A Nomogram Risk Prediction Model for Poor Outcome of Primary Brainstem Hemorrhage Based on Clinical Data and Radiographic Features

Objective: Primary brainstem hemorrhage (PBSH) is a devastating acute neurological disorder with a poor prognosis. This study aimed to identify risk factors associated with poor outcomes in PBSH patients and develop a novel nomogram for predicting prognosis, with external validation.

Methods: A total of 379 patients with PBSH were included in the training cohort. The primary outcome of interest was a modified Rankin Scale score (mRS) of 4-6 at 90 days post-onset. Multivariable logistic regression was used to construct a nomogram based on relevant variables. Model performance was tested in the training cohort and externally validated for discriminatory ability, calibration, and clinical utility at a separate institution. The nomogram was also compared to the ICH score in terms of predictive ability.

Results: The poor outcome rate at 90 days was 57.26% (217/379) in the training cohort and 61.27% (106/173) in the validation cohort. Multivariable logistic regression analysis identified age, Glasgow Coma Scale (GCS) score, and hematoma size as significant risk factors for poor outcomes. Nomograms based on these variables demonstrated good discrimination, with an area under the curve (AUC) of 0.855 and 0.836 in the training and validation cohorts, respectively. Furthermore, the nomogram showed superior predictive value to the ICH score for the 90-day outcome in both cohorts.

Conclusion: This study developed and externally validated a nomogram risk prediction model for predicting poor outcomes at 90 days in patients with PBSH, using age, GCS score, and hematoma size as predictors. The nomogram demonstrated good discrimination, calibration, and clinical validity, serving as a valuable assessment and decision-making tool.