An Automated Real-time Screening System with Decision-support Capabilities for Swift Traumatic Brain Injury Detection in Military Personnel

We propose to develop a portable, ultrasound-based imaging solution for accurate screening and real-time monitoring of traumatic brain injury (TBI) in both war and non-combat scenarios. This system will offer reliable and robust on-site analysis using an in-built decision-support framework to efficiently detect TBI in triage situations. Current ultrasound-based techniques detect TBI using a surrogate estimate of elevated intracranial pressure (EICP), i.e., the optical nerve sheath (ONS) diameter measurement with high sensitivity but low specificity. The low specificity limits the metric’s application for reliable screening. As phase I deliverables, we propose i) novel EICP predictors such as ONS pressure, as well as additional auxiliary non-invasive sensing modes for cerebral blood flow and blood oxygen content measurements, ii) novel 3D morphometric analysis of ONS, iii) an automated method for ONSD and ONSP measurement, and iv) a complete software prototype for US image analysis, 3D volume modeling and analysis incorporating i)-iii) above. The decision-support system will improve TBI classification by incorporating accurate on-site measurements that accounts for individual baseline variations and monitors temporal manifestation of TBI. In summary, this knowledge-driven system will equip a novice or a trained DOD medic with an easy-to-use, rapid and automated screening tool to improve TBI prognosis.