Genetic Testing in Prostate Cancer

Abstract

Purpose of Review

This review summarizes recent advances in prostate cancer (PCa) genetics.

Recent Findings

Upwards of 20% of metastatic castration-resistant prostate tumors (mCRPC) carry homologous recombination (HR) repair gene mutations, of which ~ 10% are germline (inherited). Another ~ 5% exhibit microsatellite instability (MSI-H) and/or mismatch repair deficiency (MMRd). Pembrolizumab is approved for tumors with MMRd, thus patients with mCRPC and MMRd are candidates for pembrolizumab. Emerging data indicate that platinum chemotherapy and poly ADP-ribose polymerase inhibitors (PARPi) are effective in PCa exhibiting HR deficiency. NCCN guidelines now recommend germline and somatic tumor testing in specific clinical scenarios due to treatment and family implications.

Summary

Genetic testing in PCa patients may inform prognosis, treatment options, and have implications for family counseling. PARPi, platinum chemotherapy, and immune checkpoint inhibitors are promising targeted therapies for PCa with specific molecular features. Therapeutic advances, along with importance to relatives, are driving genetic testing in prostate cancer.

Introduction

Prostate cancer has a significant heritable component. In the past few years, substantial strides have been made in understanding genetic factors influencing prostate cancer susceptibility. Many of the recent discoveries have extended beyond common single-nucleotide polymorphisms (SNPs) to high and moderate penetrance genetic variants, alongside new precision-directed therapeutic implications that are leading to shifts in research and practice.

More than 100 susceptibility loci for prostate cancer have been identified with GWAS (genome-wide association studies), accounting for ~ 33% of familial prostate cancer risks [1–7]. Many variants identified were high prevalence and low-penetrance, and were not clinically used to differentiate risk of aggressive from indolent prostate cancer [7]. The germline HOXB13 G84E variant was established in 2012 as a susceptibility loci that significantly increased prostate cancer risk through study of families with multiple cases of prostate cancer [8]. Germline testing of unselected prostate cancer patients showed overall low prevalence of germline pathogenic or likely pathogenic variants (hereafter referred as mutations) in BRCA2; 1.2% among men were diagnosed before 65 years old [9, 10]. Prior to 2016, clinical germline testing for prostate cancer risk was not pervasive.

The Molecular Landscape of Metastatic Prostate Cancer

Substantial changes came with a Stand Up To Cancer-Prostate Cancer Foundation Prostate Cancer Dream Team study that sequenced metastatic prostate cancer biopsies and provided molecular characterization of later evolutionary stages of tumorigenesis [11]. A notable finding was that alterations in BRCA2, BRCA1, and ATM were observed in 19.3% (29/150) of metastatic tumors—a much higher frequency compared with localized disease [9, 11]. Eight percent (12/150) of patients with metastatic prostate cancer had pathogenic germline alterations [11].

Prevalence of Germline Mutations in Metastatic Prostate Cancer

In 2016, Pritchard et al. reported a study of germline sequencing in 692 men with metastatic prostate cancer, unselected for family history or age at diagnosis, and found that 11.8% (82/692) had germline DNA repair alterations compared with 4.6% (23/499) among men with localized prostate cancer from the Cancer Genome Atlas (TCGA), and 2.7% (1434/53,105) in persons without a known cancer diagnosis (EXAC) [12••]. A subsequent cross-sectional study by Nicolosi et al. evaluated 3607 men with a personal history of prostate cancer who underwent germline genetic testing between 2013 and 2018. The study found that 4.74% of patients with prostate cancer had gBRCA2 mutations, 2.88% gCHEK2, 2.03% gATM, 1.25% gBRCA1, 1.12% gHOXB13, 0.69% gMSH2, and 0.56% gPALB2 gene mutations [13]. The higher than previously recognized prevalence of DNA repair gene mutations led to changes in clinical testing guidelines.

Emerging data suggest that patients with ductal and intraductal prostate cancer carcinoma have higher risk of having microsatellite instability [14–17] and germline homologous recombination repair mutations [17–20]. Taylor et al. performed whole exome sequencing of prostate cancer tumors from 14 patients with castration-sensitive localized prostate cancer and gBRCA2 mutations, and showed that these tumors harbor increased genomic instability, and their mutation profiles resemble metastatic tumor [18]. A case study of patients with ductal histology prostate carcinoma showed that 49% (25/51) had DNA repair gene alterations, including 20% (10/51) with germline alterations. Somatic tumor sequencing of this patient cohort reported that 14% (7/51) tumors had mismatch repair gene alterations and 31% (16/51) in homologous recombination repair genes [17].

