An individualized prediction of time to cognitive impairment in Parkinson's disease: A combined multi-predictor study. (25th September 2021)

• Record Type: Journal Article
• Authors: Tang, Chunyan; Zhao, Xiaoyan; Wu, Wei; Zhong, Weijia; Wu, Xiaojia
• Is Part Of: Neuroscience letters. Volume 762(2021)
• Publication Date: 2021-09-25
• Subjects: Parkinson's disease -- Cognitive impairment -- Radiomics -- Magnetic resonance imaging; Neurology -- Periodicals; Neurology -- Periodicals; Research -- Periodicals; Neurologie -- Périodiques; Neuroanatomie -- Périodiques; Neuropharmacologie -- Périodiques; Neurophysiologie -- Périodiques; Neurology; Periodicals; Electronic journals; 617.48
• DOI: 10.1016/j.neulet.2021.136149
• Languages: English
• ISSNs: 0304-3940

Abstract

Highlights: Radiomics plays an important role in predicting the CI of PD. The combined model could accurately predict the TTP of PD from NC to CI. A multi-predictor nomogram for individual prediction of TTP was established. Risk stratification for patients is beneficial to delay and prevent CI. Abstract: Background: Cognitive impairment (CI) is important for the prognosis of Parkinson's disease (PD). Early prediction whether and when cognitive decline from normal cognition (NC) will occur is crucial for preventing or delaying the progression timely. The current study aimed to provide a personalized risk assessment of CI by using baseline information and establishing a multi-predictor nomogram. Methods: 108 patients with PD were collected from the Parkinson's Progression Markers Initiative (PPMI), of whom 58 had progressed to CI and 50 remained NC during 5-year follow up. Radiomics signatures were obtained by using least absolute shrinkage and selection operator (LASSO) Cox regression algorithm. Clinical factors and laboratory biomarkers were selected by multivariate Cox regression analysis. The combined model of radiomics signatures and clinical risk factors was developed by a multivariate Cox proportional hazard model. A multi-predictor nomogram derived from the combined model was established for individualized estimation of time to progress (TTP) of CI. We analyzed the risk of two subgroups of the combined model by Kaplan-Meier (KM) analysis. Results: The combined model showed the best performance with a C-index of 0.988 and 0.926 in the training and validation datasets. KM analysis verified significant TTP of CI ( P ＜0.05) between two subgroups stratified by the cutoff value (−0.058). Conclusion: The combined model and its multi-predictor nomogram can be used to perfectly and individually predict the TTP of CI for patients with PD. Stratification of PD will benefit its timely clinical intervention and the delay and prevention of CI.