The STEMI-MIRI random forest model successfully stratified patients, with myocardial ischemia-reperfusion injury incidence of 11.8%, 50.0%, and 76.0% across low-, moderate-, and high-risk groups.
Cohort (n=670)
No
Does a random forest machine learning model combining inflammatory-immune markers and anatomical parameters accurately predict myocardial ischemia-reperfusion injury in STEMI patients undergoing primary PCI?
A machine learning model incorporating systemic immune-inflammation index and neutrophil-to-lymphocyte ratio accurately predicts myocardial ischemia-reperfusion injury before PCI in STEMI patients, outperforming traditional logistic regression.
valor p: p=0.032
Myocardial Ischemia-Reperfusion Injury (MIRI) remains a critical challenge following Percutaneous Coronary Intervention (PCI) in patients with ST-Segment Elevation Myocardial Infarction (STEMI); current management strategies focus on post-procedural remedial interventions rather than preemptive risk assessment. To develop and validate a machine learning model for predicting the risk of MIRI by combining admission-derived inflammatory-immune markers with post-angiography anatomical parameters in the pre-PCI period for STEMI patients who are scheduled for emergency Percutaneous Coronary Intervention (PCI). In the present study, a random forest algorithm was used to develop the predictive model from the multidimensional clinical data of 528 STEMI patients (training set: 369; internal validation: 159; temporal validation: 142). Feature importance and SHapley Additive exPlanations (SHAP) value analyses were conducted in order to identify key predictors and explain nonlinear relationships. The model demonstrated strong discriminative performance across training, internal validation, and temporal validation sets (AUC: 0.838, 0.810, and 0.775, respectively), significantly outperforming logistic regression (P = 0.032). Systemic immune-inflammation index (SII) and neutrophil-to-lymphocyte (NLR) ratio were ranked as the top two predictors. Risk stratification showed a pyramid-shaped distribution with Myocardial Ischemia-Reperfusion Injury (MIRI) incidence of 11.8%, 50.0%, and 76.0% across low-, moderate-, and high-risk groups (6.4-fold gradient, P < 0.001), enabling pre-procedural risk identification and individualized preventive strategies. The STEMI-MIRI prediction model allows for the identification of patients at high risk of MIRI before the procedure, which could help in taking preventive measures and change the strategy from remediation to prevention.
Wan et al. (Fri,) conducted a cohort in ST-Segment Elevation Myocardial Infarction (STEMI) (n=670). STEMI-MIRI prediction model (Random Forest) vs. Logistic regression was evaluated on Model discriminative performance (AUC) for predicting Myocardial Ischemia-Reperfusion Injury (MIRI) (p=0.032). The STEMI-MIRI random forest model successfully stratified patients, with myocardial ischemia-reperfusion injury incidence of 11.8%, 50.0%, and 76.0% across low-, moderate-, and high-risk groups.