NMR (MRI and MRS) as an outcome measure

With the development of quantitative approaches in imaging techniques for neuromuscular diseases, the possibility of using magnetic resonance imaging and spectroscopy as a non-invasive outcome measure to assess treatment response is becoming a reality. For this to be a realistic option in future trials, cross-centre comparability of results will be a prerequisite. In work led by Pierre Carlier's group at the Institut de Myologie, standard operating procedures (SOPs) for certain quantitative techniques have been developed by a multinational (and multi-platform) working group. A manuscript has been prepared for publication, and SOPs for the following techniques:

• Acquisition of T1w images for qualitative evaluation of muscle involvement and topography of lesions in neuromuscular diseases
• Quantification of muscle fat infiltration in neuromuscular diseases using Nuclear Magnetic Resonance imaging 3‐point Dixon or spectrally selective excitation techniques.
• Quantitative evaluation of muscle edema or inflammation in neuromuscular diseases using nuclear magnetic resonance T2 measurement techniques will appear here on the TREAT-NMD website in the near future.

The process of developing these SOPs and the collaborative work that has taken place as part of a multicentre MRI study of patients with LGMD2I (FKRP mutation) has enabled an overall consensus to be reached on the ideal protocol to apply in clinical studies to obtain quantitative results from MRI techniques, and it was therefore possible to use this work to form the foundations of the MRI section of a study protocol for a major multicentre dysferlinopathy project that is currently under funding review.

After the definition of the principal NMRI outcome measures was achieved, much optimization work had still to be performed, in order to obtain truly quantitative measurements. Post-processing procedures for B1 receiver inhomogeneities were designed and validated and are proposed as add-on for Dixon imaging. Fat contamination is a major source of bias for skeletal muscle T2 determination. A relatively simple and easily implementable algorithm has been developed, which realizes a deconvolution of the fat and water signal. B1 transmit misadjustments are sources of error in T2 determination. A method for selection of voxels with adequate magnetization nutation has been implemented and allows normal skeletal muscle T2 values to be ascribed within a 4ms range. All these optimizations are described in the SOPs and in the manuscript to be submitted for publication.