Machine learning prediction of contrast-induced AKI after PCI using the Systemic Immune-inflammation Index: insights from MIMIC-IV

This study aimed to evaluate the systemic immune-inflammation index (SII) for predicting contrast-induced acute kidney injury (CI-AKI) and to develop a machine learning model integrating SII with key risk factors. Data were derived from the MIMIC-IV database (2008-2019) for acute myocardial infarction patients undergoing percutaneous coronary intervention in the intensive care unit. Logistic regression and restricted cubic splines were used to assess the association between SII and CI-AKI. Six machine learning models were developed on 70% of the training set and validated on the remaining 30%. Among 1,334 included patients, multivariable logistic regression identified a higher SII as a significant independent predictor for CI-AKI (Q4 vs. Q1: OR = 2.90, 95% CI: 2.01-4.19, <i>p</i> < 0.001). The Random Forest model demonstrated the best performance, achieving an area under the curve (AUC) of 0.84 (95% CI: 0.78-0.91) in the validation set, significantly outperforming the traditional modified Mehran score (AUC = 0.84 vs. 0.72). Calibration and decision curve analyses confirmed the model's robustness and clinical utility. In conclusion, SII is strongly associated with CI-AKI risk, and the developed Random Forest model integrating SII offers a superior and interpretable tool for pre-procedural risk stratification, potentially guiding targeted prevention strategies.