N7 - Engineering a Lingual Pulse Oximeter: Combating Discrepancies in Oxygen Saturation Amongst Skin Tones

Introduction: : Pulse oximeters are widely used in healthcare as convenient, noninvasive tools for assessing oxygenation status. Their alternative, arterial blood gas, is invasive and may not always be practical. Since the development of the fingertip probe in 1977 , pulse oximeter design has had minimal advancement.1 This design has known limitations, particularly in its reduced accuracy in patients with darker skin pigmentation.
Pulse oximeters may overestimate oxygen saturation due to melanin interfering with light absorption and reflection of hemoglobin, leading to occult hypoxemia, arterial oxygen saturation below 88% despite readings above 92%.2 This inaccuracy may result in missed diagnoses and delayed interventions, contributing to health disparities in populations with darker skin tones. Although this discrepancy has been recognized among many studies, there remains no proposed solution to mitigate this error. To address this, this study proposes a lingual pulse oximeter that measures oxygen saturation on the tongue, an area of decreased melanocyte variability across populations. By utilizing the lingual surface, the objective is to reduce the risk of occult hypoxemia and improve accuracy and equity in oxygen saturation monitoring.
The goal of this study is to engineer a non-invasive, reliable device that functions across all skin tones and improves patient outcomes in our diverse populations.

Materials and

Methods: : In veterinary medicine practices, a lingual pulse oximeter is utilized to evade the discrepancies in oxygen saturation that is caused by fur color. Despite the use of the pulse oximeter on the tongue, there is no currently available ‘lingual pulse oximeter.’ This study designed a modeled after this idea and utilizes a latching device that will extend and clasp the tongue to account for all tongue thicknesses and decrease the risk of slippage. The device will extend out of the patient's mouth and display SpO2 and temperature of the patient with a plastic cover that can be easily disposable after use to maintain sanitation. To evaluate the proof of concept a total of 23 male and female subjects ranging in age from 22-30 years were recruited for our study (IRB23-0164). Each participant completed a questionnaire about their Fitzpatrick skin type, exercise habits, and had their arm photographed alongside a Pixel Perfect Camera Color Correction Card. A Nonin Palmsat 2500 handheld pulse oximeter with a Model 8000Q2 ear clip was utilized for placement on the participants tongue concomitantly with an Oxiline Pulse 9s Pro fingertip oximeter on the finger. To maintain hygiene, the lingual pulse oximeter clip was wrapped in Saran Wrap, which was disposed of after each subject, followed by sanitization with CaviWipes. After reaching a hypoxic state of HR of 150-170 from the American Heart Association guidelines, oxygen saturation readings were taken from both devices.

Results, Conclusions, and Discussions:: At rest, finger SpO₂ averaged 97.10 ± 2.77 % and tongue SpO₂ averaged 96.14 ± 3.58 %. A paired t‑test showed no significant difference (p = 0.188), and equivalence testing within a ±2.5 % margin confirmed the two were statistically equivalent (p = 0.019). After exercise, finger SpO₂ readings were 95.41 ± 2.09 % and tongue SpO₂ readings were 93.95 ± 3.51 %. A paired t‑test showed a significant difference (p = 0.041), and equivalence testing failed at ±2.5 % (p = 0.067), indicating that finger oximeters slightly overestimate SpO₂ compared to tongue measurements after exercise.
The initial testing posed observations and challenges that will be implemented in our device design including device fit and comfortability. The veterinary ear clip that was outfitted as our tongue pulse oximeter was not designed for the human tongue which resulted in participants needing to stabilize it with their lips and teeth. Minor discomfort was reported due to keeping their mouth closed immediately after exercise, but no safety concerns arose. Finger-based measurements generally displayed values quickly, while the tongue readings took an additional amount of time post-exercise. In addition, the tongue readings did not give an oxygen saturation reading until it was also able to record a pulse reading. Other observations were in nail interference. One participant’s acrylic nails prevented the finger pulse oximeter from obtaining any reading, emphasizing the current finger design limitation. These observations were considered for our current design iteration focusing on patient comfort, ease of use, and multifunctional capabilities to include temperature measurement in addition to oxygen saturation measurements.
This initial testing indicates that the finger probe pulse oximeter and the lingual pulse oximeter exhibit similar SpO₂ readings while at rest yet when in a mild hypoxic state, the lingual pulse oximeter was better able to detect lower oxygen saturation levels. These preliminary results are promising for the benefits of a lingual pulse oximeter. This study is currently limited due to a relatively small sample size (n=23) with not all six Fitzpatrick skin types included in the study.

Acknowledgements and/or References (Optional)::