Multimodal data-driven prognostic model for predicting new-onset ST-elevation myocardial infarction following emergency percutaneous coronary intervention

Objectives: We developed a nomogram model derived from inflammatory indices, clinical data, and imaging data to predict in-hospital major adverse cardiac and cerebrovascular events (MACCEs) following emergency percutaneous coronary intervention (PCI) in patients with new-onset ST-elevation myocardial infarction (STEMI).

Methods: Patients with new-onset STEMI admitted between June 2020 and November 2022 were retrospectively reviewed. Data pertaining to coronary angiograms, clinical data, biochemical indices, and in-hospital clinical outcomes were derived from electronic medical records. Lasso regression model was employed to screen risk factors and construct a prediction model.

Results: Overall, 547 patients with new-onset STEMI who underwent PCI were included and assigned to the training cohort (n = 384) and independent verification cohort (n = 163). Six clinical features (age, diabetes mellitus, current smoking, hyperuricemia, neutrophil-to-lymphocyte ratio, and Gensini score) were selected by LASSO regression to construct a nomogram to predict the risk of in-hospital MACCEs. The area-under-the-curve (AUC) values for in-hospital MACCEs risk in the training and independent verification cohorts were 0.921 (95% CI 0.881-0.961) and 0.898 (95% CI 0.821-0.976), respectively. It was adequately calibrated in both training cohort and independent verification cohorts, and predictions were correlated with actual outcomes. Decision curve analysis demonstrated that the nomogram was capable of predicting in-hospital MACCEs with good clinical benefit.

Conclusions: Our prediction nomogram based on multi-modal data (inflammatory indices, clinical and imaging data) reliably predicted in-hospital MACCEs in new-onset STEMI patients with emergency PCI. This prediction nomogram can enable individualized treatment strategies.