Unraveling the Predictors of Enlarged Perivascular Spaces: A Comprehensive Logistic Regression Approach in Cerebral Small Vessel Disease

Abstract

Ning Li,* Jia-Min Shao,* Ye Jiang, Chu-Han Wang, Si-Bo Li, De-Chao Wang, Wei-Ying Di Department of Neurology, Affiliated Hospital of Hebei University, Baoding, Hebei Province, People’s Republic of China*These authors contributed equally to this workCorrespondence: Wei-Ying Di, Email [email protected]: This study addresses the predictive modeling of Enlarged Perivascular Spaces (EPVS) in neuroradiology and neurology, focusing on their impact on Cerebral Small Vessel Disease (CSVD) and neurodegenerative disorders.Methods: A retrospective analysis was conducted on 587 neurology inpatients, utilizing LASSO regression for variable selection and logistic regression for model development. The study included comprehensive demographic, medical history, and laboratory data analyses.Results: The model identified key predictors of EPVS, including Age, Hypertension, Stroke, Lipoprotein a, Platelet Large Cell Ratio, Uric Acid, and Albumin to Globulin Ratio. The predictive nomogram demonstrated strong efficacy in EPVS risk assessment, validated through ROC curve analysis, calibration plots, and Decision Curve Analysis.Conclusion: The study presents a novel, robust EPVS predictive model, providing deeper insights into EPVS mechanisms and risk factors. It underscores the potential for early diagnosis and improved management strategies in neuro-radiology and neurology, highlighting the need for future research in diverse populations and longitudinal settings.Keywords: Enlarged Perivascular Spaces, cerebral small vessel disease, predictive model, risk factors, neuro-radiology, LASSO regression

Keywords

• enlarged perivascular spaces
• cerebral small vessel disease
• predictive model
• risk factors
• neuro-radiology
• lasso regression