Prediction model of local effect model-based rectal dose volume histogram in prostate cancer patients treated with carbon ion radiotherapy / 中华放射肿瘤学杂志

ABSTRACT

Objective:To establish a local effect model (LEM)-based rectal dose volume histogram (DVH) prediction model in prostate cancer patients treated with carbon ion therapy based on Japanese experience, aiming to provide reference for clinically reducing the incidence of rectal adverse reactions.Methods:The planning CT data of 76 patients with prostate cancer were collected. The microdosimetric kinetic model (MKM) was used for initial planning, and the LEM was selected to recalculate the biological dose based on the same fields to MKM. Then, the geometric features and DVH of the rectum were extracted from the LEM plans. The planning data of 61 cases were used to establish the prediction model with linear regression and the other 15 cases were used for validation.Results:The ratio of the overlapped volume between the rectum and the region of interest (ROI) expended from planning target volume by 1 cm along the left and right directions of the rectum could be proved to be the characteristic parameters for linear regression. The mean goodness-of-fit R2 of predicted and LEM plan-based DVH of 15 cases was 0.964. The results of predicted rectal adverse reactions based on predicted DVH were consistent with those of LEM plan-based DVH. Conclusions:The linear regression method used in this study can establish an accurate prediction model of rectal DVH, which may provide certain reference for reducing the incidence of rectal adverse reactions. Nevertheless, the findings remain to be further verified by clinical trials with larger sample size.