. Rare, nonpathogenic, nonsynonymous variants did not have a major impact on risk of overall prostate cancer or disease aggressiveness.

CONCLUSION

Rare pathogenic variants in DNA repair genes have appreciable effects on risk of aggressive prostate cancer in men of African ancestry. These findings have potential implications for panel testing and risk stratification in this high-risk population.

INTRODUCTION

African American men have a 70% higher incidence rate and a 130% higher mortality rate of prostate cancer (PCa) than white men.¹ Incidence rates in Uganda are among the highest of all African countries,² and men in sub-Saharan Africa have been reported to have the highest mortality rates compared with men of other races or ethnicities,³ which is likely a consequence of the high percentage of men with aggressive disease at diagnosis.⁴ Inherited susceptibility is an important contributor to PCa risk,⁵ with studies identifying 175 common variants that explain approximately 37% of familial PCa risk.^6-8 There is growing support for the role of rare or low-frequency coding variation in PCa risk from studies in men of European ancestry.⁹ In HOXB13, a missense variant has been associated with risk, with effects being greater among men with early-onset and familial PCa.¹⁰ Studies in whites have shown pathogenic variants in BRCA1 and BRCA2 to confer increased risk and to be more strongly associated with aggressive disease.¹¹ Studies in whites have also provided strong evidence supporting the role of pathogenic variants in other DNA repair genes (eg, ATM, CHEK2, PALB2, RAD51D) in aggressive PCa.^12,13

The PCa relevance of rare variants in DNA repair pathway genes has not been established for men of African ancestry. In this study, we sequenced 19 DNA repair and known cancer predisposition genes in 2,121 cases and 1,483 controls to investigate the aggregate contribution of rare pathogenic and nonpathogenic coding variation to risk of PCa and aggressive disease in African American and Ugandan men.

METHODS

Study Subjects

The men in this study are from the Multiethnic Cohort (MEC), the Los Angeles Study of Aggressive Prostate Cancer (LAAPC), and the Uganda Prostate Cancer Study (UGPCS). Men with higher Gleason and stage from the following studies were overselected to increase power to identify rare variants associated with more clinically aggressive disease. Participants all signed a written consent, and the study protocol was approved by the Institutional Review Boards of the University of Southern California and Makerere University. The human investigations were performed after approval by a local Human Investigations Committee and in accordance with an assurance filed with and approved by the Department of Health and Human Services, where appropriate.

Context

• Key Objective

• To our knowledge, this is the first study aimed to investigate the contribution of rare pathogenic and nonpathogenic coding variation in DNA repair genes to risk of prostate cancer and aggressive disease specifically in men of African ancestry.

• Knowledge Generated

• Rare pathogenic variants in BRCA2, ATM, PALB2, and NBN were significantly associated with a high risk of aggressive disease. In addition, we found a substantial number of pathogenic variants that had not been previously reported in major public databases, which highlights the importance of sequencing efforts in genetically diverse populations.

• Relevance

• These findings underscore the clinical relevance of rare pathogenic variants in DNA repair genes in men of African ancestry, which could be used to build better predictive and prognostic gene-based tests to identify men at higher risk of aggressive prostate cancer in this population.

Multiethnic Cohort.

MEC is a prospective cohort study of men and women recruited from Hawaii and California between 1993 and 1996.¹⁴ Incident patients with PCa were identified through linkage to cancer SEER registries in Hawaii and California. Information on histologic status of disease (eg, Gleason score, stage) was also obtained from the SEER registries. The MEC includes > 1,200 African American patients with PCa with biospecimens. In the current study, 760 African American PCa cases, overselected for high stage and Gleason score, and 996 controls were included. Also included were 620 African American PCa cases recruited in Los Angeles County between 2007 and 2015.

Los Angeles Study of Aggressive Prostate Cancer.

In LAAPC, incident African American PCa cases were identified through the Los Angeles County SEER cancer registry between 1999 and 2003.¹⁵ A total of 458 African American PCa cases were recruited, of whom 77 with nonlocalized disease or Gleason score > 7 were included in the current study.

Uganda Prostate Cancer Study.

