A comprehensive literature review and meta-analysis on prevalence of BRCAm, HRRm and HRD+ across tumor types.

Abstract

10589 Background: Poly (ADP-ribose) polymerase inhibitor (PARPi) may have broad application in the treatment of cancer patients with mutations in BRCA (BRCAm) or other homologous recombination repair genes (HRRm) or with homologous recombination deficiency positive (HRD+). A literature review and meta-analysis were conducted to evaluate the prevalence of BRCAm, HRRm, and HRD+ across tumor types. Methods: Comprehensive searches for eligible studies in Ovid MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and Cochrane reviews were performed in May 2020 to capture studies published in English, within 10 years for manuscripts and 3 years for abstracts across geographic regions. A weighted summary estimate was calculated for BRCAm. A summary estimate with corresponding 95% CI was calculated using random-effects models for HRD+ and HRRm. Results: A total of 342 eligible studies with at least 100 samples were included in the review of BRCA1/2m prevalence, containing a total of over 469,000 samples across 24 tumor types. The most frequent indications examined in the included studies were breast (study number n = 144), ovarian (53), prostate (17) and pancreatic (11) cancers. The prevalence of germline BRCA1m (gBRCA1m) and gBRCA2m was 5% and 4% for breast, 12% and 5% for ovarian, 1% and 3% for prostate, and 1% and 4% for pancreatic cancer, respectively. The prevalence of somatic BRCA1m (sBRCA1m) and somatic BRCA2m (sBRCA2m) was 3% and 3% for breast, 7% and 5% for ovarian, 3% and 5% for prostate, and 1% and 3% for pancreatic cancer, respectively. Few studies evaluated endometrial, lung and colon cancers, the prevalence of gBRCA1m or gBRCA2m was generally less than 1%, and the prevalence of sBRCA1m and sBRCA2 ranged 1 to 3%. Seven publications were identified where HRD+ was defined by either BRCAm or genomic instability score (GIS) ≥ 42 across breast, ovarian and pancreatic cancers. The overall HRD+ prevalence was 56% (95%CI: 48, 64), with similar prevalence observed across the 3 tumor types and was 50% (34, 66) for the 3 studies only counting GIS≥ 42. 194 studies across 26 tumor types were identified that examined HRRm as mutations in one or more HRR genes other than BRCA1/2m. The definitions of HRRm varied substantially across the studies, and ATM (2.8%), CHEK2 (1.6%), and PALB2 (1.6%) accounted for most of the observed mutations among HRR genes. Conclusions: Prevalence of BRCAm, HRRm and HRD+ varied by cancer type. This comprehensive meta-analysis enriches the knowledge in this field and demonstrates the need to standardize the measurement of HRRm and HRD.Understanding the prevalence of these biomarkers could have important clinical implications.