Denoising proton therapy Monte Carlo dose distributions in multiple tumor sites: A comparative neural networks architecture study.

Abstract

INTRODUCTION\nMonte Carlo (MC) algorithms provide accurate modeling of dose calculation by simulating the delivery and interaction of many particles through patient geometry. Fast MC simulations using large number of particles are desirable as they can lead to reliable clinical decisions. In this work, we assume that faster simulations with fewer particles can approximate slower ones by denoising them with deep learning.\n\n\nMATERIALS AND METHODS\nWe use mean squared error (MSE) as loss function to train networks (sNet and dUNet), with 2.5D and 3D setups considering volumes of 7 and 24 slices. Our models are trained on proton therapy MC dose distributions of six different tumor sites acquired from 50 patients. We provide networks with input MC dose distributions simulated using 1\xa0×\xa0106 particles while keeping 1\xa0×\xa0109 particles as reference.\n\n\nRESULTS\nOn average over 10 new patients with different tumor sites, in 2.5D and 3D, our models recover relative residual error on target volume, ΔD95TV of 0.67\xa0±\xa00.43% and 1.32\xa0±\xa00.87% for sNet vs. 0.83\xa0±\xa00.53% and 1.66\xa0±\xa00.98% for dUNet, compared to the noisy input at 12.40\xa0±\xa04.06%. Moreover, the denoising time for a dose distribution is: < 9s and \xa0<\xa01s for sNet vs. < 16s and \xa0<\xa01.5s for dUNet in 2.5D and 3D, in comparison to about 100\xa0min (MC simulation using 1\xa0×\xa0109 particles).\n\n\nCONCLUSION\nWe propose a fast framework that can successfully denoise MC dose distributions. Starting from MC doses with 1\xa0×\xa0106 particles only, the networks provide comparable results as MC doses with1\xa0×\xa0109 particles, reducing simulation time significantly.