Deep learning in proton therapy : optimizing treatment dose prediction and assessing uncertainty

We developed a new uncertainty quantification method for deep learning regression models, based on Layer Ensembles [1], which is competitive with state-of-the-art ensembling and Monte Carlo (MC) dropout techniques. The method was implemented in a U-Net-like architecture and applied to predicting 3D dose and uncertainty map for head and neck cancer patients who are treated with proton therapy. The new approach runs approximately 8 times faster than MC Dropout (20 simulations). Our statistical analysis showed no significant difference in prediction accuracy between the two different methods (p-value = 0.09). Moreover, the correlation uncertainty/error in the body is only -3%. We also investigated the use of single-number metrics, such as the AULA score [1] and the mean value of the 3D uncertainty map, as a way to quantify the uncertainty of the prediction. These findings demonstrate the potential of the new method in enabling fast and accurate uncertainty quantification for regression problems and, in particular, for proton therapy dose prediction. [1] Kaisar Kushibar, Victor Campello, Lidia Garrucho, Akis Linardos, Petia Radeva, and Karim Lekadir. Layer ensembles: A single-pass uncertainty estimation in deep learning for segmentation. Lecture Notes in Computer Science, pages 514-524, 2022.