A Study of Automatic Planning of Esophageal Cancer Based on Dose Prediction and Beam Angle Optimization

Abstract

To propose an automatic planning method for esophageal cancer based on dose volume histogram prediction and beam angle optimization in the treatment planning system (TPS). 50 IMRT plans of esophageal cancer were selected as the training set to establish a dose prediction model for organ at risk. The gantry optimization function of the three-dimensional conformal radiotherapy (3D-CRT) module was used in the TPS to provide the beam angle for automatic planning. Another 20 testing plans were optimized in the TPS and manual method respectively, the dosimetric differences and planning efficiency between the two methods were statistically compared with paired t-test. The predicted dose of organs at risk fit well with real DVH parameters in automatic plans. There was no significant dosimetric difference observed in Dmax, Dmin, CI, HI of PTV, V5 of both lungs, the Dmax of spinal cord between automatic plans and manual plans. The V20 and Dmean of both lungs generated from automatic plans were reduced by 1.1%, 0.37Gy and 1.2%, 0.38Gy, respectively, compared with those in manual plans. The V30, V40 and Dmean of heart in automatic plans were decreased significantly by 5.1%, 3.0% and 1.41Gy, respectively, compared with those in manual plans. The labor time, computer working time, and MU number of automatic plans were decreased by 65.8%, 14.1%, and 17.2%, respectively, compared with manual planning. The automatic planning script improve the efficiency of esophageal cancer planning by dose prediction of organ at risk and beam angle optimization, and provides another alternative for esophageal cancer radiotherapy planning.