Prognostic assessment of simple hysterectomy versus radical/modified radical hysterectomy in clinically apparent stage II endometrial cancer: A multi-center cohort study.

5611 Background: The extent of hysterectomy for endometrial cancer (EC) patients with suspected or gross cervical involvement remains controversial. This study aimed to compare the effect of simple hysterectomy (SH) and radical or modified radical hysterectomy (RH/mRH) on the prognosis in clinically apparent stage II endometrial cancer, stratified by aggressive and non-aggressive histological types. Methods: This multi-center retrospective study enrolled patients with clinically apparent stage II (macroscopically cervical involvement without pelvic extension based on the preoperative imaging examination and intraoperative exploration) from the largest endometrial cancer database in China between 2000 and 2019. The primary and secondary outcomes were cancer-specific survival (CSS) and disease-free survival (DFS), respectively. The prognostic difference was assessed using Kaplan-Meier analysis and Cox regression analysis, in total cohort and stratified by aggressive and non-aggressive histological subtypes. Propensity score matching (PSM) algorithm were applied. Results: A total of eligible 725 participants were enrolled, including 245 patients with aggressive histological types and 480 with non-aggressive subtypes. The Kaplan-Meier showed better CSS and DFS in the RH/mRH group than the SH group both in the before ( P=0.0014 and P=0.0038, respectively) and after PSM cohorts ( P=0.042 and P=0.046, respectively) among patients with aggressive histological types. Nevertheless, no significant difference was found (P>0.05) among the patients with the non-aggressive subtypes. Similarly, the multivariate COX analysis revealed that hysterectomy type was an independent prognostic factor for CSS (hazard ratio HR RH/mRH vs SH : 0.36, 95% confidence interval CI 0.14–0.95, P=0.039) and DFS (HR RH/mRH vs SH : 0.47, 95% CI 0.24–0.93, P=0.030) in patients with aggressive histological types. Conclusions: Compared to SH, RH/mRH were associated with better CSS and DFS in patients with aggressive histological types, but not in the non-aggressive histological types. Despite guidelines recommend SH as first-line treatment for patients with suspected or gross cervical involvement, RH/mRH might be a more reasonable and individualized choice for those with aggressive histological subtypes.