Prognostic impact of BRCA mutation on metastasis-free survival in a localized/locoregional high-risk real-world prostate cancer population

Background: Understanding the prognostic impact of BRCA mutation status across prostate cancer (PCa) patient segments is critical to deploying the right therapies at the right time in a patient’s journey. Prior studies have shown that germline BRCA mutations are associated with poorer outcomes in an all-risk localized PCa population. To examine whether this association extends to a high-risk PCa population with either somatic or germline BRCA mutations, we evaluate the prevalence and independent prognostic impact of BRCA mutations in a real-world (RW), high-risk, localized/locoregional PCa population.

Methods: A retrospective study (2016-2023) was performed using a de-identified multimodal RW dataset of patients undergoing Tempus xT next-generation sequencing. Eligible patients had localized/locoregional PCa at primary diagnosis (pDx), were biopsied prior to metastatic diagnosis, received adjuvant radiotherapy and/or had a prostatectomy, and had an evaluable Gleason score (GS). Risk groups were categorized as low/intermediate (low/int)-risk (total GS <8) and high-risk (total GS ≥8). BRCA status was defined as BRCAm if the patient had a short variant that was predicted or known to be pathogenic or a copy number loss in BRCA1 or BRCA2 and BRCAwt if no mutation was detected. Median real-world metastasis-free survival (rwMFS), defined as time from primary intervention to first metastatic diagnosis, was estimated using Kaplan-Meier methods and stratified by combined BRCA status and risk group (BRCAm high-risk, BRCAm low/int-risk, BRCAwt high-risk, BRCAwt low/int-risk). To assess the independent prognostic value of BRCA status among all patients and within a high-risk subset, two multivariate Cox models were performed adjusting for risk group, age at pDx, baseline PSA, primary intervention type, and T and N stage at pDx.

Results: Among 607 eligible patients, 67 (11%) had a BRCA mutation (11 germline, 27 somatic, 29 tumor/unknown). Median rwMFS was longest in the BRCAwt low/int-risk group (n=147) at 50 mos (95% CI: 41, NA), followed by the BRCAwt high-risk group (n=393) at 36 mos (95% CI: 31, 42), and the BRCAm high-risk group (n=61) at 22 mos (95% CI: 11, 45). Median rwMFS estimates in the BRCAm low/int-risk group were unreliable due to a small sample size (n=6). From the adjusted Cox model, BRCAm patients had higher risk of metastasis compared to BRCAwt patients in the all-risk group (HR=1.48, 95% CI: 1.06, 2.09) and in the high-risk subset (HR=1.50, 95% CI: 1.04, 2.17).

Conclusions: Our findings show that BRCAm is independently prognostic of metastasis in a RW, localized/locoregional, high-risk PCa population. We identify high-risk BRCAm patients as a subgroup with significantly elevated risk of developing metastasis; clinical trials with targeted therapeutic intervention may be needed to address the clinical unmet need.