Collectively, these studies suggest that germline DNA repair mutations are present in substantial percentages in specific populations, (1) prostate cancer patients with metastatic disease, (2) those with a family history suggestive of inherited cancer predisposition, (3) ductal and intraductal histologic subtypes. NCCN guidelines for prostate cancer now recommend offering germline genetic testing to men in these groups (Table 1) [21, 22].

Table 1.

Family history for prostate cancer patients

Family history criteria

Family history of high-risk germline mutations (e.g., BRCA1/2, Lynch syndrome)

Brother or father or multiple family members diagnosed with prostate cancer (but not clinically localized grade group 1) at < 60 years or who died from prostate cancer

Ashkenazi Jewish ancestry

≥ 3 cancers on same side of family, especially diagnosed at age

≤ 50 years, bile duct, breast, colorectal, endometrial, gastric, kidney, melanoma, ovarian, pancreatic, prostate (but not clinically localized grade group 1), small bowel, or urothelial

Family Impact of Germline Testing

Germline testing in men with prostate cancer can potentially benefit the patient in informing treatment options, and if a mutation is identified, may also guide screening of other cancers and have family implications for cascade genetic testing (testing of close relatives for the same germline mutation). Early cancer detection strategies and preventive measures may be available to relatives identified to have same germline mutations. Cascade family testing can be valuable for family members, but unfortunately cascade family testing is currently underperformed, and strategies to overcome barriers, such as lack of knowledge, family communication, lack of access to genetic services, and cost of testing, are needed.

The NCCN guidelines for Genetic/Familial High-Risk Assessment for Breast and Ovarian Cancer [23] recommend breast cancer screening starting at age 25 years for female relatives and age 36 years for male relatives with deleterious gBRCA1/2 mutations. gBRCA1/2 female carriers should discuss with their health care provider risk reduction mastectomy and salpingo-oophorectomy, which is typically recommended between 35, 36, 37, 38, 39, 40 years old, and upon completion of childbearing.

Male relatives with gBRCA2 and gBRCA1 mutations may be at increased risk of cancers such as male breast, pancreatic, colon, and prostate. A study of 173 breast-ovarian cancer families with gBRCA2 mutations showed 4.7- to 8.6-fold increased risks of prostate cancer with cumulative risk of 20–33% in US carriers by the age of 70 years [24]. A study of 913 male gBRCA1 mutation carriers reported that gBRCA1 mutation increases the relative risk of prostate cancer by 3.75-fold and results in an 8.6% cumulative risk by age 65 [25].

The IMPACT study (Identification of Men with a Genetic Predisposition to Prostate Cancer: Targeted screening in gBRCA1/2 mutation carriers and controls) conducted yearly PSA screening in families with known gBRCA1/2 mutations [26, 27]. The study enrolled 2932 men with no personal history of prostate cancer, 919 gBRCA1 carriers, 902 gBRCA2, 497 gBRCA2 noncarriers, and 709 gBRCA1 noncarriers. Preliminary results reported after 3 years of follow-up showed overall 21% positive predictive value (PPV) of PSA > 0.3 ng/ml with 31% PPV in gBRCA2 carriers and 18% in gBRCA2 noncarriers; and 23% in gBRCA1 carriers and 15% in gBRCA1 noncarriers. PPV of prostate biopsy, initiated when PSA > 3.0 ng/ml, was 39% in gBRCA2 carriers and 28% in gBRCA2 noncarriers, but no significant difference was detected between gBRCA1 carriers and gBRCA1 noncarriers. gBRCA2 carriers had higher incidence of prostate cancer diagnosed at younger age and with more aggressive disease characteristics compared with gBRCA2 noncarriers. The results for gBRCA1 carriers were not definitive, and further investigation is needed. The number needed to screen to detect one clinically significant prostate cancer was 60 for gBRCA2 carriers ages 40–54 years and 13 for carriers ages 55–69 years. Thus, the results from IMPACT suggest annual PSA screening for gBRCA2 mutation carriers between age 40 and 69, using PSA cutoff of 3 ng/ml [27]. Studies evaluating predictive value of lower PSA cutoff and prostate MRI are ongoing (e.g., www.clinicatrials.gov, NCT03805919, NCT01990521).

