Background: Friedreich Ataxia (FRDA) is the most common hereditary ataxia. Atrophy of the spinal cord is one of the hallmarks of the disease. Magnetic resonance imaging (MRI) and spectroscopy (MRS) are powerful and non-invasive tools to investigate pathological changes in the spinal cord. A handful of studies have reported cross-sectional alterations in FRDA using MRI and diffusion MRI (dMRI) in FRDA. However, to our knowledge no longitudinal MRI, dMRI or MRS results have been reported in the spinal cord in FRDA.
Objective: To investigate early-stage cross-sectional alterations and longitudinal changes in the cervical spinal cord in FRDA, using a multimodal magnetic resonance (MR) protocol comprising morphometric (anatomical MRI), microstructural (dMRI), and neurochemical (1H MRS) assessments.
Design: We enrolled 28 early-stage individuals with FRDA and 20 age- and gender-matched controls (cross-sectional study). Disease duration at baseline was 5.5±4.0 years and Friedreich Ataxia Rating Scale (FARS) total neurological score at baseline was 42.7±13.6. Twenty-one FRDA participants returned for 1-year follow-up, and 19 of those for 2-year follow-up (cohort study). Each visit consisted in clinical assessments and MR scans. Controls were scanned at baseline only.
Results: At baseline, individuals with FRDA had significantly lower spinal cord cross-sectional area (-31%, p=4.10-17), higher eccentricity (+10%, p=5.10-7), lower total N-acetyl-aspartate (-36%, p=6.10-9) and higher myo-inositol (+37%, p=2.10-6) corresponding to a lower ratio tNAA/mIns (-52%, p=2.10-13), lower fractional anisotropy (-24%, p=10-9) as well as higher radial diffusivity (+56%, p=2.10-9), mean diffusivity (+35%, p=10-8) and axial diffusivity (+17%, p=4.10-5) relative to controls.
Longitudinally, spinal cord cross-sectional area decreased by 2.4% per year relative to baseline (p=4.10-4), the ratio tNAA/mIns decreased by 5.8% per year (p=0.03), and fractional anisotropy showed a trend to decrease (-3.2% per year, p=0.08).
Spinal cord cross-sectional area correlated strongly with clinical measures, with the strongest correlation coefficients found between cross-sectional area and Scale for the Assessment and Rating of Ataxia (SARA) (R=-0.55, p=7.10-6) and between cross-sectional area and FARS total neurological score (R=-0.60, p=4.10-7). Less strong but still significant correlations were found for fractional anisotropy and tNAA/mIns.
Conclusion: We report here the first quantitative longitudinal MR results in the spinal cord in FRDA. The largest longitudinal effect size was found for spinal cord cross-sectional area, followed by tNAA/mIns and fractional anisotropy. Our results provide direct evidence that abnormalities in the spinal cord result not solely from hypoplasia, but also from neurodegeneration, and show that disease progression can be monitored non-invasively in the spinal cord.