Challenging Skin Pigmentation Bias: Application of Time-Domain Near-Infrared Spectroscopy in Tissue Oximetry

Background and Motivation

In recent years, optical technologies have been increasingly utilized in medical diagnostics and therapies. However, varying levels of skin pigmentation (which depend on melanin concentration in the skin) can significantly interfere with the accuracy of optical devices. For instance, during the COVID-19 pandemic, many clinicians observed that pulse oximeters provided less accurate oxygen saturation (SpO2) readings in patients with darker skin when their oxygen levels were low. This issue has prompted the research community to reexamine the performance of optical devices across diverse populations. However, existing studies on the influence of skin pigmentation on various optical devices remain limited, especially in the emerging field of Time-Domain Near-Infrared Spectroscopy (TD-NIRS).

TD-NIRS is an optical technology based on short-pulse lasers, fast photodetectors, and timing electronics, designed to precisely measure the photon time-of-flight distribution in tissues. This allows for quantitative analysis of tissue optical properties such as the absorption coefficient and the reduced scattering coefficient. An important potential advantage of TD-NIRS is its relatively low sensitivity to changes in superficial skin layers (including pigmentation), making it a potentially more robust tool for tissue oximetry across populations with varying skin tones. This study systematically investigates whether TD-NIRS can overcome the biases caused by skin pigmentation and demonstrates its potential in this regard.

Origin of the Study

This research was published in the Biomedical Optics Express, Volume 16, Issue 2, February 2025, under the title “Challenging the skin pigmentation bias in tissue oximetry via time-domain near-infrared spectroscopy”. The study was conducted by researchers from multiple institutions, including Politecnico di Milano, Consiglio Nazionale delle Ricerche, and Buzzi Children’s Hospital. Alessandro Torricelli was the leading author of this research.

Study Design

The study comprised a series of experiments including laboratory simulations, static and dynamic tests on healthy volunteers, and clinical trials on pediatric patients. Each step was designed to assess the influence of skin pigmentation on TD-NIRS measurements.

1. Fabrication of Skin Phantoms and Baseline Experimental Validation

The research team first created a series of skin phantoms designed to simulate different levels of pigmentation. These phantoms were made using silicone elastomers (Sylgard 184), titanium dioxide (TiO2, used as a scattering agent), and ethanol-soluble nigrosine (a black dye used as an absorber). The melanosome fraction (MF), representing pigmentation levels, was adjusted to 0%, 2%, 6%, 14%, 30%, and 43%. The final thickness of these phantoms was approximately 270 ± 10 μm, and the diameter was 60 mm, closely emulating natural skin.

These skin phantoms were placed on two solid bulk phantoms (labeled B6 and C4) with known optical properties, and the effects of skin pigmentation on the absorption coefficient (µa) and reduced scattering coefficient (µ′s) were measured. The results demonstrated that, except for a melanosome fraction of 43%, changes in µa and µ′s were negligible across the wavelength range of 620–1100 nm. This indicates that TD-NIRS is relatively insensitive to variations in superficial skin pigmentation.

2. Static In-Vivo Measurements in Healthy Volunteers

To validate the applicability of the laboratory findings, the research team conducted in-vivo measurements on six healthy volunteers with various skin tones. Measurements were taken at three different sites: the forearm, forehead, and abdomen. During tests, skin phantoms simulating different pigmentation levels were placed over the natural skin.

The experiments showed that changes in MF had minimal impact on µa in all tested sites (absolute deviation within ± 0.005 cm-1). Changes in µ′s were slightly higher but remained within acceptable limits. These results further reinforced the robustness of TD-NIRS against interference from skin pigmentation.

3. Dynamic Tests in Healthy Volunteers: Arterial Occlusion Protocol

Dynamic arterial occlusion experiments were designed to simulate blood oxygen changes under clinical conditions. Six volunteers underwent occlusion testing with a manual pneumatic cuff inflated to 250 mmHg around the upper arm. During occlusion, TD-NIRS continuously monitored tissue oxygen saturation (StO2) and total hemoglobin concentration (tHb).

The results revealed that while skin phantoms with different MF levels caused some variation in tHb measurements, StO2 measurements remained consistent. Bias analysis further confirmed that the absolute deviation in StO2 was less than 1%, indicating high stability. Variations in tHb were attributed more to probe repositioning and physiological differences between volunteers.

4. Clinical Testing in Pediatric Patients

To evaluate TD-NIRS performance in a larger and more diverse population, the study included 352 pediatric patients with varying skin tones (classified using the Fitzpatrick scale). Measurements were taken at two sites: the left frontotemporal cortex (brain region) and the left mid-upper arm (skeletal muscle). Statistical analysis showed no significant differences in the measured StO2 or tHb values across Fitzpatrick categories, further confirming TD-NIRS’s robustness.

Conclusions and Significance

This systematic study led to the following key conclusions:

1. Demonstrated Robustness: Compared to other optical devices (such as Continuous-Wave Near-Infrared Spectroscopy or photoacoustic imaging), TD-NIRS showed significantly reduced sensitivity to skin pigmentation bias.
2. Clinical Applicability: Whether in static or dynamic conditions, various levels of skin pigmentation had no significant impact on tissue oxygen saturation measurements using TD-NIRS.
3. Technological Advantages: The time-domain characteristics of TD-NIRS, which decode deeper tissue layers via photon time-of-flight analysis, substantially reduced interference from superficial skin variations.

This research not only paves the way for TD-NIRS applications in diverse clinical scenarios but also provides critical insights into developing standardized practices for optical diagnostic tools in healthcare.

Key Highlights

• Innovative Use of Bio-Mimetic Models: The study’s design of skin phantoms ensured high experimental reproducibility and control.
• Comprehensive Performance Analysis: Static and dynamic evaluations of TD-NIRS provided valuable data for future device enhancements.
• Broad Applicability: Validation across multiple population subsets illustrated TD-NIRS’s potential to address public health challenges.

Future research could expand the sample size and simulate more complex interference environments to further advance TD-NIRS technology towards commercialization and clinical adoption.