Evaluating the Role of Liquid Biopsy to Detect Pathogenic Homologous Recombination Repair (HRR) Gene Alterations in Metastatic Prostate Cancer.

BACKGROUND: Metastatic prostate cancers frequently harbour pathogenic aberrations in Homologous Recombination Repair (HRR) genes that confer sensitivity to PARP inhibitors (PARPi). Therefore, accurate identification of all eligible patients is needed. The development of a circulating tumour DNA (ctDNA) testing alternative is promising as genomic testing of archived tissue leads to a failure rate of up to 30-40% in prostate cancer. METHODS: This was a bi-institutional retrospective cohort study of patients with metastatic prostate cancer treated at the Jewish General Hospital or the McGill University Health Center, Montreal, Canada, between 2021 and 2023. Molecular data and treatment information were abstracted from a chart review. Chi-square, Fisher's exact test, and Mann-Whitney tests were used to assess differences between groups. RESULTS: We identified 484 metastatic prostate cancer patients. Somatic and germline testing for HRR was performed in 55.4% (n = 268) and 20% (n = 97) patients, respectively. Somatic testing was performed on tissue (n = 192, 71.6%) or ctDNA from liquid biopsies (n = 18, 6.7%) or both (n = 58, 21.7%). Pathogenic somatic HRR alterations were detected in 48 patients (17.9%). BRCA2 was the most frequent (n = 17), followed by ATM (n = 11), then CHEK2 (n = 5). Amongst patients with germline testing, 13/97 (13.4%) had pathogenic alterations predicted to lead to deficient HRR, mostly BRCA2 (n = 9), and three had detectable BRCA2 in tissue. Dual testing modality (tissue+ctDNA) significantly enhanced the detection rate of HRR alterations 19/58 (32.7%) vs. 29/210 (13.8%) for single testing modality (tissue or ctDNA), p = 0.008. The rate of inconclusive results was significantly lower in dual testing modality 0/58 (0%) vs. 25/210 in single testing modality (11.9%), p = 0.003. Amongst the 14 patients who had discordant results between liquid and tissue tests, HRR abnormalities were more frequently identified in ctDNA (n = 11) vs. tissue (n = 3). Patients who had HRR deficiency detected only in ctDNA had older tissue samples (median 5.6 years) compared to those who had deficient HRR detected only in tissue (median 0.2 years; p = 0.14). CONCLUSIONS: These data highlight a potential role in implementing liquid biopsy-especially in patients who only have older archival tissue available or failed tissue testing-to improve the detection rate of deficient HRR. Our ongoing prospective study will further validate whether the addition of liquid biopsy can identify more patients who are eligible to receive precision therapies by increasing the rate of detection of HRR deficiency compared to routine tissue testing alone.