Meeting Banner
Abstract #0088

Machine learning for automatic three-dimensional segmentation of the aorta in 4D flow MRI

Martijn Froeling1, Emile S. Farag2, R. Nils Planken3, Tim Leiner1, and Pim van Ooij3

1Department of Radiology, University Medical Center Utrecht, Utrecht, Netherlands, 2Department of Cardiothoracic Surgery, Amsterdam University Medical Centers, AMC, Amsterdam, Netherlands, 3Department of Radiology & Nuclear Medicine, Amsterdam University Medical Centers, AMC, Amsterdam, Netherlands

In this study we present a machine learning convolutional neural network (CNN) for automatic segmentation of the aorta used for peak systolic wall shear stress (WSS) assessment from 4D flow MRI data. The automated three-dimensional WSS profiles (WSSMACHINE) were compared with WSS calculated using manually (WSSMAN) created segmentations. Bland-Altman and orthogonal regression analysis revealed good agreement between WSSMAN and WSSMACHINE in terms of small mean differences and slopes and intercepts close to unity and zero respectively. The CNN has the ability to drastically accelerate aortic segmentation from 4D flow MRI data, which will greatly improve the clinical applicability of WSS.

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

Join Here