Abstract
INTRODUCTION: Cardiotoxicity has become a major concern in cancer patients, especially those with lung cancer, as anti-tumor therapies can significantly affect patient survival and quality of life. This study aims to develop and validate a dynamic nomogram based on the Inflammation Burden Index (IBI) to predict the risk of cardiac injury within one year after anti-tumor treatment in lung cancer patients. METHODS: This single-center, retrospective study included 1386 lung cancer patients who underwent myocardial enzyme testing between July 2018 and January 2023. The IBI was calculated as: IBI = (CRP (mg/dL) × Neutrophils (/μL)) / Lymphocytes (/μL). Statistical analysis using SPSS 22.0 and R 4.4.1, including machine learning algorithms and multivariate logistic analysis, identified independent predictors of cardiac injury. An online dynamic nomogram was developed and validated using internal validation, ROC curves, and decision curve analysis (DCA). RESULTS: The average age of the 1386 patients was 61.98 ± 9.22 years. Significant independent predictors included age, BMI, hypertension, immunotherapy, D-dimer, LDH, NSE, CKMB, and IBI. The nomogram showed strong discriminative ability with AUC-ROC values of 0.85 for the training set and 0.86 for the validation set. Calibration curves confirmed good fit, and DCA showed high clinical utility. CONCLUSION: An online dynamic nomogram based on clinical and inflammatory markers was developed to predict cardiac injury in lung cancer patients following anti-tumor therapy. The model shows strong discriminative ability and potential clinical value, which can provide vital information for oncologists when designing customized clinical treatments.