This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/mp.15290. This article is protected by copyright. All rights reserved. Title: Cascaded deep-learning based auto-segmentation for head and neck cancer patients: organs at risk on T2 weighted magnetic resonance imaging Running Title: MRI DL auto-segmentation of OAR in HNC Authors & Affiliations: James C Korte 1,2 , Nicholas Hardcastle 1,3,4 , Sweet Ping Ng, 5,6 , Brett Clark 1,2 , Tomas Kron 1,4 , Price Jackson 1,4 1 Department of Physical Science, Peter MacCallum Cancer Centre, Melbourne, Australia 2 Department of Biomedical Engineering, University of Melbourne, Melbourne, Australia 3 Centre for Medical Radiation Physics, University of Wollongong, Wollongong, Australia 4 Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, Australia 5 Department of Radiation Oncology, Peter MacCallum Cancer Centre, Melbourne, Australia 6 Department of Radiation Oncology, Olivia Newton-John Cancer and Wellness Centre, Austin Health, Melbourne, Australia Corresponding Author: Dr James C Korte James.Korte@petermac.org Peter MacCallum Cancer Centre 305 Grattan Street Melbourne, Victoria 3000 Australia Previous publication of manuscript text or data: Early results of the low-resolution auto-segmentation method from this paper were presented at the Australian Magnetic Resonance in Radiation Therapy annual meeting in 2019.