Abstract
PURPOSE: Left ventricular aneurysm (LVA) is a severe and common mechanical comorbidity with acute myocardial infarction (AMI) that can present high mortality and serious adverse outcomes. Accordingly, there is a need for early identification and prevention of patients at risk of LVA. The aim of this study was to develop and validate a risk prediction model for LVA among AMI patients in Northwest China. METHODS: A total of 509 patients with AMI were retrospectively collected between January 2018 and August 2021. All patients were randomly divided into a training group (n=356) and a validation group (n=153). Potential risk factors for LVA were screened for predictive modelling using least absolute shrinkage and selection operator regression, multivariate logistic regression, clinical relevance, and represented by a comprehensive nomogram. Receiver operating characteristic curve, calibration curve, and decision-curve analysis (DCA) were used to assess the discrimination capacity, calibration, and clinical validity, respectively. RESULTS: Seven predictors were finally identified for the establishment of prediction model, including age, cardiovascular disease history, left ventricular ejection fraction, ST-segment elevation, percutaneous coronary intervention history, mean platelet volume, and aspartate aminotransferase. The prediction model achieved acceptable areas under the curves of 0.901 (95% confidence interval [CI]=0.868-0.933) and 0.908 (95% CI=0.861-0.956) in the training and validation groups, respectively, and the calibration curves fit well in our model. The DCA result indicated that this nomogram exhibited a favorable performance in terms of clinical utility. CONCLUSION: An accurate prediction model for LVA development established, which can be applied to rapidly assess the risk of LVA in patients with AMI. Our findings will aid clinical decision-making to reduce the incidence of LVA in high-risk patients, and counteract adverse cardiovascular outcomes.