OC-0549: Effective alpha/beta concept: normal-tissue fractionation accounting for heterogeneous dose and volume effect

A.L. Hoffmann, A.E. Nahum
2013 Radiotherapy and Oncology  
Purpose/Objective: Many dose-limiting normal tissues in radiotherapy display considerable internal organ motion during a course of treatment, potentially deteriorating the relations between dosevolume parameters and morbidity. To address this issue, we have previously developed a motion simulation model and validated and applied it on the rectum geometries in the planning CT scans in a cohort of prostate cancer patients. In this present study we apply this model to compare the associations with
more » ... e associations with morbidity of the planned vs. motion-inclusive dose distributions in two cohorts of patients prospectively followed with respect to a broader spectrum of rectal morbidity endpoints. Materials and Methods: The patients in the two included cohorts (Cohort I: 192 patients; Cohort II:87 patients) were previously treated with RT for localised prostate cancer to prescribed doses of 70 vs. 74 Gy. The dose-response relations for the planned as well as simulated motion-inclusive rectal dose-volume histograms (DVHs) were associated with six rectal morbidity endpoints (Cohort I: RTOG gastrointestinal (GI) toxicity, stool frequency,mucosal loss, tenesmus, sphincter control and late rectal bleeding (LRB); Cohort II: LRB). The motion-inclusive dose distributions were obtained from simulations of small, moderate and large motion magnitudes applied to the planned dose distributions, using the previously developed and validated model.Additionally the data was evaluated with a probit NTCP model using the QUANTEC recommended rectal parameters (TD50=76.9, m=0.13, n=0.09). Results: Statistically significant associations with the studied endpoints were obtained at high doses (> 55Gy) using both the planned and the simulated motion-inclusive dose distributions, with the strongest associations observed for RTOG GI toxicity and stool frequency (Spearman's rank correlation coefficient, Rs=0.13-0.21 and Rs=0.14-0.16). Within the high dose region the associations with morbidity using the planned or the motion-inclusive DVHs with either small or moderate motion magnitudes were stronger than the associations seen when simulating large motion. Using the probit NTCP model the strongest associations were observed for tenesmus (Rs=0.14, p=0.03) and LRB (Rs=0.12, p=0.13). Conclusions: Equally strong associations with rectal morbidity were observed for high doses (>55 Gy), for the planned dose distributions and the simulated dose distributions including small and moderate rectal motion. Using a probit model with the QUANTEC recommended parameters the strongest significant associations were obtained for tenesmus.
doi:10.1016/s0167-8140(15)32855-3 fatcat:e7qkbxpadzf4pijzknhibyxrsi