Abstract
Performing absolute measurements of tissue saturation of the brain with near-infrared spectroscopy (NIRS) is a clinically desirable brain monitoring tool. Tissue oxygenation index (TOI) is an indicator of absolute tissue mixed arterial and venous oxygen saturation, and can be calculated using a NIRS technique called spatially resolved spectroscopy (SRS). SRS instruments measure the change of light attenuation with distance by using multiple light source-detector distances at two or more wavelengths. The aim of the study is to use broadband NIRS SRS data to investigate the effects on the calculation of TOI of different parameters: wavelength selection, scattering dependence, source-detector distance, and resolving for water. In total, 55 neonates with hypoxic-ischemic encephalopathy were monitored using a broadband multi-distance continuous wave NIRS system; 172 datasets were recorded. Using a "Standard" approach, TOI values between 0 and 100% ("good") were calculated in 157/172 datasets with a mean TOI of 50%. By changing the wavelength selection, the number of "good" data sets increases to 165/172 with a mean of 60%. Alteration of the dependence of scattering on wavelength acts as a constant which shifts the absolute value of TOI significantly (p < 0.05), demonstrating the importance of having a subject-appropriate estimation of scattering dependence. In general, changing the combination of source-detector distances does not significantly alter the TOI (the mean TOI ranges from 41% to 53%) which suggests that the algorithm is robust to different source-detector combinations. The study shows the broadband NIRS SRS algorithm gives the opportunity to explore the calculation of TOI and could further improve the measurement of tissue saturation in a clinical setting.