Development and Validation of a Nomogram for Predicting Cognitive Impairment in Patients with Leukoaraiosis

Background Leukoaraiosis (LA) is a common cerebral small vessel disease in elderly populations that frequently leads to cognitive impairment and may progress to vascular dementia. Early identification of cognitive dysfunction risk remains challenging due to the subtle onset and lack of specific biomarkers. Objective To identify key risk factors for cognitive impairment in LA patients and develop a logistic regression-based prediction model to facilitate early clinical identification and intervention. Methods This retrospective study included 390 LA patients admitted to the Department of Neurology between June 2020 and April 2023. Patients were classified into cognitive impairment (CI) and non-cognitive impairment (NCI) groups based on Montreal Cognitive Assessment (MoCA) scores.Data collected included demographics, medical history, biochemical markers, and neuroimaging features. The dataset was randomly split 7:3 into training (n = 273) and validation (n = 117) sets. Univariate analysis identified significant variables (p < 0.05), which were then incorporated into multivariate logistic regression analysis. A nomogram was constructed based on the final model, and performance was evaluated using receiver operating characteristic (ROC) curves and calibration plots for both training and validation sets. Results In the training set of 273 patients, 137 had cognitive impairment and 136 did not. Univariate analysis revealed that age, Fazekas score, intracranial arterial stenosis assessment (IASA), serum creatinine, and total bilirubin were significantly associated with cognitive impairment (p < 0.05). Multivariate logistic regression identified age (OR = 1.17, 95%CI: 1.11–1.24), IASA (OR = 2.52, 95%CI: 1.64–3.68), and Fazekas score (OR = 2.58, 95%CI: 1.74–3.60) as independent risk factors. The logistic regression model demonstrated excellent discrimination with AUC values of 0.873 for both training and validation sets. Calibration curves showed good agreement between predicted and observed probabilities, confirming model reliability. Conclusions Age, intracranial arterial stenosis assessment, and Fazekas score are independent risk factors for cognitive impairment in LA patients. The logistic regression model with nomogram provides a clinically practical tool for early identification and risk stratification of high-risk patients, enabling timely intervention to improve outcomes.