Meeting Banner
Abstract #3620

T2 Relaxometry

Eve LoCastro1, Sneha N. Pandya1, Xiobo Shen2, Thanh Nguyen, PhD1, Susan A. Gauthier3, Ashish Raj1

1Department of Radiology, Weill Cornell Medical College, New York, NY, United States; 2Department of Computer Science, Cornell University, Ithaca, NY, United States; 3Department of Neurology and Neuroscience, Weill Cornell Medical College, New York, NY, United States

To introduce a new software pipeline for performing clinically feasible and routine quantitative myelin imaging via T2 Relaxometry, with specific application to demyelinating diseases like Multiple Sclerosis. T2 Relaxometry uses multi-echo T2weighted images to separate the contribution of various tissue components in the brain, thereby quantifying the myelin content. We have developed a fully automated pipeline in MATLAB featuring 3 steps: a) fast 3D T2-prepared spiral MR sequence for rapid multi-echo T2 imaging, b) a new post-processing technique obtaining T2 distributions and myelin fractions, and c) fully automated atlas-based coregistration, segmentation and parcellation pipeline to enable cross-subject voxel-based analysis. The overall pipeline is depicted in Fig 1. This analysis pipeline will be made publically available to facilitate routine myelin imaging and statistical analysis in common space.

Abstract PDF Presentation Video

Keywords

acknowledgment acquisition advance allows another applied atlas audience automated axial basic bottom brain bundles clearly clinical clinically college colors combined common components computations computed computer consistent constants constraints content convex cortical cost coverage cross custom defined definition delineation depicted described detailed details developed diagnostically difficult diseases distribution echoes enable enables entire excellent execution exploited exponentially facilitate fast feasible featuring form fraction fully generated great greatly histogram homogeneous hours implemented imposing improved improves indicates individual interpret intra introduce inverse investigations involved iterative launcher lesions linear maps medical model neurological neurologists neurology noise note novel numerical overall patient pipeline potential prep prepared problem processing proposed quantifying quantitative quite radiology rapid realm reduces regional registered relaxing restricting ribbon routine science sclerosis segmentation separate separated shaped since slices software solver space spaced sparsity spatial speed spiral statistical strength subject subjects supported target template thereby therein transformations unfortunately variation variations volumetric water white whole whose written