Modeling light-tissue interactions in pulse oximetry: effect of device design and skin pigmentation

Abstract

Pulse oximeters have traditionally been used for measuring oxygen saturation (SpO2), yet their widespread adoption has led to its expanded applications, including diagnostic uses like detecting hypoxemia during the COVID-19 pandemic. Recent research has identified unexpected errors, with discrepancies in SpO2 measurement accuracy and diagnostic efficacy based on race or skin pigmentation. The positive bias in SpO2 measurements, particularly in the critical range of 85-90%, poses a significant concern for clinical care. This study employs Monte Carlo simulations to model transmittance mode pulse oximetry and investigates whether these disparities can be attributed to variations in epidermal melanin volume fraction (Mf = 2-43%). The research aims to elucidate key mechanisms that may guide evaluation of clinical devices and the improvement of technology. Additionally, this study explores the effect of sensor/bandage reflectivity on sensitivity to the epidermal region. Simulation results reveal a significant impact of skin pigmentation on transmitted signals. Modulation ratios at 660 nm and 940 nm exhibit an inverse correlation with melanin fraction, notably more pronounced at 660 nm (20-25%) compared to 940 nm (6-12%). Modulation ratios and the magnitude of pigmentation dependence are shown to vary with SaO2. Calibration curves derived from simulations representing low Mf subjects generate positive bias when applied to data representing higher pigmentation subjects. The agreement of these results with recent clinical studies provides evidence that melanin content plays a key role in determining the accuracy of pulse oximeters.