Skin Pigmentation Impacts on Established and Emerging Optical Diagnostic Devices: A Survey of Mechanisms and Effects
- Authors:
- Center:
-
Contributing OfficeCenter for Devices and Radiological Health
Abstract
Reduction or elimination of racial/ethnic disparities represents an area of major public health need. In some cases, biological characteristics correlated with race or ethnicity – such as epidermal melanin content – can contribute to disparities. In the field of biomedical optics, melanin has long been recognized as a major dermal chromophore, or absorber of visible to near-infrared light. The same optical characteristics that make melanin the chief determinant of visually identifiable pigmentation also have the potential to adversely impact the performance of medical devices based on light. Three factors make melanin a particularly challenging factor in cutaneous light-tissue interactions: it has an absorption coefficient far greater than any other dermal constituent, its concentration can vary by more than an order of magnitude across the population, and its absorption level changes rapidly across the visible to near-infrared range. Scientific journal articles contain clear evidence of the impact of melanin on detected optical signals. Recently, there has been increasing attention to the impact of this effect on optical diagnostic device performance – particularly pulse oximeters and wearable photoplethysmography (PPG) devices. However, there are a variety of other currently marketed and emerging optical modalities that interrogate skin with light – most commonly for oximetry and cancer detection applications. In a recent congressional inquiry to FDA on skin pigmentation-dependent bias in pulse oximetry, one of the issues raised was whether similar problems occur in other optical device types. To identify the range of currently marketed or emerging optical modalities that are impacted by variations in skin pigmentation, we have initiated an extensive review of literature. This work will provide a strong foundation for regulatory progress in this area, as it will help us to clarify the scope of this issue, anticipate the ways in which device performance can be impacted and the working mechanisms involved, and determine optimal approaches to mitigate the impact of melanin. In addition to pulse oximeters and PPG wearables, we have initiated a review of literature on a wide range of optical diagnostic technologies, including cerebral oximeters, photoacoustic imagers, bilirubinometers, hyperspectral imaging systems, functional near-infrared spectroscopy devices (fNIRS), and Raman spectroscopy devices. Research with reflectance spectroscopy systems has shown a strong decrease in signal intensity with increasing pigmentation level; the magnitude of this effect becomes smaller across the near-infrared (NIR) spectrum. Hyperspectral imaging systems have shown similar changes, which in some cases causes devices to produce inaccurate results in subjects with strong pigmentation. While pulse oximeters that operate in the NIR reportedly produce a small positive bias in subjects with strong pigmentation, cerebral oximeters have shown the opposite effect. Bilirubinometer manufacturers have implemented methods for optically assessing and extracting the effect of melanin in visible-wavelength measurements. fNIRS device researchers have shown significant decreases in detected signal (and thus signal-to-noise ratio) with skin pigmentation, but no impact on measurement of hemoglobin concentration levels was found. Emerging photoacoustic imaging technology has been shown to produce strong superficial acoustic signals due to melanin that have the potential to reduce image quality. Melanin is also known to produce fluorescence that can interfere with measurements of low-yield endogenous Raman signals. Our preliminary assessment of the scientific literature indicates that skin pigmentation impacts a wide range of optical diagnostic device types through different mechanisms and with different effects. Further analysis of this information will help CDRH identify trends useful for guiding regulatory science research and making regulatory decisions. For example, our initial findings indicate the need for phantom-based test methods to enable benchtop assessment of robustness to pigmentation and improved understanding of algorithms used to extract the impact of pigmentation from detected optical signals.