Does complete-staging lymphadenectomy improve survival outcomes in stage I endometrioid epithelial ovarian carcinoma? A multi-institutional retrospective study with propensity score-weighted analysis

Background: We investigated the prognostic impact of complete-staging lymphadenectomy on patients with clinically apparent Stage I endometrioid epithelial ovarian carcinoma. Methods: We conducted a regional multi-institutional retrospective study between 1986 and 2018. Amongst 4897 patients with malignant ovarian tumours diagnosed under central pathological review, 259 women with Stage I endometrioid epithelial ovarian carcinoma were eligible. We evaluated differences in survival of patients with both pelvic and para-aortic lymphadenectomy (Group A) and those with only pelvic lymphadenectomy and/or clinical lymph node evaluation (Group B). To analyse the therapeutic effects, the baseline imbalance between patients with both pelvic and para-aortic lymphadenectomy and others was adjusted with an inverse probability of treatment weighting using propensity score involving independent clinical variables. Results: In total, 145 patients (56.0%) received both pelvic and para-aortic lymphadenectomy. With propensity score-based adjustment, estimated survival was better in Group A compared with that in Group B but not significant. Pelvic and para-aortic lymphadenectomy also led to no significant improvement of overall survival in most of the subgroups. However, point estimations of the hazard ratio for lymphadenectomy in patients with an age of 45 or younger (hazard ratio, 0.304; 95% confidence interval, 0.094-0.982), a Grade 1-2 (hazard ratio, 0.441; 95% confidence interval, 0.204-0.954) and T1c2-3 tumour (hazard ratio, 0.449; 95% confidence interval, 0.164-1.231) were better compared with those with the opposite characteristics. Conclusions: Complete-staging lymphadenectomywas not a significant prognostic factor in patients with Stage I endometrioid epithelial ovarian carcinoma, where we still need to explore appropriate candidate for the procedure.