Abstract
Computed tomography (CT) brain imaging is routinely used to support clinical decision-making in patients with traumatic brain injury (TBI). Only 7% of scans, however, demonstrate evidence of TBI. The other 93% of scans contribute a significant cost to the healthcare system and a radiation risk to patients. There may be better strategies to identify which patients, particularly those with mild TBI, are at risk of deterioration and require hospital admission. We introduce a blood serum liquid biopsy that utilizes attenuated total reflectance (ATR)-Fourier transform infrared (FTIR) spectroscopy with machine learning algorithms as a decision-making tool to identify which patients with mild TBI will most likely present with a positive CT scan. Serum samples were obtained from patients (n = 298) patients who had acquired a TBI and were enrolled in CENTER-TBI and from asymptomatic control patients (n = 87). Injury patients (all severities) were stratified against non-injury controls. The cohort with mild TBI was further examined by stratifying those who had at least one CT abnormality against those who had no CT abnormalities. The test performed exceptionally well in classifications of patients with mild injury versus non-injury controls (sensitivity = 96.4% and specificity = 98.0%) and also provided a sensitivity of 80.2% when stratifying mild patients with at least one CT abnormality against those without. The results provided illustrate the test ability to identify four of every five CT abnormalities and show great promise to be introduced as a triage tool for CT priority in patients with mild TBI.