UGPCS is a case-control study of PCa in Uganda that began in 2010. As of July 2016, 664 incident PCa cases and 487 controls have been recruited.¹⁶ Cases were enrolled from 13 hospitals/clinics across the country, whereas controls were enrolled from nonurologic clinics (eg, surgery). All controls had a prostate-specific antigen (PSA) level < 4 ng/mL, and a saliva spit kit was used to collect a biospecimen at recruitment. Gleason score was available for 441 cases (68%), and PSA level at diagnosis was available for 434 (67%) cases. Information regarding family history of PCa was obtained by questionnaire at cohort entry in the MEC and at case-control enrollment in LAAPC and UGPCS.

Disease Aggressiveness

Aggressive and nonaggressive PCa was defined based on stage, Gleason score, death from PCa, and PSA level at diagnosis. African American men with high-risk PCa (n = 645; 44.6%) were defined as those with nonlocalized disease or Gleason score > 7 or those who died as a result of PCa. African American men with low-/intermediate-risk PCa (n = 802; 55.4%) were those with localized disease and Gleason score ≤ 7. In UGPCS, high-risk PCa (n = 414; 63.6%) was defined as Gleason score > 7 or PSA level > 50 ng/mL, whereas low-/intermediate-risk PCa (n = 53; 8.1%) was defined as Gleason score ≤ 7 and PSA level ≤ 50 ng/mL. The 50-ng/mL cut point for PSA level was chosen based on the advanced disease state at which most Ugandan patients are diagnosed because we had no patients with a PSA level of ≤ 20 ng/mL in our study. Because stage information was not available in UGPCS, metastatic disease was defined as a PSA level > 100 ng/mL.

Targeted Gene Sequencing

Genomic DNA libraries were prepared using a custom QIAseq amplicon-based targeted panel (Qiagen, Germantown, MD)¹⁷ of coding regions and essential splice sites for 19 cancer predisposition genes with a core function in DNA damage repair: ATM, BARD1, BRCA1, BRCA2, BRIP1, CHEK2, MLH1, MRE11A, MSH2, MSH6, NBN, PALB2, PMS2, PTEN, RAD50, RAD51C, RAD51D, TP53, and XRCC2.^12,13 The panel was designed to capture all exonic regions of the investigated genes.

Of the 3,604 samples, 3,588 were successfully sequenced with 99.6% of target bases covered ≥ 20×, whereas 25 samples were removed because of insufficient coverage (< 90% of bases covered ≥ 20×), high degree of relationship with other samples (relatedness score > 0.5), or excessive homozygosity. A total of 2,098 cases and 1,481 controls were retained for analysis.

Variant Annotation

CAVA (Clinical Annotation of Variants) v.1.1.1 was used to annotate variants according to functional effect.¹⁸ Given the focus of the study on variants that might alter protein structure and function, variants in introns (n = 215), 5′-UTR (n = 15), 3′-UTR (n = 7), in-frame insertions/deletions (n = 14), synonymous variants (n = 537), and those located > 2 bp outside of the consensus splice site (n = 147) according to CAVA were excluded. We also removed 148 variants having frequency ≥ 1% in controls or in gnomAD. The remaining variants (n = 1,163) with frequency < 1% were sorted into 2 groups based on predicted functional and pathogenic annotation (Data Supplement). Variants predicted to result in protein truncation or to significantly alter the protein sequence (frameshift insertion/deletions, gain of stop codon or loss of essential splice site donor/acceptor) and missense variants (nonsynonymous codon change, exon start/end codon change) that were reported as pathogenic or likely pathogenic in ClinVar¹⁹ by 1 or more clinical laboratories (Ambry, SCRP, InVitae, GeneDX, Emory, and InSiGHT) were classified as pathogenic (tier 1; n = 91). Missense variants not identified as being pathogenic or likely pathogenic in ClinVar were annotated with dbNSFP version 3.3a²⁰ using 5 in silico algorithms (Polyphen2-HumDiv, PolyPhen2-HumVar, LRT, Mutation Taster, and SIFT) to predict potential functional effects.²¹ Tier 2 variants (n = 566) were those predicted by ≥ 1 of the 5 dbNSFP algorithms to be putatively functionally altered, which included a small number of missense variants (n = 10) in tier 1. We subsequently annotated variants at the gene set level, with genes grouped by the DNA repair process.²²

