Germline DNA repair gene (gDRG) variants in BCG un-responsive patients: A potential genetic implication for risk stratification and target therapy.

Abstract

808 Background: The introduction of the “BCG-unresponsive” definition by the U.S. Food and Drug Administration (FDA) has provided a structured framework for classifying patients unlikely to benefit from BCG therapy, accelerating the approval of novel therapeutic agents. There is an urgent need to improve tools predicting BCG failure. To address this unmet need, we investigated the predictive role of gDRG pathogenic variants (PVs), likely pathogenic variants (LPVs) and variants of unknown significance (VUS) in patients who received BCG. Methods: This is a nested case-control study, within an ongoing prospective single center trial, which includes patients with NMIBC who have received BCG and developed a recurrence. The study was founded by AIRC - Fondazione AIRC per la Ricerca sul Cancro, supported by the Italian Ministry of University and Research (MUR) under the National Recovery and Resilience Plan (PNRR), and registered in Italy (amendment number: ID-IG-25027-V1.3). This study was Naïve patient who received BCG treatment for HG-NMIBC from June 2023 to March 2025 underwent whole-exome sequencing. They were investigated for PVs, LPVs and VUS (including TM, ATR, MRE11A, BAP1, BARD1, BRCA1, NBN, BRCA2, PALB2, BRIP1, CHEK2, RAD51C, FAM175A, RAD51D, GEN1, XRCC1, and the mismatch repair genes MLH1, PMS2, MSH2, and MSH6). The primary endpoint was to describe the prevalence of PVs, LPVs and VUS in BCG un-responsive and responsive patients. Results: From January to June 2024 to August, a total of 18 naïve patients were enrolled. Seven patients (mean age 57.5 yrs – range 38-74) showed a recurrence during the follow-up and were defined “BCG un-responsive”. The median follow-up at the time of testing was 14 months. All patients had a history of high-grade bladder cancer and received intravesical BCG adjuvant treatment according to the SWOG protocol. None of the responders harbored PVs-LPVs, while two (18.1%) had VUS. In un-responsive group 2 over 7 (71.4%) presented at least a PV, and 5 a VUS; one of this was reclassified as LPVs by literature. The main variants were: ATM, BRCA2, MSH6, ATR, BARD1 and PALP2. The two BCG-unresponsive patients with more than one variant had a significant lower age at the diagnosis: 38 and 52 yrs old, and presented carcinoma in situ at recurrence. Conclusions: These very preliminary findings seem suggesting that gDRG variants could matter in patients who failed BCG course and underscore the need for further investigation into the utility of germline genetic risk stratification as crucial role in refining patient selection for BCG and new target therapies.