Development and validation of a nomogram to predict acute aortic dissection in sudden chest pain patients

Background: The aim of this study was to construct a model by combining routine laboratory biomarkers and clinical characteristics to distinguish acute aortic dissection (AAD) patients from other sudden chest pain patients with AMI, APE and AAA. Methods and Results: Qualified patients were randomly divided into training and validation cohorts. Independent predictive factors for differentiating AAD were filtered out via backward stepwise logistic regression. A nomogram containing the included factors was constructed. The discrimination and calibration abilities were verified via receiver operating characteristic (ROC) curves and calibration curves. The clinical use of the nomogram was evaluated via DCA. A total of 860 eligible patients were randomly allocated to the training (602) and validation (258) cohorts. The WBC count, Baso%, NLR, age, DD and alcohol status were established as independent factors for patients with AAD after multiple logistic regression analysis. A nomogram was constructed. The AUC values were 0.775 (0.733--0.817) and 0.709 (0.637--0.781) for the training and validation cohorts, respectively. The Hosmer–Lemeshow test revealed no significant difference (P>0.05), indicating that the nomogram was reliable. DCA showed favorable clinical benefit. Conclusion. This study constructed a prediction model for AAD. Validation revealed excellent discrimination and calibration, indicating that the nomograms may provide clinical reference information and increase the diagnostic efficiency of AAD.