Abstract
OBJECTIVE: Aortic dissection (AD) is a life-threatening cardiovascular emergency. Delayed diagnosis frequently leads to treatment delays, elevated mortality, and complications. This study investigates the factors contributing to the misdiagnosis of AD and proposes strategies for improving its early diagnosis. METHODS: A retrospective analysis of 801 patients with AD identified 219 cases for inclusion, which were split into a training set (131 cases) and a validation set (88 cases). A binary logistic regression model was used to identify factors influencing misdiagnosis, while a Nomogram prediction model was developed. RESULTS: The analysis revealed that factors such as the timing and suddenness of symptom onset, typical back pain, walk-in clinic visits, and laboratory results (D-dimer, fibrinogen, and white blood count) were significant in predicting misdiagnosis. The Nomogram model showed high predictive accuracy with an Area under the ROC curve (AUC) of 0.924 in the training set and 0.912 in the validation set, demonstrating good sensitivity and specificity. CONCLUSION: The model offers potential for improving diagnostic accuracy and clinical outcomes in AD cases.