Evaluating performance of a near-infrared-spectroscopy-based and machine-learning-powered bio-optical sensitivity parameters in identifying intracranial hemorrhages in TBI across different age-groups.

Primary objective: To evaluate the accuracy of an innovative machine-learning-powered near-infrared spectroscopy (mNIRS)-based bio-optical sensitivity parameters, namely specific tissue optical index (STOI) and intracranial tissue optical index (ITOI) for effective triaging of patients with suspected traumatic brain injury (TBI) across different age-groups to facilitate timely intervention and curtail the silent epidemic.

Research design: Prospective, observational, double-blinded, cross-sectional single-center.

Methods: A total of 240 subjects suspected with TBI and recommended for head CT scan were enrolled. Findings of CT were compared with the bio-optical sensitivity parameters (STOI+ITOI) generated by the mNIRS system to detect intracranial hemorrhages (ICHs).

Statistical software used: spss (ibm version 21.0) results: Bio-optical sensitivity parameters (STOI+ITOI) analysis of 1288 scanned lobes showed high specificity of 97% (CI 95-98%), sensitivity of 96% (CI 92-99%), and accuracy of 97% (CI 96-98%) across all ages, maintaining robust performance for detecting subdural, epidural, subarachnoid hemorrhages, and contusions. Also, the reliability in the diagnosis of ICH was evidenced by Youden's index of 0.93 and positive and negative likelihood ratios of 29.13 and 0.04, respectively.

Conclusion: The mNIRS-based STOI+ITOI proves to be an effective triage tool for TBI, exhibiting superior diagnostic performance across all age-groups.