Abstract
Purpose Deep learning-based knowledge-based planning (KBP) methods have been introduced for radiotherapy dose distribution prediction to reduce the planning time and maintain consistent high-quality plans. This paper presents a novel KBP model using an attention-gating mechanism and a three-dimensional (3D) U-Net for intensity-modulated radiation therapy (IMRT) 3D dose distribution prediction in head-and-neck cancer. Methods A total of 340 head-and-neck cancer plans, representing the OpenKBP-2020 AAPM Grand Challenge data set, were used in this study. All patients were treated with the IMRT technique and a dose prescription of 70 Gy. The data set was randomly divided into 64%/16%/20% as training/validation/testing cohorts. An attention-gated 3D U-Net architecture model was developed to predict full 3D dose distribution. The developed model was trained using the mean-squared error loss function, Adam optimization algorithm, a learning rate of 0.001, 120 epochs, and batch size of 4. In addition, a baseline U-Net model was also similarly trained for comparison. The model performance was evaluated on the testing data set by comparing the generated dose distributions against the ground-truth dose distributions using dose statistics and clinical dosimetric indices. Its performance was also compared to the baseline model and the reported results of other deep learning-based dose prediction models. Results The proposed attention-gated 3D U-Net model showed high capability in accurately predicting 3D dose distributions that closely replicated the ground-truth dose distributions of 68 plans in the test set. The average value of the mean absolute dose error was 2.972 +/- 1.220 Gy (vs. 2.920 +/- 1.476 Gy for a baseline U-Net) in the brainstem, 4.243 +/- 1.791 Gy (vs. 4.530 +/- 2.295 Gy for a baseline U-Net) in the left parotid, 4.622 +/- 1.975 Gy (vs. 4.223 +/- 1.816 Gy for a baseline U-Net) in the right parotid, 3.346 +/- 1.198 Gy (vs. 2.958 +/- 0.888 Gy for a baseline U-Net) in the spinal cord, 6.582 +/- 3.748 Gy (vs. 5.114 +/- 2.098 Gy for a baseline U-Net) in the esophagus, 4.756 +/- 1.560 Gy (vs. 4.992 +/- 2.030 Gy for a baseline U-Net) in the mandible, 4.501 +/- 1.784 Gy (vs. 4.925 +/- 2.347 Gy for a baseline U-Net) in the larynx, 2.494 +/- 0.953 Gy (vs. 2.648 +/- 1.247 Gy for a baseline U-Net) in the PTV_70, and 2.432 +/- 2.272 Gy (vs. 2.811 +/- 2.896 Gy for a baseline U-Net) in the body contour. The average difference in predicting the D-99 value for the targets (PTV_70, PTV_63, and PTV_56) was 2.50 +/- 1.77 Gy. For the organs at risk, the average difference in predicting the Dmax${D_{max}}$ (brainstem, spinal cord, and mandible) and Dmean${D_{mean}}$ (left parotid, right parotid, esophagus, and larynx) values was 1.43 +/- 1.01 and 2.44 +/- 1.73 Gy, respectively. The average value of the homogeneity index was 7.99 +/- 1.45 for the predicted plans versus 5.74 +/- 2.
95 for the ground-truth plans, whereas the average value of the conformity index was 0.63 +/- 0.17 for the predicted plans versus 0.89 +/- 0.19 for the ground-truth plans. The proposed model needs less than 5 s to predict a full 3D dose distribution of 64 x 64 x 64 voxels for a new patient that is sufficient for real-time applications. Conclusions The attention-gated 3D U-Net model demonstrated a capability in predicting accurate 3D dose distributions for head-and-neck IMRT plans with consistent quality. The prediction performance of the proposed model was overall superior to a baseline standard U-Net model, and it was also competitive to the performance of the best state-of-the-art dose prediction method reported in the literature. The proposed model could be used to obtain dose distributions for decision-making before planning, quality assurance of planning, and guiding-automated planning for improved plan consistency, quality, and planning efficiency.