Work

Classifying Parkinson Disease Progression via Latent Trajectories of Striatal Specific Binding Ratios from 123 I FP-CIT SPECT in Association with Levodopa-Induced Dyskinesia

Year: 2026

Type: article

Abstract: BACKGROUND AND PURPOSE: Levodopa-induced dyskinesia (LID) remains one of the most challenging complications of long-term dopaminergic therapy in Parkinson’s disease (PD), but reliable markers for identifying patients at higher risk remain uncertain. The purpose of this study was to determine whether longitudinal imaging trajectories derived from I-123 FP-CIT SPECT using latent class growth analysis (LCGA) show associations with LID occurrence, when considered with clinical characteristics. METHODS: 276 PD patients underwent I-123 FP-CIT scans at baseline, 1-year, 2-year, and 4-year follow-up periods. Specific binding ratios (SBRs) of the caudate and putamen were measured. LCGA was performed to classify patients into trajectory-based subgroups according to the patterns of dopaminergic denervation. Clinical variables including demographic characteristics, levodopa equivalent daily dose (LEDD), MDS-UPDRS scores, and motor subtypes (tremor-dominant, PIGD, and indeterminate), were assessed. Occurrence of LID within 4 years was the primary endpoint. Multivariate logistic regression was used to assess associations between baseline clinical/imaging variables and 4-year LID occurrence. RESULTS: LCGA identified two distinct trajectory classes among 276 patients, a gradual decline group (Class 1, n = 241, 87.3%) and a rapid decline group (Class 2, n = 35, 12.7%). Class 1 showed significantly lower baseline caudate (1.84 ± 0.46 vs 2.90 ± 0.37, p < 0.001) and putamen (0.70 ± 0.23 vs 1.33 ± 0.37, p < 0.001) SBRs compared with Class 2, but 4-year decline of SBRs was steeper in Class 2 (caudate −0.43 vs −0.92; putamen −0.20 vs −0.61). Class 2 was independently associated with LID in multivariate logistic regression (OR = 11.41, 95% CI 1.04–125.35, p = 0.047), along with the indeterminate motor subtype (OR = 10.56, 95% CI 1.809–61.682, p = 0.009). AUC of the multivariate model was 0.876 for 4-year LID occurrence. Conclusions: Longitudinal imaging trajectories demonstrate contemporaneous associations with LID occurrence in PD and may provide clinically relevant insights when combined with clinical characteristics.

Source: American Journal of Neuroradiology

DOI: 10.3174/ajnr.a9213

Authors: Kee Yeun Lee, Eun Hye Jeong, Won Woo Lee, Yoo Sung Song

Institutions: Ulsan National Institute of Science and Technology, Bundang Jesaeng Hospital, Seoul National University, Seoul National University Bundang Hospital

Language: English

Topic: Parkinson's Disease Mechanisms and Treatments

Subfield: Neurology

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: hybrid

Funder: Seoul National University Bundang Hospital