Abstract
BACKGROUND AND OBJECTIVES: Non-valvular atrial fibrillation (NVAF) significantly increases the risk of acute ischemic stroke (AIS). Current risk prediction models have limitations in comprehensively capturing the multidimensional factors contributing to stroke risk. This study aimed to establish a novel nomogram model for predicting AIS in NVAF patients by integrating comprehensive parameters including clinical characteristics, cardiac anatomical features, functional indices, electrophysiological patterns, hemodynamic parameters, and serum biomarkers. METHODS: We conducted a retrospective study of 415 NVAF patients from Northern Jiangsu People's Hospital. After applying inclusion and exclusion criteria, 374 patients (193 with AIS) were randomized into 7:3 training/testing cohorts. Variables with P < 0.2 in univariate analysis were entered into LASSO regression, followed by review and validation by three senior clinical experts in neurology and cardiology, and then subjected to multivariate logistic regression to identify independent risk factors for nomogram construction. Model performance was comprehensively evaluated through receiver operating characteristic (ROC) curve analysis, calibration curves, and decision curve analysis (DCA) with bootstrap resampling (1,000 iterations). The predictive performance of the new nomogram model was compared with the CHA(2)DS(2)-VASc score using net reclassification improvement (NRI), integrated discrimination improvement (IDI), ROC analysis, calibration curves, and decision curves. RESULTS: Eight variables were identified as independent predictors of AIS in NVAF patients: age, admission systolic blood pressure (SBP), history of stroke, anticoagulant therapy, left atrial diameter (LAD), left atrial appendage (LAA) filling defect, white blood cell count (WBC), and D-dimer levels (all P < 0.05). The nomogram incorporating these parameters demonstrated excellent discrimination (AUC: 0.852 in training cohort, 0.847 in testing cohort), calibration, and clinical utility. Compared to the CHA(2)DS(2)-VASc score, the new model showed superior predictive performance across all evaluation metrics. CONCLUSION: The developed nomogram model, which integrates clinical, anatomical, functional, and laboratory parameters, demonstrates superior prediction performance compared to the conventional CHA(2)DS(2)-VASc score for AIS risk stratification in NVAF patients. This multidimensional approach may facilitate more personalized and precise risk assessment to guide preventive strategies.