Abstract:Background Clinical decision-making in spinocerebellar ataxia spectrum diseases (SCAs) has mainly been based on genetic tests, not considering the SCAs' imaging and clinical heterogenicity. Objective To identify SCAs phenogroups by analysis and hierarchical clustering of infratentorial morphological MRI for unveiling pathophysiological differences among common SCA subtypes. Methods We prospectively enrolled 119 (62 women; mean age 37 years) genetically diagnosed SCAs (SCA1 n = 21, SCA2 n = 10, symptomatic SCA3 n = 59, presymptomatic SCA3 n = 22, SCA6 n = 7) and 35 healthy controls (HCs). All patients underwent MRI and detailed neurological and neuropsychology examinations. The width of each cerebellar peduncle (CP) and anteroposterior diameter of the spinal cord and pontine were measured. Twenty-five SCAs patients (15 women; mean age 35 years) were followed for at least a year (17 (15, 24) months), whose MRI and the Scale for the Assessment and Rating of Ataxia (SARA) were collected. Results Infratentorial morphological MRI measurements could significantly discriminate SCAs from HCs, even among SCA subtypes. Two mutually exclusive and clinically distinct phenogroups were identified. Despite similar (CAG) n , phenogroup 1 (n = 66, 55.5%) presented more atrophied infratentorial brain structures and more severe clinical symptoms with older age and earlier age of onset when compared with phenogroup 2. More importantly, all SCA2, most of SCA1 (76%), and symptomatic SCA3 (68%) were classified into phenogroup 1, whereas all SCA6 and all presymptomatic SCA3 were in phenogroup 2. The right middle CP had the highest diagnostic value in predicting phenogroup 2 (AUC = 0.99; P < 0.01) with high specificity (95%). Consistent with the significantly increased SARA (7.5 vs 10, P = 0.021), the bilateral inferior CP, spinal cord, and pontine tegmentum were more atrophy during the follow-up ( P < 0.05). Conclusion SCAs were with significant infratentorial brain atrophy than HCs. We identified two different SCAs phenogroups associated with substantial differences in infratentorial brain atrophy, clinical presentation, and may reflect the underlying molecular profiles to some extent, paving the way for a more personalized diagnostic and treatment approach.

Analysis and hierarchical clustering of infratentorial morphological MRI identifies SCAs phenogroups