Clinical assessment of a Near-Infrared spectroscopy device for rapid triage in traumatic brain injury.

BACKGROUND

Traumatic Brain Injury (TBI) poses a significant public health challenge in India, with nearly 2 million cases annually and limited CT availability causing delays in diagnosis. This study evaluated CEREBO®, a machine learning-enhanced near-infrared spectroscopy (mNIRS) device, as a rapid, non-invasive triage tool for TBI.

METHODOLOGY: A prospective quasi-experimental study was conducted at AIIMS, New Delhi, enrolling 202 suspected TBI patients. Participants were divided into control (standard care) and experimental (standard care + CEREBO®) groups. CEREBO® findings were compared with CT as the gold standard. Diagnostic performance (sensitivity, specificity, accuracy), time to preliminary diagnosis, and potential of impact on triage decisions were assessed.

RESULTS

CEREBO® demonstrated high diagnostic performance with 98% sensitivity, 90% specificity, and 93.9% accuracy for detecting intracranial pathology. In a post-hoc exploratory simulation, using CEREBO®'s negative result as a hypothetical criterion for deferring CT suggested that CT imaging could have been potentially avoided in 82.1% of red (critical condition) cases and 80.6% of yellow (stable but urgent) cases. These simulations also indicated that CEREBO® may help identify patients whose urgency of care could be underestimated by initial triage classification.

CONCLUSION

The NIRS device demonstrated rapid, non-invasive assessment capability across a broad patient population, supporting its potential utility as an adjunct screening tool in resource-limited trauma settings. These findings suggest that the device may help streamline triage, optimize CT utilization, and improve workflow efficiency in high-volume emergency environments.