Longitudinal Quantitative Neuromuscular MRI in Neuropathic Patients

In a context where numerous therapic strategies are being evaluated for the treatment of peripheral nervous system diseases, the clinical and electrophysiological scores currently available are proving inadequate for the detection of early change or a positive therapeutic effect. On the basis of a limited number of studies, quantitative MRI could provide much more sensitive therapeutic monitoring data over short periods of time which is crucial for future therapeutic trials.

The most interesting MRI biomarker to date would therefore be the FF, which represents the percentage of fatty infiltration of muscles following pathological nerve damage. Other MRI biomarkers, such as quantified magnetization transfer ratio (MTR), proton density (PD), water transverse relaxation time (wT2) and three-dimensional volume, enable us to study the degenerative and inflammatory phenomena at work in these neuropathies from different angles, detailing the nerve and muscle damage in patients compared with healthy controls or presymptomatic patients. In patients with Charcot-Marie-Tooth (CMT) neuropathy in particular, qMRI is the only tool to detect significant longitudinal variation over a one-year period, revealing an average increase in FF in the lower limbs of +1.5% over 12 months. Confirmation of these results and their extension to other neuropathies such as hereditary amyloid neuropathies (ATTR-PN) or acquired demyelinating neuropathies (ADN) is therefore justified, but requires the implementation of standardized longitudinal studies including these different pathologies.

As strong correlations between these MRI biomarkers and the main clinical scores have been demonstrated in several cross-sectional studies, the clinical value of this non-invasive tool is beyond doubt.

The main limitation to the clinical deployment of this technology remains the time required for the manual segmentation step to delineate the areas of interest. The ongoing development of new image analysis techniques, as well as the contribution of artificial intelligence and deep learning to the imaging data extraction process, are well on the way to solving this problem, but require continuous updating of practices to identify the most interesting MRI biomarkers, facilitate their extraction and thus measure their clinical application with a view to future therapeutic trials.