Nomogram Models for Predicting Poor Prognosis in Lobar Intracerebral Hemorrhage: A Multicenter Study.

OBJECTIVE

We aimed to investigate the prognostic factors associated with lobar intracerebral hemorrhage (ICH) and to construct convenient models to predict 3-month unfavorable functional outcomes or all-cause death.

METHODS

Our study included 322 patients with spontaneous lobar ICH from 13 hospitals in Beijing as a derivation cohort. The clinical outcomes were unfavorable functional prognosis, defined as a modified Rankin Scale (mRS) score of 4-6, or all-cause death. Variable selection was performed using the Least Absolute Shrinkage and Selection Operator (LASSO) analysis, and two nomogram models were constructed. Additionally, multivariable logistic regression analysis was conducted to identify the factors associated with unfavorable prognosis. Finally, the Area Under The Receiver Operating Characteristic Curve (AUROC), calibration curve, and decision curve analyses (DCA) were performed to evaluate the models in both the derivation and external validation cohorts.

RESULTS

Predictive factors for unfavorable functional outcomes in lobar ICH included age, dyslipidemia, ICH volume, NIHSS score, Stroke-Associated Pneumonia (SAP), and lipidlowering therapy. The model included age, GCS score, NIHSS score, antihypertensive therapy, in-hospital rehabilitation training, and ICH volume to predict all-cause mortality. Our models exhibited good discriminative ability, with an AUC of 0.897 (95% CI: 0.862-0.933) for unfavorable functional outcomes and 0.894 (95% CI: 0.870-0.918) for death. DCA and calibration curves confirmed the models' excellent clinical decision-making and calibration capabilities.

CONCLUSION

Nomogram models for predicting 3-month unfavorable outcomes or death in patients with lobar ICH were developed and independently validated in this study, providing valuable prognostic information for clinical decision-making.