Results of radical radiotherapy for recurrent endometrial cancer.

OBJECTIVES: The aims of this study were to determine the overall survival (OS) and local control (LC) achieved in patients developing a locoregional recurrence of endometrial carcinoma and to define those prognostic factors that predict for improved LC and OS. METHODS: Between 1984 and 1988, 958 women were referred to Princess Margaret Hospital (PMH) with a diagnosis of endometrial carcinoma. Of these, 58 were treated for recurrent disease with radical radiotherapy (RT). Forty-two were referred with recurrence and 16 relapsed during follow-up at PMH for their primary tumor. None had received prior RT. The majority (n = 49) were treated with combined external beam RT followed by an intracavitary cesium insertion. RESULTS: The median time to relapse from original diagnosis was 1.3 years (range 0.2-13.4 years). The actuarial 5- and 10-year OS was 53 and 41%, respectively. The respective results for LC were 65 and 62%. All end-points were measured from the time of relapse. The median total dose received was 81.5 Gy. Univariate analysis showed that favorable histological features at original diagnosis (<50% myometrial involvement, grade 1-2, P = 0.007) and Perez modified staging (P = 0.02) were significant predictors for OS. The Perez staging (P = 0.02) and size of recurrence (<2 cm versus >/=2 cm, P = 0.04) were predictors for LC. CONCLUSION: Patients with localized relapse of endometrial carcinoma in whom radical radiotherapy can be administered should be treated aggressively and may be cured in over half the cases treated. Pathological findings in the original surgical specimen, size of recurrent disease, and a modified vaginal carcinoma staging system are significant predictors of local pelvic control and survival.