Prognostic Impact of DNA Repair Gene Mutations

Prostate cancer patients may benefit from germline and/or somatic genetic testing to inform disease prognosis and treatment decisions, with somatic testing being potentially more relevant for treatment decisions. Several studies showed that germline BRCA1/2 mutations are associated with poor prognosis and worse outcomes in prostate cancer. Castro et al. showed that prostate cancer patients with gBRCA1/2 mutations were more likely to have a Gleason score ≥ 8, T3/4 stage, nodal involvement, and metastases at the time of diagnosis compared with noncarriers. Moreover, gBRCA1/2 carriers had shorter cancer-specific survival (CSS) compared with noncarriers (15.7 vs 8.6 years) [28]. gBRCA1/2 carriers with localized prostate cancer have worse outcomes after conventional treatment with surgery or radiation compared with noncarriers, 5-year metastasis free survival 72% vs 94%, 5 year CSS 76% vs 97% [29]. A retrospective case study by Na et al. found that proportion of men carrying gBRCA1/2 or gATM mutations was significantly higher in men who died from prostate cancer compared with men with localized disease (6.07% vs 1.44%). This study also showed that among patients with lethal prostate cancer rate of gBRCA1/2 and gATM mutations was higher in patients who died younger, 10% among those who died < 60years, 9% for age of death 61–65; 8% 66–70 years, 4% 71–75 years, and 2.97% among those who died > 75 years old [30•]. The prospective PROREPAIR-B found that gBRCA2 status is an independent prognostic factor for CSS in metastatic castration-resistant prostate cancer (mCRPC) patients (17.4 vs 33.2 months, p = 0.027) [31••].

Somatic DNA repair gene mutations are present in about 8–10% of localized prostate cancer cases and 20–25% of mCRPC cases, and Marshall et al. reported that presence in localized disease is associated with higher Gleason grade group (≥ 3) and more advance clinical stage (≥ cT3 disease) at the time of diagnoses [32].

Recommendations for Genetic Testing in Prostate Cancer

Based on these and other studies, NCCN guidelines for Prostate Cancer recommend offering germline testing, ideally with genetic counseling access, to the following groups of prostate cancer patients [22]:

1. Men with high-risk, very-high risk localized prostate cancer

2. Men with metastatic prostate cancer

3. Men with intraductal histology

4. Men with Ashkenazi Jewish ancestry

5. Men meeting family history criteria (see Table 1)

Somatic tumor sequencing should also be considered in prostate cancer patients especially with advanced disease. Sequencing of tumor metastatic biopsies is preferred when available, as tumor clones evolve over time, and primary prostate cancer tissue might miss alterations developed later in the disease course. Men should also be counseled that somatic tumor testing could potentially suggest presence of a germline mutation, and if the case, referral to genetic counseling to discuss dedicated germline testing would be advised. NCCN guidelines for Prostate Cancer recommend that germline and somatic testing panels should include Lynch syndrome associated genes (MLH1, MSH2, MSH6, and PMS2) and homologous recombination genes (BRCA1/2, ATM, PALB2, CHEK2) [21, 22]. Other genes might be appropriate for testing in certain scenarios, such as potential enrollment into clinical trials, where significance of newer gene/mutations as biomarkers for new treatments are being explored.

Treatment Implications of Genetic Testing

Advanced Disease

Targeted therapies are being investigated in clinical trials for prostate cancer patients with specific DNA repair gene alterations in tumor and/or germline. At present, the treatment implications of genetic testing are arguably greatest in metastatic disease, as this is the disease space with the majority of therapeutic clinical trials.

Some studies are available in earlier disease states of prostate cancer and more are expected to follow. Patients with homologous DNA repair mutations are candidates for clinical trials using poly ADP-ribose polymerase inhibitors (PARPi) and/or platinum chemotherapy, and novel combinations. There are several ongoing clinical trials evaluating the role of these agents as monotherapy or in various combination therapies in different stages of prostate cancer (e.g., www.clinicatrials.gov, NCT02975934, NCT02952534, NCT02854436, NCT02987543, NCT03413995). Below is a summary of current data on targeted therapy in prostate cancer patients with DNA repair gene mutations, i.e., immunotherapy, PARPi, platinum chemotherapy.

Immunotherapy

Pembrolizumab received the first tumor agnostic FDA approval in 2017 for metastatic solid tumors with microsatellite instability (MSI-H) and mismatch repair deficiency (MMRd) [33, 34]. About 5% of mCRPC cases are estimated to be MSI-H/MMRd and would qualify for treatment with pembrolizumab [35•, 36–38]. In a single institution case series of 1033 prostate tumors undergoing next generation sequencing, 3.1% (32/1033) were found to be MSI-H/MMRd, 78% (25/32) tumors had somatic mutations, and 22% (7/32) were found to have germline Lynch-associated mutations [35•]. Somatic and germline testing for MSI-H/MMRd is recommended in certain prostate cancer patient populations (see above) and has direct clinical implications as MSI-H/MMRd patients are eligible for pembrolizumab in second line of mCRPC treatment.

