Propensity score matching confirms that primary surgery or neoadjuvant chemotherapy result in equivalent survival within a comprehensive cohort of patients with high-grade serous ovarian cancer.

Abstract Objective Our objective was to investigate whether trial evidence showing that neoadjuvant chemotherapy is non inferior to primary surgery for the primary treatment of advanced ovarian cancer could be extrapolated to groups of patients that were not included in the trials. Methods Using a detailed retrospective cohort of all patients managed through a single tertiary hospital we carried out a propensity score analysis, principal component analysis, and cox proportional hazard analysis to compare survival in matched cohorts. Results A propensity score analysis showed that for at least 41% of all patients with advanced high-grade serous cancer neoadjuvant chemotherapy is non inferior to primary surgery (median survival primary surgery: 38 months, neoadjuvant chemotherapy: 35 months. P = 0.39). However, principal component analysis, supported by cox modelling, suggests that for some subgroups, including patients with subdiaphragmatic nodal disease, primary surgery may be associated with improved survival (HR 0.11, CI 0.026–0.48). Conclusions We have shown that the findings of previous trials can be extrapolated to a wider population and that statistical modelling can be used to identify groups or patients who benefit from specific modalities of treatment.