Germline genetic testing for inherited prostate cancer in practice: Implications for genetic testing, precision therapy, and cascade testing

Genetic testing capability and guidelines are rapidly expanding to assess inherited prostate cancer (PCA). Clinical genetic data from multigene testing can provide insights into the germline pathogenic variant (PV) spectrum and correlates in men with PCA unselected for metastatic disease to optimize identification of men for genetic evaluation and management.


| INTRODUCTION
Genetic testing for inherited prostate cancer (PCA) is increasing with expansion of genetic testing guidelines and increasing capability of clinical multigene testing. [1][2][3] The NCCN Genetic/Familial High-risk Assessment: Breast and Ovarian (Version 2.2019) recommends BRCA testing for men with metastatic PCA or men with Gleason score>= 7 who have any of the following: (1)>= 1 close blood relative with ovarian, pancreatic, metastatic PCA, or breast cancer at age <50, (2) > = 2 close blood relatives with breast or prostate cancer at any age, or  [1][2][3][4][5][6][7] Furthermore, genetic results particularly for germline DNA repair pathogenic variants are increasingly informing therapeutic, clinical trial, and management options for men with PCA. 1,4 Germline pathogenic variants in DNA repair genes including BRCA2, BRCA1, ATM, PALB2, and CHEK2 have been reported in up to 12% of men with advanced or metastatic PCA, 5,7 with data regarding clinical activity of PARP inhibitors emerging. 4,8 In 2016, olaparib was granted "Break-Though Therapy" designation by the United States FDA for BRCA1/2 or ATM mutated metastatic PCA. Furthermore, the identification of mismatch repair deficiency across tumor types is becoming important due to the potential role of immunotherapy in this setting. 9 Prior studies have focused on men with metastatic PCA or who meet strict study eligibility criteria. [5][6][7] Broader genetic testing data from clinical practice settings can expand to our understanding of the estimated scope of men with inherited PCA irrespective of metastatic disease, with subsequent impact of germline genetic results for patient management, screening, and cascade testing, as well as lend support to expanding guidelines. Since multigene panels are now commercially available for genetic testing for PCA in clinical practice, this represents a unique opportunity to learn about pathogenic variant rates in key genes of interest in PCA biology and predisposition to inform patient management and lend supporting data to emerging guidelines.
This study was performed utilizing de-identified genetic testing data from a commercial clinical genetic testing laboratory in the United States and represents a "real-world" data analysis among men with PCA unselected for metastatic disease. The goal of the study was to estimate overall pathogenic variant rates among men with PCA, with a focus in pathogenic variants in DNA repair genes where there are significant clinical implications for treatment and screening for men with PCA and their families. [1][2][3][4] Correlates of pathogenic variant rates were also assessed to optimize identification of men with PCA for genetic testing and inform implications for precision therapy, cancer screening, and cascade testing in families.

| MATERIALS AND METHODS
De-identified genetic results were received from a commercial genetic testing laboratory (Invitae, San Francisco, CA) in the United States.
Samples for this analysis included those with ICD-10 codes indicating a personal history of PCA (C61, Z85.46). The use of ICD codes is a validated method to determine diagnoses or patient characteristics for research. 16,17 However, because the ICD-10 codes for PCA do not distinguish different stage of disease, no categorization by stage was possible using these codes as in prior published studies. 16  DNA repair genes were also assessed for pathogenic variants, and were defined as: BRCA1, BRCA2, ATM, BRIP1, CHEK2, NBN, MSH6, PMS2, RAD50, PALB2, and FANCA as per prior studies. 5 Fisher's exact test (with correction for false-discovery) was used to evaluate the association between DNA repair gene pathogenic variants with personal history/family history (as indicated by ICD-10 codes) and Gleason score (which was abstracted from test request forms). Race, age at diagnosis, and stage information was not available through claims codes and were not systematically entered on test request forms. All analyses were performed in SAS 9.4 (SAS Institute Inc., Cary, NC).

| RESULTS
As of August 2017, de-identified genetic test results were available from 1328 men with ICD-10 codes indicating a personal diagnosis of PCA. Family history of breast cancer was reported most commonly (n = 453), followed by family history of PCA (n = 369) and family history of GI cancers (n = 299). Among 898 men with Gleason score information available, 6.2% had a Gleason score >= 8.
Overall rate of pathogenic variants was 15.6% and rate of VUS was 37.2%. Overall rate of pathogenic variants in DNA repair genes was 10.9%. Pathogenic variant rates by genes tested is depicted in Figure 1.
BRCA2 pathogenic variants were the most commonly identified (4.5%), and RAD51D (0.1%). Table 1  Family history of breast cancer was significantly associated with an approximate two-fold risk of carrying germline DNA repair pathogenic variants and may therefore be an important predictor to identify men with PCA for genetic testing. This finding reinforces the need for ordering providers, increasingly urologists and oncologists, to perform intake of broad family history for appropriate patient referrals to cancer genetics and to understand and deliver family history-based recommendations to men with PCA and their families. Gleason score> = 8 was also associated with an approximate two-fold increased risk for carrying germline DNA repair pathogenic variants. While Gleason data were available in a subset of the cohort, our results in an unselected population confirm and expand upon prior reports which described association of higher Gleason score to DNA repair gene pathogenic variants in men with metastatic disease. 5    There are some limitations to consider. The dataset did not include age at PCA diagnosis, stage, race, or age at diagnosis in family history, which was not systematically collected on test request forms and therefore has limited reliability and data quality. Therefore, ICD-10 claims codes were used to annotate personal history of PCA and family history, which has been performed by prior studies. 16  Family history data were derived from ICD-10 codes and may be incomplete. Therefore "not reported" may not be equivalent to "no family history." The use of a single lab as a data source is also a consideration.
There is known inter-lab variability of variant pathogenicity designation which could impact rates of pathogenic variants and VUS reported here. Additional factors may also influence results, such as differences in genetic testing practices between providers, patients tested, use of specific laboratories, or panel utilization. Thus, our results need to be confirmed for generalizability.
Since prior data have shown that PCA patients with metastatic disease have higher rates of germline DNA repair pathogenic variants which can inform therapy such as PARP inhibitors, 4,5 an additional consideration is that advanced cases may have made up a significant proportion of our dataset leading to uncharacterized selection bias.
Furthermore, genetic testing was performed prior to expansion of NCCN guidelines to include broad family history beyond cancers linked with HBOC. 1 Updated analyses in the era of expanded genetic testing guidelines will be needed.

| CONCLUSION
Our report includes one of the largest sample sets of "real-world" genetic testing data to inform strategies for identification of men for genetic evaluation for inherited PCA. As genetic testing for men with PCA increases, provider education for urologists and oncologists regarding genetic results interpretation and family history-based recommendations for men with PCA and their families will be important to provide appropriate cancer genetic education and care delivery.

AUTHOR CONTRIBUTIONS
VNG had access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design by VNG, KEK, and LGG. Acquisition, analysis, or interpretation of data by VNG, SEH, CH, EOL, JG, KEK, WKK, LGG.

CONFLICTS OF INTEREST
EOL and JG have stock ownership in Invitae. The use of Invitae data were not biased by involvement of these authors. All other authors have no conflict of interest.