Risk factors and dose-effects for bladder fistula, bleeding and cystitis after radiotherapy with imaged-guided adaptive brachytherapy for cervical cancer: An EMBRACE analysis.

ABSTRACT

PURPOSE: To identify patient- and treatment-related risk factors for fistula, bleeding, cystitis, pain and difficulty in voiding in locally advanced cervical cancer patients treated with radio(chemo)therapy and image-guided adaptive brachytherapy (IGABT). MATERIAL AND METHODS: Morbidity within the EMBRACE-I study was prospectively reported for physician-assessed (CTCAE) fistula, bleeding and cystitis and patient-reported (EORTC) pain and difficulty in voiding. Analysis of risk factors was performed in patients without bladder infiltration. Risk factors were tested with Cox regression for grade (G) ≥ 3 cystitis, for G ≥ 2 fistula, bleeding and cystitis, and for EORTC "very much" and "quite a bit" or worse. RESULTS: Of 1416 patients enrolled, 1153 and 884 patients without bladder infiltration were evaluable for the analysis of CTCAE and EORTC items, respectively. Median follow-up was 48[3-120] months. Crude incidence rates for G ≥ 2 fistula, bleeding and cystitis were 0.7%, 2.7% and 8.8%, respectively, and 16% and 14% for "quite a bit" or worse pain and difficulty in voiding, respectively. Baseline urinary morbidity and overweight/obesity were significant risk factors for most endpoints. Bladder D2cm3 correlated with G ≥ 2 fistula, bleeding and cystitis, while ICRU bladder point dose correlated with EORTC pain "quite a bit" or worse. An increase from 75 Gy to 80 Gy in bladder D2cm3 resulted in an increase from 8% to 13% for 4-year actuarial estimate of G ≥ 2 cystitis. CONCLUSION: Clinical and treatment-related risk factors for bladder fistula, bleeding and cystitis were identified within a prospective and multi-institutional setting. A dose-effect was established with bladder D2cm3, reinforcing the importance of continued optimization during individualized IGABT planning.