Beam Angle Optimization for Double-Scattering Proton Delivery Technique Using an Eclipse Application Programming Interface and Convolutional Neural Network

To automatically identify optimal beam angles for proton therapy configured with the double-scattering delivery technique, a beam angle optimization method based on a convolutional neural network (BAODS-Net) is proposed. Fifty liver plans were used for training in BAODS-Net. To generate a sequence of input data, 25 rays on the eye view of the beam were determined per angle. Each ray collects nine features, including the normalized Hounsfield unit and the position information of eight structures per 2° of gantry angle. The outputs are a set of beam angle ranking scores (Sbeam) ranging from 0° to 359°, with a step size of 1°. Based on these input and output designs, BAODS-Net consists of eight convolution layers and four fully connected layers. To evaluate the plan qualities of deep-learning, equi-spaced, and clinical plans, we compared the performances of three types of loss functions and performed K-fold cross-validation (K = 5). For statistical analysis, the volumes V27Gy and V30Gy as well as the mean, minimum, and maximum doses were calculated for organs-at-risk by using a paired-samples t-test. As a result, smooth-L1 loss showed the best optimization performance. At the end of the training procedure, the mean squared errors between the reference and predicted Sbeam were 0.031, 0.011, and 0.004 for L1, L2, and smooth-L1 loss, respectively. In terms of the plan quality, statistically, PlanBAO has no significant difference from PlanClinic (P >.05). In our test, a deep-learning based beam angle optimization method for proton double-scattering treatments was developed and verified. Using Eclipse API and BAODS-Net, a plan with clinically acceptable quality was created within 5 min.

Comprehensive Output Estimation of Double Scattering Proton System with Analytical and Machine Learning Models

The beam output of double scattering proton system is difficult to be accurately modeled by treatment planning system (TPS). This study aims to design an empirical method using the analytical and machine learning (ML) models to estimate proton output in a double scattering proton system. Three analytical models and three ML models using Gaussian process regression (GPR) were generated on a training dataset consisting of 1544 clinical measurements, and the accuracy of each model was validated against additional 241 clinical measurements as testing dataset. Two most robust models (polynomial model and the ML GPR model with exponential kernel) were selected, and these two independent models agreed with less than 2% deviation using the testing dataset. The minimum number of samples needed for either model to achieve sufficient accuracy (± 3%) was determined by evaluating the mean average percentage error (MAPE) with increasing sample number, and the differences between the estimated outputs using the two models were also compared for 1000 proton beams with randomly generated range, and modulation for each option. These two models can be used as an independent output prediction tool for a double scattering proton beam, and a secondary output check tool for cross check between themselves.