Comprehensive deep learning-based framework for automatic organs-at-risk segmentation in head-and-neck and pelvis for MR-guided radiation therapy planning

Introduction: The excellent soft-tissue contrast of magnetic resonance imaging (MRI) is appealing for delineation organs-at-risk (OARs) as it required radiation therapy planning (RTP). In the last decade there has been an increasing interest in using deep-learning (DL) techniques to shorten labor-intensive manual work and increase reproducibility. This paper focuses on automatic segmentation 27 head-and-neck 10 male pelvis OARs with methods based T2-weighted MR images. Method: proposed method uses 2D U-Nets localization 3D U-Net various structures. models were trained public private datasets evaluated only. Results discussion: Evaluation ground-truth contours demonstrated that can accurately segment majority indicated similar or superior performance state-of-the-art models. Furthermore, auto-contours visually rated by clinicians Likert score average, 81% them was found clinically acceptable.