Jin Jingkun, Ding Jiahui, Zhang Xishen, Wang Linsheng, Zhang Xudong, Li Wenhua, Li Shanshan
The First School of Clinical Medicine, Xuzhou Medical University, Xuzhou, People's Republic of China.
Department of Cardiology, The Affiliated Hospital of Xuzhou Medical University, Xuzhou, People's Republic of China.
Clin Interv Aging. 2025 Jul 25;20:1085-1098. doi: 10.2147/CIA.S534736. eCollection 2025.
ST-segment elevation myocardial infarction (STEMI), the most severe form of acute coronary syndrome (ACS), requires timely percutaneous coronary intervention (PCI) to restore coronary blood flow. However, contrast-induced acute kidney injury (CI-AKI), the third most common cause of hospital-acquired renal failure, remains a critical complication of PCI.
To develop a machine learning model predicting CI-AKI risk in elderly patients with STEMI patients using clinical features.
Data from 2120 elderly patients with STEMI treated with PCI at Xuzhou Medical University Affiliated Hospital (2019-2023) were used for model development and testing. An external validation cohort, comprising 236 individuals, was derived from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database (2008-2019). Lasso regression selected predictors, and nine Machine Learning (ML) algorithms were evaluated via Receiver Operating Characteristic (ROC) analysis. Overlapping top-ranked features from high-performing models (AUC >0.8) informed a nomogram. Performance was assessed using AUC and decision curve analysis (DCA).
The final model included five independent predictors: lymphocyte-to-monocyte ratio, diuretic use, residual cholesterol, serum creatinine, and blood urea nitrogen. This model was developed as a simple-to-use online dynamic nomogram. It demonstrated robust discrimination, with C-statistics of 0.782 in the testing dataset and 0.791 in the validation dataset. DCA confirmed its clinical utility across a wide range of risk thresholds.
A new online dynamic nomogram was developed to provide a practical tool for CI-AKI risk stratification in elderly STEMI patients, aiding personalized prevention strategies.
ST段抬高型心肌梗死（STEMI）是急性冠状动脉综合征（ACS）最严重的形式，需要及时进行经皮冠状动脉介入治疗（PCI）以恢复冠状动脉血流。然而，造影剂诱导的急性肾损伤（CI-AKI）是医院获得性肾衰竭的第三大常见原因，仍然是PCI的关键并发症。
利用临床特征开发一种机器学习模型，预测老年STEMI患者发生CI-AKI的风险。
使用徐州医科大学附属医院（2019 - 2023年）2120例接受PCI治疗的老年STEMI患者的数据进行模型开发和测试。一个由236名个体组成的外部验证队列来自重症监护医学信息集市-IV（MIMIC-IV）数据库（2008 - 2019年）。套索回归选择预测因子，并通过受试者工作特征（ROC）分析评估9种机器学习（ML）算法。来自高性能模型（AUC > 0.8）的重叠顶级特征形成了一个列线图。使用AUC和决策曲线分析（DCA）评估性能。
最终模型包括五个独立预测因子：淋巴细胞与单核细胞比值、利尿剂使用、残余胆固醇、血清肌酐和血尿素氮。该模型被开发为一个易于使用的在线动态列线图。它显示出强大的辨别力，测试数据集中的C统计量为0.782，验证数据集中为0.791。DCA证实了其在广泛风险阈值范围内的临床实用性。
开发了一种新的在线动态列线图，为老年STEMI患者的CI-AKI风险分层提供了一种实用工具，有助于制定个性化预防策略。
