Magnetization transfer imaging of the whole spinal cord in multiple sclerosis patients

International audienceContext: Magnetization transfer ratio (MTR) has shown promise to assess tissue microstructure modificationin MS patients. To date, such exploration has been limited to the brain and the cervical spinal cord (SC) portions of thecentral nervous system (CNS). Studying the MTR abnormalities in the whole SC could provide a better association withambulatory disability in MS patients.Objectives: i) to compare mean MTR values in MS patients and healthy controls (HC) according to the SC level; ii ) todescribe the link between MTR measurements at the cervical and thoracic SC levels; iii) to evaluate the link between MTRmeasures and disability according to the spinal cord level.Methods: 21 relapsing remitting MS (RRMS; median EDSS=2.5), 10 progressive MS patients (PMS; median EDSS=5.25) and13 HC were scanned on a 3T Siemens MRI scanner. The imaging protocol included 3 MT imaging acquisition slabs to coverthe whole SC. For each subject, MTR maps and vertebra labeling were computed using the SCT toolbox. MTR means werecomputed in semi-automatic delineated SC for the following vertebral levels: C4 to C6, T4 to T6, T9 to T10. Groupdifferences as well as correlations with lesions in the whole SC and EDSS were assessed controlling for age.Results: Evidence of group difference was only found in the cervical SC (C4C6; mean MTR=41.7pu, 39.4pu, 35.4pu for HC,RRMS and PMS resp.; p<.001). No evidence for group difference was found in the thoracic SC. A positive association wasfound between the mean MTR in the cervical SC and in the thoracic SC (r=.45, p=.01 for T4T6 and r=.54, p=.002 for T9T10)in MS patients. We observed negative associations between mean MTR in the cervical SC and the EDSS score (r=-.51,p=.004) and between mean MTR in the cervical SC and the SC lesion load (r=-.6, p<.001), while no clear evidence ofcorrelation was found between SC lesion load and EDSS score (r=.35; p=.084). No evidence of correlation was foundbetween mean MTR in the thoracic cord and EDSS score.Conclusions: The microstructural damage in the SC of MS patients seems to be predominant in the cervical SC and is linkedto the lesion load and the disability. In our sample data, the added value of exploring thoracic SC in addition to cervical SCusing MTR to explain disability in MS patients seems limited. Potential explanations could be the presence of highervariabilities in MTR measurement in the thoracic SC or the preferential location of MS lesions in the cervical SC