AUTHOR=Czipczer Vanda , Kolozsvári Bernadett , Deák-Karancsi Borbála , Capala Marta E. , Pearson Rachel A. , Borzási Emőke , Együd Zsófia , Gaál Szilvia , Kelemen Gyöngyi , Kószó Renáta , Paczona Viktor , Végváry Zoltán , Karancsi Zsófia , Kékesi Ádám , Czunyi Edina , Irmai Blanka H. , Keresnyei Nóra G. , Nagypál Petra , Czabány Renáta , Gyalai Bence , Tass Bulcsú P. , Cziria Balázs , Cozzini Cristina , Estkowsky Lloyd , Ferenczi Lehel , Frontó András , Maxwell Ross , Megyeri István , Mian Michael , Tan Tao , Wyatt Jonathan , Wiesinger Florian , Hideghéty Katalin , McCallum Hazel , Petit Steven F. , Ruskó László
TITLE=Comprehensive deep learning-based framework for automatic organs-at-risk segmentation in head-and-neck and pelvis for MR-guided radiation therapy planning
JOURNAL=Frontiers in Physics
VOLUME=11
YEAR=2023
URL=https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2023.1236792
DOI=10.3389/fphy.2023.1236792
ISSN=2296-424X
ABSTRACT=Introduction: The excellent soft-tissue contrast of magnetic resonance imaging (MRI) is appealing for delineation of organs-at-risk (OARs) as it is required for radiation therapy planning (RTP). In the last decade there has been an increasing interest in using deep-learning (DL) techniques to shorten the labor-intensive manual work and increase reproducibility. This paper focuses on the automatic segmentation of 27 head-and-neck and 10 male pelvis OARs with deep-learning methods based on T2-weighted MR images.
Method: The proposed method uses 2D U-Nets for localization and 3D U-Net for segmentation of the various structures. The models were trained using public and private datasets and evaluated on private datasets only.
Results and discussion: Evaluation with ground-truth contours demonstrated that the proposed method can accurately segment the majority of OARs and indicated similar or superior performance to state-of-the-art models. Furthermore, the auto-contours were visually rated by clinicians using Likert score and on average, 81% of them was found clinically acceptable.