Comprehensive genomic profiling of treatment resistant metastatic castrate sensitive prostate cancer reveals high frequency of potential therapeutic targets.

Introduction: While comprehensive genomic profiling (CGP) data is becoming increasingly important in the management of prostate cancer, it remains under-utilized in the setting of metastatic castrate sensitive prostate cancer (mCSPC). We aimed to explore the feasibility and potential utility of CGP in mCSPC.

Patients And Methods: Patients with mCSPC were prospectively recruited at the Princess Margaret Cancer Centre to the OCTANE trial (NCT02906943). The objective was to assess the feasibility of profiling archival standard diagnostic tumor tissue using next generation sequencing with a custom hybridization capture DNA-based or a targeted DNA/RNA amplicon-based panel. Clinical data were extracted from electronic health records.

Results: Among 39 mCSPC patients enrolled, 21 (54%) had sufficient archival tissue for CGP. Most had high volume (71%) or de novo (71%) mCSPC, with the majority being androgen deprivation therapy (ADT) naïve. In total, 62% of patients had a pathogenic and/or a likely pathogenic variant, many of which involved DNA damage repair (DDR, 19%), cell cycle (24%), and Androgen Receptor (AR, 10%) pathways. After median follow-up of 32.1 months, 18 of 21 patients progressed, with median time to mCRPC of 14.3 months (95% CI 10.2-21.0). Patients with AR and DDR variants seemed to have shorter median time to mCRPC; 10.2 (95% CI 9.50-NR) and 10.3 months (95% CI 6.6-14.3) respectively.

Conclusion: In this cohort of highly treatment resistant mCSPC, most of which were ADT-naïve, CGP using archival tumor tissue was feasible for over half of patients, and 62% of patients profiled had a pathogenic and/or a likely pathogenic variant. The presence of de novo variants provides biological basis for evaluating intensification strategies of systemic therapy. This highlights the potential role of routine CGP in biomarker development and clinical trial design in the setting of mCSPC.