CBI-M framework shows strong TBI outcome discrimination in two cohorts

Researchers analyzed prospective observational data from 4,509 patients in the European CENTER-TBI study and 2,552 in the US TRACK-TBI study. They assessed how the CBI-M framework and its four pillars related to CT-detected intracranial injury, hospital and intensive care admission, intracranial pressure monitoring, major cranial surgery, and six-month outcomes. Performance was compared with the IMPACT-Lab and CRASH-CT prediction models in their respective target populations.

For diagnostic and care-pathway endpoints, full CBI-M models had AUCs of 0.83–0.95 in CENTER-TBI and 0.88–0.98 in TRACK-TBI. For six-month mortality, AUCs were 0.89 and 0.91, respectively; for unfavorable outcome (GOSE ≤4), they were 0.92 and 0.95. Health-related quality-of-life performance was modest. The biomarker pillar contributed most to identifying CT abnormalities, while imaging contributed most to prediction of major cranial surgery.

The findings support further evaluation of CBI-M as a multimodal approach to characterizing acute TBI, but do not establish readiness for routine clinical implementation. The authors call for prospective validation and optimization. Generalizability is limited by the high-income-country cohorts, three unavailable Modifier-pillar elements, and possible reverse causality in some analyses.