PARPi

Patients with DNA repair mutations have higher response rates to PARPi and platinum chemotherapy [39]. The clinical activity of PARPi in prostate cancer was first reported in the TOPARP-A trial (Trial of Olaparib in Patients with Advanced Castrate Resistant Prostate Cancer), which evaluated olaparib in mCRPC patients who failed multiple lines of therapy (98% received prior abiraterone or enzalutamide, 58% cabazitaxel) [40]. TOPARP-A reported that 88% (14/16) of patients with DNA repair mutations had response to olaparib therapy, where response was defined as a reduction in the PSA by 50% (PSA₅₀), a RECIST-defined objective response rate or circulating tumor cell reduction. Only 2 out 32 patients without DNA repair gene alterations responded to olaparib in TOPARP-A trial. As a result of this trial, olaparib received FDA breakthrough therapy designation in January 2016 for patients with BRCA2-, BRCA1-, or ATM-mutated mCRPC who had received prior taxane and either enzalutamide or abiraterone.

The phase 2 TRITON2 (NCT02952534) study evaluated rucaparib in mCRPC patients with homologous recombination gene mutation progressing after 1–2 lines of androgen receptor–directed therapy and 1 prior line of taxane. Preliminary results demonstrated that among BRCA1/2 carries 44% (11/25) had radiographic response, and 51% (23/45) had PSA₅₀ response [41••]. Based on these results, FDA granted rucaparib breakthrough designation in October 2018 [42].

The phase 2 TOPARP-B is a trial of PARPi in mCRPC patients with DNA damage repair alterations progressing after at least one taxane (n = 98) [43••]. The overall median progression-free survival (mPFS) was 5.4 months. Subgroup analyses per altered gene indicated following response rates (defined as in TOPARP-A study), BRCA1/2 83% (25/30; mPFS 8.1 months); PALB2 57% (4/7; mPFS 5.3 months); ATM 37% (7/19; mPFS 6.1 months); CDK12 25% (5/20; mPFS 2.9 months); other (ATRX, CHEK1, CHEK2, FANCA, FANCF, FANCG, FANCI, FANCM, RAD50, WRN) 20% (4/20; mPFS 2.8 months). The highest PSA₅₀ response rates were observed in the BRCA1/2 (22/30; 73%) and PALB2 (4/6; 67%) mutated subgroups.

The first phase 3 randomized clinical trial evaluating PARPi in mCRPC was recently reported [44••]. PROfound enrolled mCRPC patients who progressed on abiraterone or enzalutamide and had mutations in homologous recombination DNA repair genes, identified by tumor sequencing; cohort A, patients with BRCA1/2 and ATM mutations; cohort B, patients with mutations in other homologous recombination repair genes. Patients were randomized to olaparib 300 mg BID or physician’s choice of enzalutamide or abiraterone. The results showed improved radiographic PFS (rPFS) in the olaparib arm of cohort A (7.39 vs 3.55 months, homologous recombination (HR) 0.34 p < 0.001), with objective response rate of 33.3% in olaparib and 2.3% in physician’s choice arm of cohort A (OR 20.86 p < 0.001). Despite the crossover design, overall survival (OS) was 18.5 months in olaparib arm compared with 15.11 months in abiraterone/enzalutamide arm of cohort A, although statistical significance was not reached at time of initial reporting. There was also rPFS benefit in combined cohorts A and B, 5.82 months in olaparib arm vs 3.52 months in physician’s choice arm (HR 0.49 p < 0.001). We anticipate that the FDA will approve olaparib for subset of mCRPC patients on the basis of the PROfound trial results.

Table 2 summarizes currently available study results reporting response rates to PARPi in prostate cancer.

Table 2.

