Dose Distribution Prediction for Optimal Treamtment of Modern External Beam Radiation Therapy for Nasopharyngeal Carcinoma

摘要

In Intensity-modulated radiation therapy, the planning of the optimal dose distribution for a patient is a complex and time-consuming process. This paper proposes a new automatic method for predicting of dose distribution of Nasopharyngeal carcinoma (NPC) from contoured computer tomography (CT) images. The proposed method consists of two phases: (1) predicting the 2D optimal dose images of each beam from contoured CT images of a patient by convolutional deep neural network model, called OTNet, and (2) integrating the optimal dose images of all the beams to predict the dose distribution for the patient. From the experiments using CT images of 80 NPC patients, our proposed method achieves a good performance for predicting dose distribution compared with conventional predicted dose distribution methods.