Meeting Banner
Abstract #1915

Segmentation of the Cortex and Medulla in Multiparametric Magnetic Resonance Images of the Kidney using K-Means Clustering

David M Morris1, Mhairi Donnelly2, Fiona J Gifford1, Philip Dunne2, Peter C Hayes2, Kenneth J Simpson2, Jonathan A Fallowfield1, and Scott IK Semple3

1MRC Centre for Inflammation Research, University of Edinburgh, Edinburgh Bioquarter, Edinburgh, United Kingdom, 2Liver Unit, Royal Infirmary of Edinburgh, Edinburgh Bioquarter, Edinburgh, United Kingdom, 3BHF Centre for Cardiovascular Science, University of Edinburgh, Edinburgh Bioquarter, Edinburgh, United Kingdom

Magnetic Resonance Imaging (MRI), including T1, T2*, Apparent diffusion coefficient (ADC) and Arterial Spin Labelling (ASL) perfusion, represents a powerful tool for renal investigations. This allows for simultaneous assessment of structure and function in pathologies from Acute Kidney Injury (AKI) to Chronic Kidney Dysfunction (CKD). Differentiation of the medulla and cortex is essential as these tissues have different biomarker distributions. Currently, segmentation of the biomarker histograms is carried out with manual definition of thresholds. Here, we applied K-means clustering to segment the maps and showed that this produced physiologically meaningful results while improving biomarker precision compared with whole kidney regions.

This abstract and the presentation materials are available to members only; a login is required.

Join Here