Author: Ahssan Balawi, Peter Jermain, Timothy Kearney, Sonali Rudra, Michael H. Shang, Markus Wells, Mohammad Zarenia 👨🔬
Affiliation: Department of Radiation Medicine, MedStar Georgetown University Hospital 🌍
Purpose: To investigate the applicability and accuracy of a deep learning (DL) model in predicting radiation dose distribution for breast cancer patients treated with pencil-beam-scanning proton radiotherapy.
Methods: A total of 102 proton therapy clinical plans generated by the treatment planning staff for breast cancer patients, previously treated at our clinic using the MEVION S250i pencil beam proton system, were used to construct a plan database, of which 92 plans were used to train and 10 plans to validate a 3D dense U-Net machine learning model for predicting 3D dose distributions. Inputs to the model include dose prescriptions, target and OAR contours. The model performance was assessed by comparing the DVH values of the predicted dose distributions with that of the clinical plans for 15 separate breast cancer treatment plans.
Results: The predicted dose distributions in the test set demonstrated comparable quality to the clinical plans. The average value of the mean absolute dose difference was 1.56 ± 0.29 Gy. The percentage dose differences in the predicted target metrics of D5% and D95% were 3.02% ± 2.02% and 7.29% ± 4.15%, respectively. For the OARs, the predicted mean and maximum dose differences were Dmean=1.23 ± 0.52 Gy and Dmax =5.02 ± 2.04 Gy for the heart, and Dmean=1.36 ± 1.01 Gy and Dmax=1.66 ± 0.83 Gy for the lung.
Conclusion: We developed 3D DL models capable of rapidly predicting high-quality dose distributions for breast cancer proton therapy. The predicted dose distributions can be imported into a treatment planning system for generation of treatment plans with the use of scripting functions for beam placement and plan parameter optimization that produce the predicted dose distributions. The proposed DL-based planning approach can be used for decision-making before planning, individualized assessment of plan quality, and guiding automated planning to improve planning consistency and efficiency.