Significant correlation between rectal DVH and late bleeding in patients treated after radical prostatectomy with conformal or conventional radiotherapy (66.6 –70.2 Gy

Abstract

Purpose/Objective: Investigating the correlation between dosimetric/clinical parameters and late rectal bleeding in patients treated with adjuvant or salvage radiotherapy after radical prostatectomy. Materials/Methods: Data of 128 consecutive patients postoperatively irradiated at San Raffaele H Scientific Institute during the period 1994–2000 (36 months median follow up; range: 12–72 months) including 3D treatment planning and dose-volume-histograms (DVHs) of the rectum were retrospectively analyzed. All patients were treated at a nominal dose of 66.6–70.2 Gy (1.8 Gy/fraction, ICRU dose: 68–72.5 Gy, median: 70 Gy, ) with conformal (n = 76) or conventional (n = 52) 4-fields 18 MV X-Rays techniques. A number of clinical parameters such as diabetes mellitus, acute rectal bleeding, hypertension, age and hormonal therapy were considered. Late rectal bleeding was scored using a modified RTOG scale and patients experiencing grade ≥ 2 toxicity were considered as bleeders. Mean and median rectal dose were considered together with rectal volume and the% fraction of rectum receiving more than 50, 55, 60, 65 Gy (V50, V55, V60, V65, respectively). Median and quartile values of all parameters were taken as cut-off for statistical analysis. Univariate (Log-rank) and multivariate (Cox hazard model) analyses were performed. Results: 14/128 patients experienced grade ≥ 2 late bleeding (3-year actuarial incidence: 10.5%). A significant correlation between a number of cut-off values and late rectum bleeding was found. In particular, a mean dose ≥ 54 Gy, V50 ≥ 63%, V55 ≥ 57%, V60 ≥ 50% were highly predictive of late bleeding (p = 0.01). A rectum volume < 60 cc and the type of treatment (conventional vs. conformal) were also significantly predictive of late bleeding (p = 0.05). Concerning clinical variables, acute bleeding (p < 0.001) was significantly related with late bleeding and a trend was found for hypertension (p = 0.11). Once grouped the patients in those with V50 ≥ 63% and in those with V50 < 63% (see Fig 1) data were fit with a Cox regression hazard model using DVH grouping, rectal volume and the main clinical parameters as independent variables. Results of the analysis showed that DVH grouping (RR: 3.3; p = 0.04) and acute bleeding (RR: 7.1; p = 0.001) are independently predictive of late bleeding. Conclusions: Rectal DVHs and dose statistics are significantly correlated with late bleeding for patients irradiated at 66.6–70.2 Gy after radical prostatectomy. In particular, the results of this analysis confirm the predictive value of V50 which was previously found to be strongly correlated with late bleeding in a multi-centric retrospective study on a much wider and more heterogeneous population of patients.