Left atrial wall segmentation using clinically correlated metrics
Document Type
Conference Proceeding
Publication Date
1-1-2017
Journal
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume
10124 LNCS
First Page
201
Last Page
210
URL with Digital Object Identifier
10.1007/978-3-319-52718-5_22
Abstract
The thickness of the left atrium wall may be an important parameter in atrial fibrillation disease mechanisms and subsequent treatment by catheter ablation. We have previously developed a simple, threshold-based, direct wall thickness measure from CT that has been found to correlate with clinical out- comes. In this paper, we describe the application of this method to the seg- mentation of the left atrium wall in the 2016 STACOM Left Atrium Wall Thickness Challenge. Our original method sought to partially automate the way a clinical researcher manually measures left atrial wall thickness to increase precision and repeatability. We have adapted our method to create a segmented volume instead of individual measurements in order to meet the challenge goals. We apply the method to the ten contrast-enhanced CT images provided.