Generalisability of Automated CNN-based Renal Segmentation for Multi-Vendor Studies
Alexander J Daniel1, Charlotte E Buchanan1, David M Morris2, Hao Li3, Rebecca Noble1, João Sousa4, Steven Sourbron4, David L Thomas5,6,7, Andrew N Priest3,8, and Susan T Francis1
1Sir Peter Mansfield Imaging Centre, University of Nottingham, Nottingham, United Kingdom, 2Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, United Kingdom, 3Department of Radiology, University of Cambridge, Cambridge, United Kingdom, 4Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom, 5Neuroradiological Academic Unit, UCL Queen Square Institute of Neurology, University College London, London, United Kingdom, 6Dementia Research Centre, UCL Queen Square Institute of Neurology, University College London, London, United Kingdom, 7Wellcome Centre for Human Neuroimaging, UCL Queen Square Institute of Neurology, University College London, London, United Kingdom, 8Department of Radiology, Addenbrooke’s Hospital, Cambridge, United Kingdom
Manual segmentation of the kidneys is very time consuming and reader dependent, this renders measurements of total kidney volume (TKV) in large multi-site studies impractical. Here we use a convolutional neural network (CNN), trained on data from a single MRI vendor, to segment the kidneys of volunteers scanned with a harmonised FSE image protocol on MR scanners from three different vendors (GE, Philips and Siemens). The kidneys were manually segmented by two readers, both of which demonstrated a significant difference in TKV across vendors; no significant difference in TKV was found in the segmentations produced by the CNN.
This abstract and the presentation materials are available to members only;
a login is required.