Graduate Student Seminar Series – Megh Rathod
March 24 @ 5:10 pm - 5:25 pm EDT
Graduate Student Seminar Series
Please ensure you invite your Principal Investigator by adding their email via the ‘Add Guest’ button and they will also be notified of your presentation.
Location: HS610 – 155 College St, Room 610
Presentation Title: Towards Equity in Wearable Health Technology: Addressing Skin Pigmentation Bias in Optical Sensors for Advancements in Pulse Oximetry
Abstract:
Numerous studies have shown that pulse oximeters are less accurate for individuals with darker skin, which negatively
impacts patient outcomes. This underscores the need for more equitable pulse oximetry strategies that account for
pigmentation to reduce this bias. Here, we developed a wearable reflective pulse oximeter to gather raw
photoplethysmography signals. Conventional and modified pulse oximeter algorithms were applied based on the ratio-of-
ratio’s method and a melanin term, derived from quantitative histology of the epidermis. This melanin-inclusive pulse
oximetry expression was developed and validated in a Hampshire porcine model (n=4) using adjacent areas of darkly and
lightly pigmented skin. The animals were exposed to controlled hypoxia with saturation levels ranging from 100% to 70%.
Blood gas analysis was performed and synchronized to the pulse oximeter readings on dark and light sections, with
epidermis biopsies sampled and Fontana-Masson stained for pigment quantification. The melanin-inclusive approach
reduced the bias in both the 100-85% (from 1.85% to 0.62%) and 84-70% saturation ranges (from 2.70% to 0.16%). This
was a marked improvement over conventional pulse oximetry and reference medical-grade devices, which exhibited higher
biases (3.08% to 4.98%). This study suggests that the bias in oximeter readings may be correlated with physical quantities
of melanin in the epidermis. Future work will entail obtaining a melanin coefficient from the wearable device itself for a
complete on-board non-invasive solution. This melanin-inclusive method is a promising strategy to reduce pulse oximeter
bias and improve the accuracy of a widely-used medical device, potentially enhancing care across diverse patient
populations.
Supervisor Name: Daniel Franklin
Year of Study: 4
Program of Study: PhD
Powered by Calendly.com
