Integrated risk stratification for ICI-associated myocarditis: a baseline hematological profile and a combined ECG and enzymatic signature at onset

Background Immune checkpoint inhibitor-associated myocarditis (ICI-associated myocarditis) is a rare but fatal immune-related adverse event. Early identification of high-risk patients remains challenging. This study aimed to identify risk factors and develop models for predicting both the occurrence and severity of ICI-associated myocarditis. Methods This retrospective unmatched case-control study stratified patients receiving ICIs into ICI-associated myocarditis (stratified into mild and severe subgroups) and controls. Comparative analysis of baseline and onset-phase data was performed, with logistic regression used to identify risk factors for the development of ICI-associated myocarditis and the severe myocarditis. Results In this cohort of 196 patients (98 myocarditis cases vs. 98 controls), a two-tiered risk stratification was established. Patients with myocarditis were further stratified into mild (n=71) and severe (n=27) subgroups. For predicting the occurrence of ICI-associated myocarditis, a baseline model incorporating elevated eosinophil ratio, reduced lymphocyte ratio, and elevated myoglobin demonstrated an area under the ROC curve (AUC) of 0.699 (95% CI, 0.626-0.772, P < 0.001). Upon onset, for predicting severe myocarditis, a model combining electrocardiographic abnormalities (T-wave changes, bundle branch blocks) and marked CK elevation (>10× ULN) achieved a higher AUC of 0.769 (95% CI, 0.655-0.882, P < 0.0001). Severe cases presented significantly earlier than mild cases (33 vs . 63 days, P = 0.043) and had higher rates of symptoms and concurrent immune-related adverse events. Conclusion A baseline profile of elevated eosinophil ratio, reduced lymphocyte ratio, and elevated myoglobin collectively may help identify patients at risk for ICI-associated myocarditis. At myocarditis onset, a combination of specific electrocardiographic abnormalities and marked CK elevation may predict severe cases. These findings suggest a two-stage approach for early risk stratification and targeted management.