PARPi maintenance may attenuate the adverse prognosis of residual disease after debulking surgery in HRD-positive advanced epithelial ovarian cancer: an exploratory real-world analysis

This study aims to identify independent predictors of survival in HRD-positive advanced epithelial ovarian cancer (EOC) and to examine the real-world association of PARPi maintenance therapy and surgical cytoreduction outcomes with survival. We reviewed records of 70 HRD-positive patients with FIGO stage III or IV EOC treated at Kaohsiung Chang Gung Memorial Hospital from January 2021 to December 2024. HRD status was confirmed by validated next-generation sequencing as part of the CareHRD Project in Taiwan. The optimal post-chemotherapy CA-125 cutoff for predicting disease progression was identified using ROC analysis with Youden’s index. Multivariate Cox proportional hazards regression was performed to identify independent predictors of survival outcomes. An exploratory four-group cross-tabulation (PARPi ± × Optimal debulking ±) with pairwise comparisons and a formal Cox interaction test was conducted to explore the joint association of PARPi maintenance and surgical cytoreduction with survival. Of the 70 HRD-positive patients, 52 (74.3%) received PARPi maintenance, and 18 (25.7%) did not. The median follow-up was 34.1 months. In multivariate analysis, three factors were independently associated with progression-free survival (PFS) (C-index 0.744): PARPi maintenance (HR 0.280, 95% CI 0.122–0.645, p = 0.003), post-chemotherapy CA125 ≥ 10.9 U/mL (HR 3.328, 95% CI 1.365–8.115, p = 0.008), and neoadjuvant chemotherapy (HR 2.899, 95% CI 1.224–6.867, p = 0.016). Optimal surgical cytoreduction was significant in univariate but not multivariate analysis ( p = 0.269). In the exploratory four-group analysis (overall log-rank p = 0.007), the best outcomes were observed in PARPi+/Optimal+ patients (median PFS 44.6 months), and the worst in PARPi−/Optimal− patients (median PFS 9.9 months). Neither a formal Cox interaction between PARPi and cytoreduction nor the pairwise difference between PARPi+/Optimal− (23.1 months) and PARPi−/Optimal+ (28.0 months) was significant. Among patients with residual disease after surgery (R1/R2), PARPi maintenance was associated with markedly longer PFS than no maintenance (median 24.4 vs. 8.4 months; log-rank p < 0.001). The CA125 cutoff of 10.9 U/mL was very close to the previously reported 10.7 U/mL, indicating consistency across groups. Among HRD-positive patients, PARPi maintenance was independently associated with improved PFS, whereas the independent prognostic contribution of optimal cytoreduction was attenuated after adjustment. Most notably, among patients with residual disease after surgery, PARPi maintenance was associated with substantially longer PFS, suggesting that it may partially attenuate—rather than neutralize or replace—the adverse prognosis of incomplete cytoreduction. Post-chemotherapy CA125 was confirmed as a reliable risk marker. Further studies with longer follow-up are needed to confirm these results, particularly in patients in whom optimal cytoreduction cannot be achieved.