Response to PARPi in mCRPC patients stratified by HR gene mutations

Study	Agent used	Response measured by	Number of patients responded to PARPi by mutation status
			BRCA 1/2	ATM	CDK12	Other HRD mut	No HRD mut
PROfound, Hussain et al. [44••]	Olaparib vs abiraterone/enzalutamide	Imaging rPFS (months)	84/162 vs 42/83 7.39 vs 3.55		~ 3/94 vs	~4/48	N/A
TOPARP-A, Mateo et al. [40]	Olaparib	Imaging PSA50 CTC	8/8	4/5	N/A	2/3	2/33
TOPARP-B, Mateo et al. [43••]	Olaparib	Imaging PSA50 CTC	24/30	7/19	5/20	8/27	N/A
TRITON2, Abida et al. [41••]	Rucaparib	Imaging	11/25	0/5	0/8	2/8	N/A
		PSA50	23/45	0/18	1/13	2/9	N/A
NCI study (50) Karzai et al. [45]	Durvalumab + olaparib	Imaging PSA50	7/11	N/A	N/A	N/A	2/6
GALAHAD, Smith et al. [46]	Niraparib	Imaging PSA50 CTC	18/29	N/A	N/A	5/21	N/A
KEYNOTE-365, Yu et al. [47] 2019	Olaparib + pembrolizumab	Imaging	N/A	N/A	N/A	N/A	8/28
		PSA50	N/A	N/A	N/A	N/A	5/41
Retrospective analysis, Marshall et al. [48]	Off-label olaparib	PSA50	13/17	0/6	N/A	N/A	N/A

Mut, mutations; CTC, circulating tumor cell DNA: PSA₅₀ decline of prostate-specific antigen by 50% from baseline; Imaging, radiographic response measured by RECIST criteria; rPFS, radiographic progression-free survival

Platinum chemotherapy

Platinum chemotherapy has been proven to be effective in BRCA1/2 mutated breast and ovarian cancers [49, 50]. Our single institution retrospective case series showed that 3/3 prostate cancer patients with biallelic inactivation of BRCA2 had exceptional response to platinum chemotherapy after progressing on several therapies [51]. The findings were supported by a retrospective study that showed 75% (6/8) of mCRPC patients with gBRCA2 mutation had PSA₅₀ response to platinum chemotherapy compared with 17% (23/133) of mCRPC patients without gBRCA2 mutations [52].

Localized Disease

In an active surveillance cohort, patients with localized prostate cancer and gBRCA2, gBRCA1, or gATM mutations were more likely to experience grade reclassification compared with non-mutation carriers [53]. Further studies are warranted, but these data suggest that this group of patients will be monitored closely, ideally on a clinical trial, or consider a definitive treatment approach. The role of PARPi in localized prostate cancer is being evaluated in several ongoing clinical trials (www.clinicaltrials.gov, NCT03570476, NCT02324998 and NCT03432897).

Biochemical Recurrence

Patients historically classified as having biochemically recurrent (BCR) prostate cancer are now moved to the metastatic group with the use of more sensitive treatment modalities such as PSMA-PET and fluciclovine PET. Advances in imaging has changed the BCR patient population. There is currently no standard of care treatment implications of DNA repair mutations in BCR group, but the use of PARPi is currently being studied in phase 2 clinical trial (NCT03047135), evaluated PSA₅₀ response to olaparib in BCR prostate cancer patients. The study is enrolling patients unselected for DNA repair mutation status with an adaptive plan to enrich study population with DNA repair mutation carriers if response is low in first 20 patients.

Conclusions

There has been significant advancement in prostate cancer genetics in the last 5 years. In metastatic castration-resistant prostate cancer, ~ 20% of tumors harbor homologous recombination repair gene mutations, and 5% harbor mismatch repair gene mutations; alternations in both pathways have clinical implications. The NCCN Prostate Cancer guidelines recommend germline testing for men with high-risk, very high-risk localized prostate cancer, all metastatic prostate cancer patients, patients with intraductal histology of prostate cancer, and for patients meeting family history criteria. Somatic tumor testing should also be considered for advanced disease as it may inform treatment decisions.

Patients with homologous recombination-deficient prostate cancer appear to respond to PARPi and platinum chemotherapy, although more individual gene-specific data is needed. The PROFound study, a phase 3 randomized clinical trial of olaparib in mCPRC patients with BRCA1/2 and ATM mutations showed improved rPFS. Patients with MSI-H/MMRd are eligible for immune checkpoint inhibitor therapy in second-line treatment for mCRPC. Both germline and somatic tumor genetic testing are recommended for prostate cancer patients in specific clinical scenarios. Clinical trials are ongoing to evaluate treatment implications of alterations in mismatch repair genes and homologous recombination genes, and we expect new indications and combinations of PARPi and immune checkpoint inhibitors to be explored by these trials. Further investigation is needed to identify individual gene contributions to treatment response prediction and germline risk of prostate and other cancers.