Meeting Banner
Abstract #3769

Automated Myocardial Segmentation for Quantitative Analysis of First-Pass Cardiac Perfusion MRI

Qi Duan1, Ricardo Otazo1, Daniel Kim1, Daniel K. Sodickson1

1Radiology, Center for Biomedical Imaging, NYU School of Medicine, New York, NY, USA


This abstract presents a fast automated segmentation of cardiac contours for quantitative analysis of first-pass cardiac perfusion MRI. The segmentation method was based on a previously developed novel segmentation framework for real-time segmentation, with an extension to vector images with multi-object segmentation. The myocardial segmentation and the resulting signal-time curves were in good agreement with those produced by manual tracing. The proposed method only takes 31ms to segment the myocardium from the multi-repetition data set.