Multimodal Prediction of Progression Toward Brain Death After Out-of-Hospital Cardiac Arrest

Area of Science:

• Neurology
• Critical Care Medicine
• Radiology

Background:

• Severe hypoxic-ischemic brain injury following out-of-hospital cardiac arrest (OHCA) can lead to brain death.
• Conventional neurological outcome classifications do not adequately capture this trajectory.
• A multimodal approach is needed for accurate prediction.

Purpose of the Study:

• To develop and evaluate a multimodal model to predict progression toward brain death (PTBD).
• The model integrates quantitative brain CT, serum neuron-specific enolase (NSE), and clinical variables.
• To assess the model's performance in predicting PTBD within 48 hours.

Main Methods:

• Retrospective analysis of prospectively collected data from the Korean Hypothermia Network registry.
• Inclusion of comatose OHCA survivors treated with targeted temperature management.
• Development of multivariable logistic regression models and internal validation using discrimination, calibration, and Brier score.

Main Results:

• The multimodal model demonstrated strong predictive performance (AUC 0.895, corrected AUC 0.890 in the total cohort).
• Key predictors included younger age, non-shockable rhythm, low gray-to-white matter ratio (GWR), and higher NSE at 48h.
• The model significantly outperformed individual predictors in the poor-outcome subgroup.

Conclusions:

• Progression toward brain death (PTBD) is a distinct clinical trajectory in OHCA patients.
• The developed multimodal model shows promising internal validation for early risk stratification.
• Further external validation is necessary before clinical implementation for neuroprognostication.