Ethnic disparities among men with prostate cancer undergoing germline testing

Highlights

• About 13.8% of men with prostate cancer have pathogenic/likely pathogenic germline variants.

• Racial/ethnic minority men are under-represented in studies of germline testing.

• Germline variants of uncertain significance are frequent in racial/ethnic men.

Abstract

Background

Prostate cancer is among the most heritable cancers, and clinical testing for germline genetic variants based on ethnicity, disease features, and family history has recently become standard of care for men with advanced disease. It is not established whether prevalence of germline variants varies based on ethnicity or race.

Methods

We retrospectively examined germline genetic and clinical data of men reporting a diagnosis of prostate cancer referred to Color Genomics by a healthcare provider for testing of 30 genes associated with hereditary cancer risk. Variants were classified as pathogenic (P), likely pathogenic (LP), variant of uncertain significance (VUS), likely benign, or benign. P/LP and VUS prevalence was compared among subgroups classified by age at diagnosis, self-reported ethnicity, family history, and history of other cancer.

Results

We identified 1,351 men reporting a diagnosis of prostate cancer of any stage who underwent germline testing. Overall, 78% of men were Caucasian, 11% Ashkenazi Jewish, 3% African-American/Canadian (AAC), 2% Hispanic, 2% Asian/Pacific Islander (API), and 4% Other (multiple, unknown, Native-American). One-hundred eighty-seven men (13.8%) carried a P/LP variant, and the most prevalent P/LP variants were in BRCA2 (3.4%), CHEK2 (2.8%), MUTYH (1.8%), and ATM (1.7%). Age at diagnosis, ethnicity, type of family member with prostate cancer, and type of second cancer were not associated with risk of carrying any P/LP variant. Ashkenazi Jewish men (6.7%) were more likely to carry P/LP BRCA2 variants than Caucasian men (2.8%) (P < 0.05). Two-hundred eighty-four men (21.0%) carried a VUS, and AAC (36.6%) and API (33.3%) men were most likely to carry a VUS (P < 0.01).

Conclusions

P/LP germline variants are prevalent in men with prostate cancer. AAC, Hispanic, and API men with prostate cancer are under-represented in studies of germline testing, potentially contributing to higher rates of VUS relative to Caucasian and Ashkenazi Jewish men. Further studies in these groups will facilitate reclassification of VUS, increasing opportunities for early detection, cancer risk modification, and targeted therapeutics.

Introduction

Prostate cancer is the most frequently diagnosed cancer among men in the United States. In 2019, there will be an estimated 174,650 new cases accounting for 20% of all cancers, and an estimated 31,620 prostate cancer-related deaths [1]. Prostate cancer is also one of the most heritable cancers, with an estimated 57% of incident cases deemed heritable [2].

Familial susceptibility is thought to be due to a spectrum of rare to common variants with inversely associated effect sizes [3]. Genome-wide association studies have identified over 140 common variants of loci that confer a slightly increased risk of prostate cancer, generally in a polygenic risk model, and explain 28% of excess familial prostate cancer risk [3,4], some of which may have an active role in tumorigenesis [5]. Conversely, variants of genes such as BRCA1/2, CHEK2, ATM, PALB2, and HOXB13, though less common, may confer a substantially greater risk of prostate cancer than common variants [6], [7], [8], [9], [10].

In 2016, a study of men with metastatic prostate cancer demonstrated a high prevalence (11.8%) of germline mutations in DNA repair genes independent of age at diagnosis or family history [11], stimulating a new era of multigene germline testing, and prompting professional societies to develop guidelines on screening for germline mutations. The National Comprehensive Cancer Network Prostate Cancer Version 2.2019 guidelines recommend cancer predisposition next-generation sequencing for the homologous recombination genes BRCA1, BRCA2, ATM, PALB2, and CHEK2; and MLH1, MSH2, MSH6, and PMS2 for Lynch syndrome in the following situations: family history of high-risk germline mutations (e.g., BRCA1/2, Lynch mutation), suspicious family history, and presence of intraductal carcinoma on biopsy. Germline testing can also be considered for high- or very-high risk prostate cancer, and additional genes such as HOXB13 can be tested depending on clinical context [12]. The Philadelphia Prostate Cancer Consensus and Germline Genetics Working Group of the Prostate Cancer Clinical Trials Consortium presented similar recommendations, though with minor differences in which genes to test (e.g., HOXB13), and age cutoffs for testing [13,14]. While expanded use of multigene testing has created great opportunities to widen the understanding of germline mutations in men with prostate cancer, it has also proven challenging to identify all men at high risk of carrying germline mutations while minimizing unnecessary testing in men at low risk [13,15].

Multiple cohort studies subsequent to Pritchard et al. have described germline variants among men with prostate cancer [11,15,16]. However, few ethnic/racial minority men have been studied (e.g., only 5.8% of men in Pritchard et al. were non-Hispanic black) [11], and whether prevalence of germline variants varies based on race or ethnicity has not been reported. Herein, we aim to determine whether prevalence of germline variants varies by ethnicity. We hypothesize that men with prostate cancer of African-American/Canadian (AAC), Hispanic, or Asian/Pacific Islander (API) descent are under-represented, and carry higher rates of variants of unknown significance than men of Caucasian or Ashkenazi Jewish descent.