Risk of Late Urinary Complications Following Image Guided Adaptive Brachytherapy for Locally Advanced Cervical Cancer: Refining Bladder Dose-Volume Parameters.

PURPOSE: To study correlations between dose-volume parameters of the whole bladder and bladder trigone and late urinary toxicity in locally advanced cervical cancer patients treated with pulsed-dose-rate brachytherapy. METHODS AND MATERIALS: Patients with locally advanced cervical cancer treated with chemoradiation therapy and pulsed-dose-rate brachytherapy from 2004 to 2015 were included. Cumulative dose-volume parameters of the whole bladder and bladder trigone were converted into 2-Gy/fraction equivalents (EQD2, with α/β = 3 Gy); these parameters, as well as clinical factors, were analyzed as predictors of toxicity in patients without local relapse.
RESULTS: A total of 297 patients fulfilled the inclusion criteria. The median follow-up period was 4.9 years (95% confidence interval 4.5-5.3 years). In patients without local relapse (n = 251), the Kaplan-Meier estimated grade 2 or higher urinary toxicity rates at 3 years and 5 years were 25.4% and 32.1%, respectively. Minimal dose to the most exposed 2 cm3 of the whole bladder [Formula: see text] , bladder International Commission on Radiation Units & Measurements (ICRU) (BICRU) dose, and trigone dose-volume parameters correlated with grade 2 or higher toxicity. At 3 years, the cumulative incidence of grade 2 or higher complications was 22.8% (standard error, 2.9%) for bladder [Formula: see text]  < 80 GyEQD2 versus 61.8% (standard error, 12.7%) for [Formula: see text]  ≥ 80 GyEQD2 (P = .001). In the subgroup of patients with bladder [Formula: see text]  ≤ 80 GyEQD2, a trigone dose delivered to 50% of the volume (D50%) > 60 GyEQD2 was significant for grade 2 or higher toxicity (P = .027). The probability of grade 3 or higher toxicities increased with bladder [Formula: see text]  > 80 GyEQD2 (16.7% vs 1.6%; hazard ratio [HR], 5.77; P = .039), BICRU dose > 65 GyEQD2 (4.9% vs 1.3%; HR, 6.36; P = .018), and trigone D50% > 60 GyEQD2 (3.1% vs 1.2%; HR, 6.29; P = .028). Pearson correlation coefficients showed a moderate correlation between bladder [Formula: see text] , BICRU dose, and bladder trigone D50% (P < .0001).
CONCLUSIONS: These data suggest that [Formula: see text]  ≤ 80 GyEQD2 should be advised for minimizing the risk of severe urinary complications (<15%). Bladder trigone dose was also predictive of severe late urinary toxicity. These constraints need further confirmation in a multicenter prospective setting.