Establishment and verification of a clinical prediction model for hematoma expansion in the acute phase of hypertensive intracerebral hemorrhage

To establish and validate a nomogram prediction model for hematoma expansion (HE) in the acute phase of hypertensive intracerebral hemorrhage (HICH), based on clinical data, laboratory results, and imaging features, providing a new theoretical basis for the diagnosis and treatment of HICH. A retrospective analysis of 569 HICH patients was performed. Patients were divided into two groups based on the presence or absence of HE. Thirty-one potential influencing factors were screened using least absolute shrinkage and selection operator (LASSO) regression to establish a nomogram prediction model. Receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) were used to evaluate the model's discrimination, calibration, and clinical utility. Among the 569 HICH patients,132 experienced HE (incidence rate:23.2%). LASSO regression identified six predictors (the time from onset to initial CT (FirstCT), delayed intraventricular hemorrhage, blend sign, black hole sign, GCS score, and INR) to construct the nomogram. The area under the ROC curve was 0.784. The Hosmer-Lemeshow test showed P = 0.994, indicating good agreement between the predicted and actual outcomes. DCA demonstrated clinical applicability within a certain probability range, providing a theoretical basis for early clinical intervention. The prediction model based on clinical data, laboratory results, and imaging features has good predictive performance and can aid in early identification and individualized treatment of HE in HICH patients.