Statistical Analysis

Principal components (PCs) were calculated using 311 weakly correlated (mean r², 0.01) common single nucleotide polymorphisms (frequency ≥ 1% and call rate ≥ 0.98) using EIGENSTRAT.²³ Tier 1 and tier 2 variants were evaluated in gene burden testing, with the minor alleles summed at the gene level in each individual and counts compared between cases and controls. In addition, we assessed the association across all genes (polygenic burden test) and gene sets grouped by DNA repair process (pathway-level burden test). We report odds ratios (ORs) and 95% CIs from logistic regression and P values from a likelihood ratio test, adjusting for age, ethnicity (Ugandan v African American), and the first 2 PCs, primarily capturing variation in African ancestry. Statistical significance was set at P < .05 in 2-sided tests, because many of these genes have been associated with PCa risk in men of European ancestry.^12,13

RESULTS

The mean age at diagnosis of PCa was 66.71 years for African American cases and 70.77 for Ugandan cases (Table 1). Based on the overselection of cases with high stage and high Gleason score at diagnosis (see Methods), 44.6% of African American cases and 63.6% of Ugandan cases were classified as having high-risk PCa.

TABLE 1.

Descriptive Characteristics of African American and Ugandan Prostate Cancer Cases and Controls

In total, 647 rare (< 1%) putatively disruptive variants (tier 1 plus tier 2; see Methods) were identified in the 19 genes (Data Supplement). Variants were annotated as nonsynonymous (84.4%), frameshift insertions/deletions (6.5%), stop-gain (4.5%), exon start/end codon changes (3.1%), and essential splice sites (1.5%; Table 2). We identified 111 variants (41 in African Americans and 71 in Ugandans), all with frequency ≤ 0.1%, not previously reported in gnomAD or ClinVar (Table 2). Sixty-four variants (9.9%) were shared across the 2 groups, whereas 419 (64.8%) and 164 (25.3%) variants were observed exclusively in African Americans and Ugandans, respectively.

TABLE 2.

Annotation of 647 Rare Putatively Disruptive Variants in 3,579 Men of African Ancestry

We identified 91 tier 1 pathogenic variants in 16 of the 19 genes. Ten variants in 7 genes (ATM, BRCA2, CHEK2, MLH1, MSH6, PMS2, NBN) were observed in more than 1 sample (Data Supplement), with the most common being c.1100delC (rs555607708) in CHEK2 that was present in 6 African American men (OR, 0.16; 95% CI, 0.02 to 1.51; P = .066). BRCA2 had the largest number of pathogenic variants (n = 23; 25.3% of tier 1) followed by ATM (n = 16; 17.6%), with the remaining genes having ≤ 8 variants each (Table 3).

TABLE 3.

Gene Burden Results for 91 Tier 1 Variants and Prostate Cancer Risk

In gene burden testing, pathogenic variants in BRCA2 (OR, 3.92; 95% CI, 1.34 to 11.47; P = .0045), ATM (OR, 3.83; 95% CI, 1.09 to 13.41; P = .018), and PALB2 (4 cases and 0 controls; P = .034) were significantly associated with risk, whereas an association with NBN was borderline significant (OR, 3.50; 95% CI, 0.75 to 16.35; P = .074; Table 3; Fig 1). Positive associations with these genes were observed in both African Americans and Ugandans, although the associations were only statistically significant in Ugandans. Variants in CHEK2, which included c.1100delC, were under-represented in cases versus controls (OR, 0.47; 95% CI, 0.15 to 1.45; P = .18). However, exclusion of c.1100delC altered the risk estimate for CHEK2 toward the null (OR, 0.99; 95% CI, 0.22 to 4.48; P = .99; data not shown).

FIG 1.

