Early Differential Diagnosis of Epstein–Barr Virus-Associated Hemophagocytic Lymphohistiocytosis and Macrophage Activation Syndrome in Children: A Clinical Prediction Model Based on 106 Patients

The study screened clinical and laboratory indicators accessible within 48 h of admission for children with Epstein–Barr virus-associated hemophagocytic lymphohistiocytosis (EBV-HLH) and macrophage activation syndrome (MAS) to establish an efficient and convenient differential diagnosis model. The study retrospectively analyzed the clinical data and laboratory results obtained within 48 h of admission from 106 pediatric patients initially hospitalized and diagnosed with EBV-HLH or MAS at the Children’s Hospital of Soochow University from January 2019 to November 2024. Univariate logistic regression and LASSO regression filtered the variables alongside cross-validation and dimensionality reduction. The final regression model constructed a predictive nomogram, calibration curve, ROC curve, and DCA curve to evaluate the discrimination capacity and net clinical benefit of the model. Univariate logistic regression, LASSO regression with five-fold cross-validation, and stepwise multivariate logistic regression identified skin rash, bone marrow hemophagocytosis, fibrinogen, neutrophil percentage, and hepatosplenomegaly as the core predictors for differentiating EBV-HLH from MAS. The constructed nomogram and diagnostic calculator demonstrated favorable discriminative ability, with an area under the receiver operating characteristic curve of 0.938 (95% CI: 0.894–0.974). The calibration curve showed good agreement between predicted probabilities and actual observed outcomes, yielding a Brier score of 0.099. Decision curve analysis indicated that the model provided a significant positive net clinical benefit across a risk threshold range of 5% to 95%. A simple scoring system based on common clinical indicators effectively distinguished EBV-HLH from immune-related HLH in the early stages and guided initial treatment decisions.