Abstract 4019: Clinical validation of WES-based HRD scoring and its prognostic value for PARP inhibitor maintenance therapy in ovarian cancer

Abstract Background: Homologous recombination deficiency (HRD) predicts PARP inhibitor (PARPi) sensitivity in ovarian cancer. We evaluated the clinical and analytical performance of a WES-based genomic profiling assay generating HRD scores and BRCA status, and assessed its prognostic value for PARPi maintenance. Methods: We included 81 consecutive patients with epithelial ovarian cancer (FIGO I-IV) who had received PARPi therapy between 2017-2024 at Seoul National University Hospital. After tumor purity estimation and microdissection, formalin-fixed, paraffin-embedded (FFPE) surgical tumor tissues and matched blood were analyzed using the CancerProfiler™ WES panel. HRD scores were derived from genomic scar metrics (LOH, TAI and LST) following the scarHRD algorithm. Germline and somatic BRCA1/2 variants were profiled concurrently. HRD-high status was defined at ≥42 and ≥50. Tumor purity and fraction were estimated from WES data and incorporated into HRD computation to ensure quantitative stability. Results: Median age was 56; 96% had high-grade serous carcinoma and 94.8% FIGO III-IV disease. Optimal cytoreduction (1 cm residual) was achieved in 61%, and 52% received neoadjuvant chemotherapy. By the WES panel assay, a total of 46 patients had germline or somatic BRCA mutations. HRD-high status was observed in 77.8% at ≥42 and 66.7% at ≥50; the ≥50 cutoff showed a stronger association with germline BRCA (OR 3.15; p=0.021). Among 73 PARPi maintenance recipients (52 first-line and 23 second-line), HRD ≥50 showed differential predictive value by NAC. In first-line recipients, PFS was similar between HRD-high and -low after NAC (15.3 vs 11.3 mo, p=0.77) but markedly longer in HRD-high without NAC (median not reached vs 7.66 mo, p=0.001). This pattern persisted in the second-line setting (NAC: 19.48 vs 6.49 mo, p=0.049; no NAC: 17.78 vs 2.43 mo, p0.001). BRCA mutation predicted benefit only in NAC-treated first-line patients (17.7 vs 4.5 mo, p=0.007) and not elsewhere. HRD-high was an independent predictor of longer PFS (first-line HR=0.26, p=0.02; second-line HR=0.20, p=0.029). In multivariable analyses, the predictive value of PARP inhibitors was BRCA-dependent only in the first-line setting (HR=0.09, p=0.001), with no significant association observed in the second-line (HR=0.28, p=0.226). By contrast, HRD-high status remained a consistent independent predictor in both first- (HR=0.26, p=0.02) and second-line therapy (HR=0.02, p=0.029), emerging as the sole significant determinant after recurrence. Conclusions: HRD ≥50 is a robust predictor of PARPi benefit—particularly in treatment-naïve tumors—whereas BRCA status shows limited and NAC-dependent predictive value. These findings support prospective multi-center validation of WES-based HRD assays as scalable alternatives to array-based platforms. Citation Format: DongSoo Kyung, Kyung Hee Han, Yongjun Cha, Won Yeong Ko, Hanseong Roh, Noh Hyun Park, Cheol Lee, Se Ik Kim, Tae-You Kim. Clinical validation of WES-based HRD scoring and its prognostic value for PARP inhibitor maintenance therapy in ovarian cancer abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4019.