Tier 1 variants in genes significantly associated with prostate cancer risk. Each variant (DNA position and resulting amino acid change) in these genes and protein domains is presented by lollipop plots, with the variant type indicated by color. Variants predicted to result in protein truncation or significantly alter the protein sequence (frameshift insertion/deletions, stop codon gained or essential splice site donor/acceptor) and missense variants (nonsynonymous coding, exon start/end codon change) that were reported as pathogenic or likely pathogenic in ClinVar are coded in violet and green, respectively. On the graph of each gene, the x-axis represents the number of amino acid residues, and the y-axis represents the total number of variants identified. Protein domains are also distinguished by color. FAT domain, focal adhesion targeting domain; FATC domain, focal adhesion targeting carboxyterminal domain; FHA, forkhead-associated domain; Nbs1_C, Nijmegen breakage syndrome C-terminal domain.

The positive associations observed with pathogenic variants were limited primarily to high-risk disease (BRCA2: OR, 5.29; 95% CI, 1.74 to 16.06; P = .0014; ATM: OR, 5.15; 95% CI, 1.40 to 18.93; P = .0067; PALB2: 3 cases and 0 controls, P = .021) and Gleason score > 7 tumors (BRCA2: OR, 7.70; 95% CI, 2.43 to 24.43; P = .00017; ATM: OR, 7.11; 95% CI, 1.81 to 27.94; P = .0029; PALB2: 2 cases and 0 controls, P = .025; Data Supplement; Fig 2a). In African Americans, the association with BRCA2 was statistically significant for Gleason score > 7 (OR, 5.23; 95% CI, 1.23 to 22.30; P = .021; Data Supplement; Fig 2b). In Ugandans, BRCA2 and ATM were significantly associated with Gleason score > 7 (BRCA2: OR, 11.99; 95% CI, 1.35 to 106.71; P = .0078; ATM: 3 cases and 0 controls; P = .015) and with metastatic disease defined by PSA level > 100 ng/mL at diagnosis (BRCA2: OR, 14.94; 95% CI, 1.69 to 132.44; P = .0027; ATM: 4 cases and 0 controls; P = .007; Data Supplement; Fig 2c). Other genes were not strongly associated with aggressive phenotypes, except for a significantly higher number of CHEK2 variants in controls compared with cases with high Gleason score (0 cases and 8 controls, P = .036). The risk estimates for BRCA2 and ATM remained statistically significant when including cases with a Gleason score of 7 in the high-risk category (data not shown).

FIG 2.

Gene burden associations of tier 1 variants with prostate cancer risk by case-control status and strata of disease phenotype. Histograms are used to represent associations of rare pathogenic variants in ATM, BRCA2, PALB2, and NBN genes with prostate cancer risk and aggressiveness in (A) African Americans and Ugandans, (B) African Americans only, and (C) Ugandans only. In each plot, the x-axis reflects the genes and counts of individuals by subgroup, and the y-axis reflects the burden frequency for each subgroup. Numbers above the bars represent risk estimates for cases and aggressive phenotypes compared with controls, with stars indicating the strength of association (P value). Case-control status and strata of disease phenotype are distinguished by color. (*) .05 < P ≤ .1; (**) .01 < P ≤ .05; (***) .001 < P ≤ .01; (****) P ≤ .001.

In aggregating all pathogenic variants, 75 of 2,098 cases (3.6%) and 31 of 1,481 controls (2.1%) were carriers (OR, 1.82; 95% CI, 1.19 to 2.79; P = .0044; Table 4; Data Supplement), with the risk being higher for high-risk disease (OR, 2.27; 95% CI, 1.43 to 3.61; P = .00046) and metastatic PCa (OR, 3.10; 95% CI, 1.54 to 6.23; P = .0022). Associations were even larger when restricting the analyses to ATM+BRCA2+PALB2+NBN or the 2 most significantly associated genes, BRCA2 and ATM, with ORs of approximately 4 for overall PCa and 5.3 to 9.2 for aggressive disease phenotypes, and with the higher risks observed in Ugandans (ORs, 9.1 to 22.2; Table 4; Data Supplement). The aggregation of pathogenic variants by pathway showed higher risk with genes linked with Fanconi anemia (OR, 3.92; 95% CI, 1.50 to 10.26; P = .0015) and homologous recombination (OR, 2.02; 95% CI, 1.25 to 3.26; P = .0026; Data Supplement). However, these associations were no longer significant (P > .05) after removing BRCA2 (Data Supplement), except for a borderline increased risk of high-risk disease (OR, 1.77; 95% CI, 1.05 to 2.99; P = .031; Data Supplement).

