Frontiers in Physics (Sep 2023)
Comprehensive deep learning-based framework for automatic organs-at-risk segmentation in head-and-neck and pelvis for MR-guided radiation therapy planning
- Vanda Czipczer,
- Bernadett Kolozsvári,
- Borbála Deák-Karancsi,
- Marta E. Capala,
- Rachel A. Pearson,
- Emőke Borzási,
- Zsófia Együd,
- Szilvia Gaál,
- Gyöngyi Kelemen,
- Renáta Kószó,
- Viktor Paczona,
- Zoltán Végváry,
- Zsófia Karancsi,
- Ádám Kékesi,
- Edina Czunyi,
- Blanka H. Irmai,
- Nóra G. Keresnyei,
- Petra Nagypál,
- Renáta Czabány,
- Bence Gyalai,
- Bulcsú P. Tass,
- Balázs Cziria,
- Cristina Cozzini,
- Lloyd Estkowsky,
- Lehel Ferenczi,
- András Frontó,
- Ross Maxwell,
- István Megyeri,
- Michael Mian,
- Tao Tan,
- Jonathan Wyatt,
- Florian Wiesinger,
- Katalin Hideghéty,
- Hazel McCallum,
- Steven F. Petit,
- László Ruskó
Affiliations
- Vanda Czipczer
- GE Healthcare, Budapest, Hungary
- Bernadett Kolozsvári
- GE Healthcare, Budapest, Hungary
- Borbála Deák-Karancsi
- GE Healthcare, Budapest, Hungary
- Marta E. Capala
- Department of Radiation Oncology, Erasmus MC Cancer Institute, Rotterdam, Netherlands
- Rachel A. Pearson
- Northern Institute for Cancer Research, Newcastle University, Newcastle, United Kingdom
- Emőke Borzási
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Zsófia Együd
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Szilvia Gaál
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Gyöngyi Kelemen
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Renáta Kószó
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Viktor Paczona
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Zoltán Végváry
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Zsófia Karancsi
- GE Healthcare, Budapest, Hungary
- Ádám Kékesi
- GE Healthcare, Budapest, Hungary
- Edina Czunyi
- GE Healthcare, Budapest, Hungary
- Blanka H. Irmai
- GE Healthcare, Budapest, Hungary
- Nóra G. Keresnyei
- GE Healthcare, Budapest, Hungary
- Petra Nagypál
- GE Healthcare, Budapest, Hungary
- Renáta Czabány
- GE Healthcare, Szeged, Hungary
- Bence Gyalai
- GE Healthcare, Szeged, Hungary
- Bulcsú P. Tass
- GE Healthcare, Budapest, Hungary
- Balázs Cziria
- GE Healthcare, Budapest, Hungary
- Cristina Cozzini
- GE Healthcare, Munich, Germany
- Lloyd Estkowsky
- GE Healthcare, Milwaukee, WI, United States
- Lehel Ferenczi
- GE Healthcare, Budapest, Hungary
- András Frontó
- GE Healthcare, Budapest, Hungary
- Ross Maxwell
- Northern Institute for Cancer Research, Newcastle University, Newcastle, United Kingdom
- István Megyeri
- GE Healthcare, Szeged, Hungary
- Michael Mian
- GE Healthcare, Milwaukee, WI, United States
- Tao Tan
- GE Healthcare, Eindhoven, Netherlands
- Jonathan Wyatt
- Northern Institute for Cancer Research, Newcastle University, Newcastle, United Kingdom
- Florian Wiesinger
- GE Healthcare, Munich, Germany
- Katalin Hideghéty
- Department of Oncotherapy, University of Szeged, Szeged, Hungary
- Hazel McCallum
- Northern Institute for Cancer Research, Newcastle University, Newcastle, United Kingdom
- Steven F. Petit
- Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, Netherlands
- László Ruskó
- GE Healthcare, Budapest, Hungary
- DOI
- https://doi.org/10.3389/fphy.2023.1236792
- Journal volume & issue
-
Vol. 11
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.
Keywords
