Development of a nomogram model for predicting acute stroke events based on dual-energy CTA analysis of carotid intraplaque and perivascular adipose tissue

Objective To evaluate the predictive value of dual-energy CT angiography (DECTA) parameters of carotid intraplaque and perivascular adipose tissue (PVAT) in acute stroke events. Methods A retrospective analysis was conducted using clinical, laboratory, and imaging data from patients who underwent dual-energy carotid CTA and cranial MRI. Acute cerebral infarctions occurring in the ipsilateral anterior circulation were classified as the symptomatic group (STA group), while other cases were categorized as the asymptomatic group (ATA group). LASSO regression was employed to identify key predictors. These predictors were used to develop three models: the intraplaque model (IP_Model), the perivascular adipose tissue model (PA_Model), and the nomogram model (Nomo_Model). The predictive accuracy of the models was evaluated using receiver operating characteristic (ROC) analysis, calibration curves, and decision curve analysis. Statistical significance was defined as p < 0.05. Results Seventy-five patients (mean age: 68.7 ± 8.7 years) were analyzed. LASSO regression identified seven significant variables (IP_Zeff, IP_40KH, IP_K, PA_FF, PA_VNC, PA_Rho, PA_K) for model construction. The Nomo_Model demonstrated superior predictive performance compared to the IP_Model and PA_Model, achieving an area under the curve (AUC) of 0.962, with a sensitivity of 95.8%, specificity of 82.4%, precision of 82.6%, an F1 score of 0.809, and an accuracy of 88.0%. The clinical decision curve analysis further validated the Nomo_Model’s significant clinical utility. Conclusion DECTA imaging parameters revealed significant differences in carotid intraplaque and PVAT characteristics between the STA and ATA groups. Integrating these parameters into the nomogram (Nomo_Model) resulted in a highly accurate and clinically relevant tool for predicting acute stroke risk.