TABLE 4.

Polygenic Association Analysis of 91 Tier 1 Variants and Prostate Cancer Risk by Strata of Disease Phenotype

Of the 566 tier 2 nonsynonymous variants (see Methods) identified, 360 (63.6%) were unique to 1 sample. As for the tier 1 pathogenic variants, 2 of the largest genes harbored the largest number of tier 2 variants: ATM (n = 78; 13.8%) and BRCA2 (n = 68; 12.0%; Data Supplement).

In gene burden testing, tier 2 variants in NBN were inversely associated with overall risk (OR, 0.61; 95% CI, 0.39 to 0.94; P = .027; Data Supplement) and with aggressive phenotypes (Gleason score > 7: OR, 0.37; 95% CI, 0.15 to 0.87; P = .011; high risk: OR, 0.54; 95% CI, 0.30 to 0.95; P = .026; Data Supplement). Variants in PTEN were only found in African American cases (4 v 0 controls; P = .041), whereas ATM variants were associated with risk in Ugandans (OR, 1.52; P = .032), but not in African Americans (OR, 0.91; P = .52). Suggestive associations in subgroup analyses were also observed, including with XRCC2 with Gleason score > 7 in African Americans (OR, 3.63; 95% CI, 1.52 to 8.65; P = .0042; Data Supplement).

Restricting tier 2 variants to those reported as pathogenic or likely pathogenic in ClinVar or predicted to be functionally relevant by all 5 algorithms (n = 220; see Methods) did not affect the results; a positive association with MLH1 was noted in African Americans (OR, 6.68; 95% CI, 0.85 to 52.80; P = .023; Data Supplement). There was limited evidence of association with tier 2 variants across the full set of genes and gene sets grouped by DNA repair process (Data Supplement).

DISCUSSION

In men of African ancestry, rare pathogenic variants in BRCA2, ATM, PALB2, and NBN were significantly associated with a high risk of aggressive PCa, with effect sizes ranging from approximately 4 to 15 in the combined study sample of African American and Ugandan men. Sequencing at high coverage in men of African ancestry revealed a substantial number of pathogenic variants that had not been previously reported in major public databases (18.5% with frequency ≤ 0.1%), highlighting the importance of risk allele discovery efforts in genetically diverse populations.

Overall, 5.7% of men of African ancestry defined as having metastatic PCa were carriers of pathogenic variants in the 19 DNA repair genes analyzed, which is lower than the 11.8% prevalence observed among white men with metastatic disease sequenced using a similar 20-gene panel.¹³ The strongest evidence of association was observed with BRCA2, which is consistent with previous studies in men of European ancestry that found BRCA2 germline variants to be associated with aggressive PCa.^13,24 However, we observed a lower frequency of BRCA2 variants in metastatic cases (2.1%) compared with those reported in white men (5.3%).¹³ Our findings are also in agreement with a prior study in men of European ancestry that identified deleterious variants in ATM associated with aggressive PCa.^13,25 More comparable frequencies were found for metastatic cases with ATM variants (1.8%) compared with white men (1.6%).¹³ We also observed an enrichment of pathogenic variants in PALB2 and NBN among cases versus controls, which also supports a number of previous studies in men of European ancestry.^13,25,26 However, overall pathogenic variants in ATM, BRCA2, PALB2, and NBN genes collectively accounted for only 3.9% of metastatic cases in our study versus 7.7% of metastatic cases in white men.¹³

Although pathogenic variants were associated with risks of aggressive disease in both African American and Ugandan men, differences in effect size were noted. For African American men, 3.4% of cases with Gleason score > 7 (v 1.0% with Gleason score ≤ 7) carried pathogenic variants in these genes. These percentages were greater for Ugandan men, where 5.2% of cases with Gleason score > 7 (v 1.3% with Gleason score ≤ 7) and 4.8% of cases with PSA level > 100 ng/mL (v 1.6% with PSA level < 50 ng/mL) were pathogenic variant carriers. These population differences in rare variation frequency and the larger effect size in gene burden testing observed in Ugandans reflects the higher proportion of metastatic disease at diagnosis and PSA screening used in the selection of controls in the Ugandan study (see Methods).

