Neurochemical Endpoints to Inform Early‐Stage Trials of Spinocerebellar Ataxia 2 and 3 in a Multisite Setting

OBJECTIVE: Neurochemical levels measured by brain MR spectroscopy (MRS) have been proposed as endpoints for clinical trials in early-stage spinocerebellar ataxia (SCA) trials. We tested their trial-readiness by quantifying neurochemicals in three affected brain regions in early-stage cohorts of SCA2 and SCA3, examining their reproducibility in both patients and controls in a multisite/multivendor setting, and determining if the same neurochemical outcomes can be used as endpoints for different SCA genotypes. METHODS: Seventy-two participants (18 SCA2, mean Scale for the Assessment and Rating of Ataxia (SARA) = 9 +/- 5; 25 SCA3, mean SARA = 10 +/- 6; 29 controls) underwent 3 T scans at four sites using scanners from two vendors and a harmonized MRS protocol. Scans were performed twice with a maximum 2-week interval. We measured neurochemical concentrations in the cerebellum, pons, and putamen, and assessed their within- and between-site reproducibility and associations with clinician- and patient-reported outcomes. RESULTS: Test-retest coefficients of variance were comparable across sites and vendors, but concentration estimates had to be adjusted for sites to account for small vendor-based biases. Total N-acetylaspartate (tNAA) and glutamate were lower, and myo-inositol (mI) and total creatine (tCr) were higher in the cerebellum and pons in both SCAs vs. controls. Pontine tNAA and mI, and cerebellar tNAA, mI, and glutamate levels were significantly different between SCA2 and SCA3, with strong effect sizes (|Cohen's d| = 1.1-2.1). The tNAA/tCr ratio had the strongest associations with clinical outcomes (|r| ≥ 0.5), with stronger associations in the cerebellum for SCA2 and in the pons for SCA3. INTERPRETATION: These data support the use of tNAA and its ratios in a genotype-specific manner in early-stage SCA2 and SCA3 trials in the multisite/multivendor setting.