Joint Modeling Study Identifies Blood-Based Transcripts Link to Cognitive Decline in Parkinson's Disease.

Abstract

Cognitive decline in Parkinson's disease (PD) is prevalent, insidious, and burdensome during the progression of the disease. We aimed to find transcriptome-wide biomarkers in blood to predict cognitive decline and identify patients at high risk with cognitive impairment in PD. We carried out joint modeling analysis to characterize transcriptome-wide longitudinal gene expression and its association with the progression of mild cognitive impairment (MCI) in PD patients. The average time-dependent area under the curves (AUCs) were used for evaluating the accuracy of the significant joint models. A cognitive survival score (CogSs) derived from joint model was leveraged to predict the occurrence of MCI. All predicting models were built in a discovery cohort with 272 patients and replicated in an independent cohort with 177 patients. We identified five longitudinal varied expression of transcripts that were significantly associated with MCI progression in patients with PD. The most significant transcript IGLC1 joint model accurately predicted the progression of MCI in PD patients in the discovery and replication cohorts (average time-dependent AUCs >0.82). The CogSs derived from the optimal IGLC1 joint model had a high accuracy at early study stage in both cohorts (AUC ≥0.91). Our transcriptome-wide joint modeling analysis uncovered five blood-based transcripts related to cognitive decline in PD. The joint models will serve as a useful resource for clinicians and researchers to screen PD patients with high risk of development of cognitive impairment and pave the path for Parkinson's personalized medicine. © 2022 International Parkinson and Movement Disorder Society.