Disease-Specific Metabolic Patterns in Progressive Supranuclear Palsy with Richardson's Syndrome and Variants: A Longitudinal Study.

BACKGROUND

Progressive supranuclear palsy (PSP) is a rare tauopathy with multiple clinical phenotypes. The PSP-related metabolic pattern (PSPRP) has proved useful for differentiating PSP from other parkinsonian syndromes, but type-specific metabolic patterns remain incompletely characterized.

OBJECTIVE

This study aimed to identify type-specific metabolic patterns in PSP and evaluate their ability to monitor progression and predict conversion of variant PSP (vPSP) to PSP with Richardson's syndrome (PSP-RS).

METHODS

PSP-RS, PSP with predominant parkinsonism (PSP-P), PSP with predominant gait freezing (PSP-PGF), and PSP with predominant postural instability (PSP-PI) underwent baseline and follow-up positron emission tomography with 18F-fluorodeoxyglucose (18F-FDG PET) imaging. First, the PSP-RS-related pattern, PSP-P-related pattern, PSP-PGF-related pattern (PGFRP), and PSP-PI-related pattern (PIRP) were identified by scaled subprofile modeling/principal-component analysis. Furthermore, we evaluated the similarity of these patterns to previously reported PSPRP and tested their ability to classify clinical phenotypes. Finally, we evaluated these patterns longitudinally for monitoring disease progression and predicting conversion of vPSP to PSP-RS.

RESULTS

All type-specific patterns showed significant topographic similarity and association with PSPRP. The PGFRP exhibited more extensive involvement in the striatum than other type-specific patterns. In general, both type-specific patterns and PSPRP demonstrated similar performance in distinguishing PSP-RS from vPSP. Pattern expressions increased at follow-up and correlated with clinical progression. Higher baseline pattern expressions significantly predicted conversion from vPSP to PSP-RS.

CONCLUSIONS

PSP phenotypes exhibit type-specific metabolic covariance patterns that largely overlap with PSPRP but show phenotype-relevant differences. These patterns hold promise for longitudinal monitoring of disease progression and prediction of vPSP conversion to PSP-RS. © 2026 International Parkinson and Movement Disorder Society.