Distribution of MRI-derived T2 values as a biomarker for in vivo rapid screening of phenotype severity in MDX mice

Abstract

Motivation: The frequently-used mdx model of Duchenne muscular dystrophy exhibits wide variation in disease severity, confounding detection of treatment effects. Goal(s): We sought to design a rapid, noninvasive imaging/analysis pipeline to prescreen animals and balance disease severity across treatment groups. Approach: Axial MR images and T2 maps were obtained in the hindlimbs of mdx and wildtype mice. A neural network was trained to speed segmentation. The distribution of muscle T2 values was analyzed. Results: Semiautomated segmentation reduced image processing time ~tenfold. Pearson Skew and interdecile range of muscle T2 distributions were repeatable indicators of muscle disease severity and correlated with Evans Blue dye uptake. Impact: Use of this rapid, non-invasive, semi-automated MRI/analysis pipeline has the potential to improve the sensitivity of preclinical treatment studies by enabling screening of dystrophic mice prior to study enrollment to increase uniformity of muscle pathology across treatment groups.