Near infrared spectroscopy monitoring in prehospital traumatic brain injury care

Abstract

Traumatic brain injury (TBI) is a significant cause of morbidity and mortality across all age groups. The impact of these injuries can be lifelong and affect not just the individual but those around them and society itself. This has made TBI an area of interest for researchers and clinicians over decades with minimal change in patient outcomes.

This thesis focuses on the moments after the injury. Prehospital medicine disrupts assessment and reduces treatment options. There is a monitor that is missing from prehospital care - a non-invasive monitor to guide therapy by providing information about the brain itself. This could then guide advanced therapies at the roadside including anaesthesia, intubation and ventilation, transfusion and administration of vasoactive agents.

The physics of near-infrared spectroscopy (NIRS) allows optodes placed on the skin to measure local tissue oxygen saturation and blood volume with a high degree of temporal resolution. Although the role of NIRS has been explored in contexts including TBI and surgery its precise clinical role is still being sought.

This thesis explores the role of NIRS tissue oximetry as a non-invasive monitor of cerebral tissue in prehospital TBI care. The monitor demonstrated technical feasibility, however standard analytic approaches did not find meaningful associations between rSO2 values and either short-term and long-term outcomes or pathology. This may be a result of a requirement for much larger studies to establish adequate power along with significant gaps in the monitoring data. Our further exploratory analysis suggests that deep learning approaches may be able to discern significant patterns within NIRS oximetry data. Further development of this approach requires engineering improvements to monitors, larger datasets and clearer evidence of the best use of clinically linked monitoring data from the hospital environment to justify further research at the accident scene.