Abstract
OBJECTIVE: The study aimed to evaluate site-specific recurrent patterns via competing risks analysis and hazard function to provide evidence for adjuvant treatment and follow-up for early staged endometrial cancer (EC). METHODS: A total of 858 patients with International Federation of Gynecology and Obstetrics stage I-II EC who received adjuvant radiotherapy at our institution (2000-2017) were included. The radiotherapy modality comprised external beam radiotherapy (EBRT) with or without vaginal brachytherapy (VBT) or VBT alone. Competing risks analysis and hazard rate function were employed to evaluate the recurrence rate according to the ESMO-ESGO-ESTRO risk classification. RESULTS: The 5-year overall survival rates of the low-risk (LR), intermediate-risk (IR), high-intermediate risk (HIR), and high-risk (HR) groups were 96.1%, 95%, 93%, and 89.7%, respectively (p = 0.018). Sixty-eight patients developed recurrence. The 5-year incidence of distant recurrence was the highest in the HR group (14.87%), followed by the HIR (7.71%), IR (5.27%), and LR (1.26%) groups (Gray's test, p < 0.001). The LR and IR groups showed late metastasis behaviors for distant metastasis. The HR group presented a large magnitude of distant metastasis with an early peak that increased beyond 3 years. Subgroup analysis revealed that EBRT±VBT tended to reduce the locoregional relapse rate compared with VBT in the HIR-HR group (2.36% vs. 7.73%, Gray's test, p = 0.08). CONCLUSION: The established competing risk modeling demonstrated different recurrence patterns across the risk groups and radiotherapy modes. A better understanding of the change in site-specific recurrence behavior allows more targeted adjuvant treatment and surveillance regimens.