To identify risk factors for 3-year major adverse cardiovascular events (MACE) in patients with acute myocardial infarction (AMI) after percutaneous coronary intervention (PCI) using machine learning algorithms, least absolute shrinkage and selection operator (LASSO) regression, and differential analysis, and to develop a corresponding predictive model.

A total of 400 AMI patients who underwent PCI at Xiang'an Hospital of Xiamen University between January 2020 and December 2021 were retrospectively enrolled. Patients were divided into a MACE group (n = 102) and a good prognosis group (n = 298) according to the occurrence of adverse cardiovascular events within 3 years after PCI, and were further divided into a training set (n = 280) and a validation set (n = 120) in a 7 ∶ 3 ratio. Chi-square test, independent samples t-test, and Mann-Whitney U test were used for preliminary screening of potential risk factors. Random forest and LASSO regression were further applied to identify important clinical features associated with poor prognosis. A Venn diagram was used to obtain the intersection variables from the three methods. The "rms" package was used to construct a nomogram based on the selected variables. The predictive performance of the model was evaluated using receiver operating characteristic (ROC) curve analysis and decision curve analysis with the "pROC" and "rmda" packages.

The differential analysis and LASSO regression respectively identified 11 risk factors. Among the 15 features selected by the random forest model, five overlapping variables were identified, including heart rate, age, serum creatinine, hemoglobin, and left ventricular end-diastolic diameter. The nomogram constructed based on these five variables demonstrated good predictive performance, with an accuracy of 0.643 in the training set and 0.683 in the validation set. The area under the ROC curve was 0.649 [95% confidence interval (CI) (0.571, 0.726), P < 0.001] in the training set and 0.796 [95%CI (0.699, 0.892), P < 0.001] in the validation set.

Through machine learning, LASSO regression, and differential analysis, key factors influencing adverse events within 3 years after PCI in AMI patients are identified. The constructed predictive model has high value in forecasting such events.