Abstract #1052

Comparison of Automated Cerebral Infarct Segmentation Techniques using DWI and FLAIR MRI

Ryan A. Rava^1,2, Muhammad Waqas^2,3, Kenneth V. Snyder^2,3, Elad I. Levy^2,3, Adnan H. Siddiqui^2,3, Jason M. Davies^2,3, Xiaoliang Zhang¹, and Ciprian N. Ionita^1,2,3

¹Biomedical Engineering, University at Buffalo, Buffalo, NY, United States, ²Canon Stroke and Vascular Research Center, University at Buffalo, Buffalo, NY, United States, ³Neurosurgery, University at Buffalo, Buffalo, NY, United States

FLAIR MRI has the potential to provide more accurate ground truth infarct labels compared with DWI for the purpose of software validation and determination of ischemic stroke patient eligibility for thrombectomy. Currently, accurate segmentation of infarct has hindered the use of FLAIR infarct labels due to skull and erroneous image intensity values being similar to those of infarct lesions. In this study, an automated segmentation technique was developed for segmentation of infarct tissue from FLAIR MRI and performance metrics comparing this method to manually segmented infarct (FLAIR Sorenson-Dice=0.8168, DWI Sorenson-Dice=0.7922) indicate this technique is non-inferior to the current standard (DWI).

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

Join Here