Texture analysis of high-resolution MRI to assess muscle denervation in polyneuropathy

Motivation: Texture analysis of high-resolution fat-suppressed gradient echo may serve as monitoring biomarkers to detect the slow disease progression in polyneuropathy. Goal(s): To develop a radiomics method and evaluate texture features that correlate with CMT examination scores (CMTES), a clinical assessment of disease severity, and the Mercuri score, a radiological score of muscle denervation severity. Approach: Individual muscles at mid-thigh level of 67 participants were manually segmented. Texture features were extracted using the FAE software. Features were correlated with CMTES and the Mercuri score. Results: Certain features were highly correlated with the Mercuri score, and mildly correlated with CMTES. Impact: This study demonstrates a radiomics method for high-resolution leg muscle MRI, as well as the sensitivity of texture features to radiological and clinical severity. These texture features may be used as monitoring biomarker for disease progression in patients with polyneuropathy.