Abstract
PURPOSE: This study used a deidentified nationwide (US-based) high-grade serous ovarian cancer (HGSOC) clinicogenomic database (CGDB) to enrich our understanding of HGSOC's genomic heterogeneity and assess the utility of comprehensive genomic profiling (CGP) in clinical settings. PATIENTS AND METHODS: We conducted a retrospective observational analysis on 856 patients with HGSOC profiled with CGP genomic tests, retrieved from CGDB from January 2011 to September 2023. RESULTS: In addition to BRCA1 (11.7%) and BRCA2 (6.5%) variants, CGP revealed further potentially actionable alterations (amplifications and/or mutations) in CCNE1 (16%), FGFR1/2/3/4 (6.5%), PIK3CA (3.9%), TP53(Y220C) (3.7%), ERBB2 (3.5%), CDK12 (2.3%), ARID1A (2.2%), KRAS (2.1%), and BRAF (1%) genes. Then, 439 patients were selected, presenting both CGP test performed on specimen collected at the time of surgery and initiation of first-line therapy within ±8 months from surgery, and categorized into no (NS, n = 74), interval (IS, n = 157), and upfront surgery (n = 208) groups, each comparable by clinical features. The CGP revealed BRCA mutations, at similar frequency in the three groups, in 54/439 patients (12.3%). Patients with pathogenic BRCA mutations had better event-free survival (EFS) compared with those with BRCA(wt). Loss-of-heterozygosity (LOH) ≥16 (LOH-positive patients) was found in 142/433 (32.8%) patients, with different prevalence across treatment groups (12.8% NS; 9% IS; 8.8% upfront surgery). Patients treated with poly (ADP-ribose) polymerase inhibitors (PARPi) had improved EFS (hazard ratio for other drugs v PARPi 1.77 [95% CI, 1.21 to 2.58]). Interestingly, in 206 BRCA(wt) and LOH-negative patients, not eligible for PARPi, CGP detected potentially targetable alterations in 99 of them (48%). CONCLUSION: Overall, our study provides evidence that CGP significantly improves the identification of molecular targets in HGSOC, supporting its importance in the clinical practice to provide patients with more therapeutic options.