Rectal sequelae after conformal radiotherapy of prostate cancer: dose-volume histograms as predictive factors

Purpose: To identify clinically relevant parameters predictive of late rect al bleeding derived from cumulative dose-volume histograms (DVHs) of the re ctum after conformal radiotherapy of prostate cancer. Materials and methods: One hundred and nine patients treated with 3D confor mal radiotherapy between 1/1994 and 1/1996 for localized prostate cancer (c linical stage T1-T3) were available for analysis. All patients received a t otal dose of 66 Gy/2 Gy per fraction (specified at the International Commis sion on Radiation Units and Measurements ICRU reference point). DVHs of the contoured rectum were analyzed by defining the absolute (aV) and relative (rV) rectum volume that received more than 30% (V-30), 50% (V-50), 70% (V-7 0) 80% (V-80), 90% (V-90) and 100% (V-100) of the prescribed dose. Addition ally, a new aspect of DVH analysis was investigated by calculation of the a rea under the DVH-curve between several dose levels (area under the curve ( AUC)-DVH). DVH-variables were correlated with radiation side effects evalua ted in 3-6 months intervals and graded according to the EORTC/RTOG score. T he median follow-up was 30 months (12-60 months). Results: Univariate and multivariate stepwise Cox-Regression analysis inclu ding age, PTV, rectum size, rV(100), rV(90), rV(80), rV(70); rV(50) rV(30) and aV(30), to aV(100) were calculated. Late rectal bleeding (EORTC/RTOG gr ade 2) was significantly correlated with the percentage of rectum volume re ceiving greater than or equal to 90% of the prescribed dose (rV(90)) (P = 0 .007) and inversely correlated in a significant way with the size of contou red rectum (P = 0.006) in multivariate analysis. In our series, a proportio n of the rectum volume greater than or equal to 57% were included in the 90 %-isodose (rV(90) greater than or equal to 57%) in one half of the patients , with an actuarial incidence of 31% of late rectal bleeding at 3 years. In the other half of the patients, when rV(90) < 57%, the 3-year actuarial in cidence was 11% (P < 0.03). Conclusion: Our data demonstrate a dose-volume relationship at the referenc e dose of 60 Gy (similar to 90% of the prescribed dose) with respect to lat e rectal toxicity. The rV(90) seems to be the most useful and easily obtain ed parameter when comparing treatment plans to evaluate the risk of rectal morbidity. (C) 2001 Elsevier Science Ireland Ltd. All rights reserved.