Abstract
BACKGROUND: Differentiating demyelinating from axonal polyneuropathies is essential for accurate diagnosis and treatment. We hypothesized that multiparametric quantitative MRI (qMRI) of peripheral nerves can differentiate demyelination from axonal loss. This retrospective study leveraged genetically defined demyelinating and axonal polyneuropathies to test this concept. METHODS: Multiparametric qMRI data of proximal (sciatic) and distal (tibial) nerves were acquired on 3T MRI, including magnetization transfer ratio (MTR), MT saturation index (MTsat), T(2) (*), T(1), proton density (PD), fractional anisotropy (FA), mean/axial/radial diffusivities (MD, AD, RD), and fascicular volume (fVol). Data were analyzed from patients with Charcot-Marie-Tooth type 1 (CMT1, de-/dys-myelinating, n=19), CMT2 (axonal, n=12), hereditary neuropathy with liability to pressure palsies (HNPP, a cohort who often has intermediate changes between the two classifications, n=25), and health controls (HC, n=25). A composite qMRI score, as CMT Imaging Score (CMTIS), was developed to predict disease severity using the CMT Neuropathy Score version-2 (CMTNSv2) as a clinical reference. Receiver operating characteristic (ROC) analyses assessed diagnostic performance. RESULTS: CMT1 showed significantly increased fVol versus HCs, while CMT2 demonstrated reduced T(2) (*). Both CMT1 and CMT2 exhibited reduced FA, MTsat, and AD, along with elevated T(1) and RD, with larger abnormalities in CMT1. ROC analyses demonstrated strong discrimination of CMT1 and CMT2 (AUCs: 0.95 and 0.85 for sciatic; 0.89 and 0.73 for tibial nerves). CMTIS correlated strongly with CMTNSv2 (r=0.67 sciatic; r=0.72 tibial; r=0.79 combined). CONCLUSIONS: Multiparametric qMRI identifies distinct imaging signatures of demyelinating versus axonal hereditary polyneuropathies. The CMTIS shows strong potential as a biomarker for disease monitoring.