To our knowledge, this study represents the first large investigation aimed at identifying and testing rare pathogenic variants in DNA repair genes in association with PCa risk for men of African ancestry. This empirical evidence highlights the clinical relevance of pathogenic variants in some of the genes previously associated with aggressive disease in white men as major risk factors for aggressive PCa in men of African ancestry. Thus, guidelines regarding gene panel testing for men at high risk for metastatic disease that have been based primarily on studies among white men are appropriate to consider for men of African ancestry. Additional large studies of aggressive PCa phenotypes and more extensive gene panels will help to build better predictive and prognostic gene-based tests of aggressive PCa and to better guide patient management for men of African ancestry.

The study has a few limitations, including the small number of clinically defined cases with metastatic PCa and the lack of complete clinical data for cases (eg, stage for Ugandans). Grade migration over time may have resulted in misclassification of men with lower Gleason scores being categorized into higher Gleason score categories. If this is the case, the associations between pathogenic variants and aggressive PCa would likely be biased toward the null, with true effect sizes being larger. However, the positive associations observed in aggressive disease groups defined by Gleason score, stage, and PSA suggest that grade migration did not bias the results. Another limitation of this study is that we did not investigate large indels that could result in loss of function of the entire protein.²⁷ Whole-genome sequencing will be needed to explore a more complete spectrum of pathogenic variants, including noncoding variation and large indels, in PCa in men of African ancestry.

In conclusion, this study provides strong support for the clinical importance of rare pathogenic variants in DNA repair genes in the development of aggressive PCa for men of African ancestry. Quantifying the prevalence and penetrance of pathogenic germline variants in these and other DNA repair pathway genes will be required to better define men at high risk for clinically severe disease who would benefit from more regular screening and targeted prevention or therapeutic approaches.

AUTHOR CONTRIBUTIONS

Conception and design: Marco Matejcic, Raed Samara, Sue A. Ingles, Stephen Watya, Fergus J. Couch, Christopher A. Haiman

Financial support: Fergus J. Couch, Christopher A. Haiman

Administrative support: Vicky Kiddu, Benon Masaba, Fergus J. Couch

Provision of study materials or patients: Kun Y. Lee, Alexander Lubmawa, Dan Namuguzi, Kuteesa Job, Sue A. Ingles, Loic Le Marchand, Stephen Watya, Christopher A. Haiman

Collection and assembly of data: Marco Matejcic, Jenna Lilyquist, Chunling Hu, Kun Y. Lee, Siddhartha Yadav, Lucy Xia, Xin Sheng, Alexander Lubmawa, Vicky Kiddu, Benon Masaba, Dan Namuguzi, Kuteesa Job, Henry M. Dabanja, Sue A. Ingles, Lynne Wilkens, Loic Le Marchand, Stephen Watya, Fergus J. Couch, Christopher A. Haiman

Data analysis and interpretation: Marco Matejcic, Yesha Patel, Jenna Lilyquist, Chunling Hu, Rohan D. Gnanaolivu, Steven N. Hart, Eric C. Polley, Nicholas J. Boddicker, Xin Sheng, George Mutema, Sue A. Ingles, Fergus J. Couch, David V. Conti, Christopher A. Haiman

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).

Eric C. Polley

Research Funding: Grail

Raed Samara

Employment: Qiagen

Stock and Other Ownership Interests: Qiagen

Alexander Lubmawa

Employment: Uro Care Hospital (I)

Research Funding: Non-government organization (I)

Travel, Accommodations, Expenses: Uro Care Hospital (I)

Benon Masaba

Other Relationship: 7 Hills Medical Center

Stephen Watya

Employment: Uro Care Hospital

Leadership: Uro Care Hospital

Stock and Other Ownership Interests: Uro Care Hospital

Fergus J. Couch

Consulting or Advisory Role: AstraZeneca

Speakers' Bureau: Ambry Genetics, Qiagen

Research Funding: Grail

Travel, Accommodations, Expenses: Grail, Qiagen

Other Relationship: Ambry Genetics

No other potential conflicts of interest were reported.