This research introduces a 3D foundation model that transforms emergency head CT interpretation by detecting diverse neuro-trauma findings with high accuracy, addressing radiologist shortages and increasing demand.

• Leverages large language models (LLMs) for automatic, comprehensive multi-label annotations
• Provides rapid assessment of critical neurological conditions like hemorrhage, midline shift, and cerebral edema
• Demonstrates potential to reduce interpretation times while maintaining high diagnostic accuracy
• Creates a scalable approach for emergency radiology workflows

This innovation has significant clinical implications by enabling faster triage decisions in emergency settings, potentially improving patient outcomes while alleviating workload pressure on radiologists.

A Non-contrast Head CT Foundation Model for Comprehensive Neuro-Trauma Triage