Towards achieving even distributions of participant skin tones when verifying pulse oximeter performance

Towards achieving even distributions of participant skin tones when verifying pulse oximeter performance | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Short Report Towards achieving even distributions of participant skin tones when verifying pulse oximeter performance Wim Verkruysse, Bernice E. Rogowitz, David H. Milkes, Paul Mannheimer, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7112852/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 5 You are reading this latest preprint version Abstract Purpose. This communication describes a a skin tone characterization methodology for the international standard ISO 80601-2-61:2026 (Ed 3) on pulse oximeter basic safety and essential performance. The methodology’s goal is to create the appropriate proportional distributions of skin tones for participants in pulse oximeters performance verification studies. We elucidate the techniques and associated metrics used for characterizing skin tone. This standard is currently in its final draft and voting stage before becoming formally adopted. Methods Various methods to quantify pigmentation were evaluated for their suitability for the purpose. Evaluation criteria included precision, accuracy, standardizability and efficacy for communication. The evaluations and selections are the result of ISO Joint Working Group activity over a period of about two years. Results The individual typology angle (ITA), was identified as a preferred objective measurement method of pigmentation. It is colorimetry based, is consequently highly standardized and has high precision and accuracy. The Monk Skin Tone scale was selected as a subjective method to complement the objective method as it is much more relatable than the ITA. Conclusion A standardized method for assessing skin tone and standardized expectation for balanced participant enrollment across the world’s range of skin tones was defined. A colorimetry based metric is used for precision and accuracy in combination with a color scale for optimal communication and recruitment. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction As an organization, International Organization for Standardization (ISO) aims to improve the performance of pulse oximeters, a device that non-invasively estimates the hemoglobin oxygen saturation of arterial blood. Stakeholders in the ISO Oximeters joint work groups (JWG) , include clinicians, manufacturers, regulators and other experts who collaboratively develop and maintain the standards. A recently published draft (1) on standardization of pulse oximetry (PO) includes significant changes to address the potential health inequity related to differences in SpO 2 readings between light and dark skin groups, that have been observed in real world studies (RWS) (2–4). In the process of updating the ISO standard to its third edition, the JWG agreed that the underlying root cause(s) for the observed bias differences remain(s) unknown. More specifically, whether the degree of melanin pigmentation is the direct root cause, or merely a strong correlator with another confounding variable, is not known. The JWG identified two main objectives, each with a slightly different working hypothesis on the root cause. In both hypotheses it is acknowledged that SpO 2 bias may have been missed in earlier device verification studies, where dark skin individuals were under-represented. The first objective was to ensure a balanced and wider distribution of skin tone among the participants enrolled in pulse oximeter verification studies. The second objective was to create the methodology to characterize the degree of difference in SpO 2 readings between dark and light skin participants, defined as pigmentation differential bias , observed in these verification studies. While these two objectives complement one another, they address different aspects of the task – study participant enrollment and device characterization. Here, we focus on the first objective: study participant enrollment. The main goal of this paper is to share the rationale, and a description of the pigmentation metrics used as they are likely unfamiliar to many. In the section Enrollment Methodology, these metrics are presented in a concise manner. They will be placed in a broader context in the Discussion, for a more thorough understanding of the metrics. Since PO and the metrics to quantify pigmentation are both rooted in tissue/skin optics, we briefly address a few basic elements, shown in Figure 1. Skin has two dominant chromophores: hemoglobin (Hb) in blood, and melanin. Deoxy-hemoglobin and oxy-hemoglobin (HHb and O2Hb, respectively) are the most relevant Hb species for the purpose of PO. Melanin, located in the epidermis, varies strongly in concentration between people, resulting in skin tones that span from very light to very dark (5). Absorption of light by blood and melanin give skin its characteristic range of colors. Besides absorption, light scattering also influences the penetration depths of each wavelength; as a general principle, the longer wavelengths penetrate deeper in skin. As a result, both the concentration and the distribution of chromophores have an impact on the skin color (6). The reflectance spectra shown in Figure 1B illustrate that the impact of blood on the skin color becomes gradually smaller with increasing levels of melanin: the oxy-hemoglobin absorption peaks at 545nm and 577nm have a less pronounced effect for darker skin. The concept of PO is shown in Figure 1C. Red and infra-red light emitted by light emitting diodes (LEDs) and detected in transmission (or reflectance, in devices with different form factor) is modulated in time by the pulsatile volume of arterioles. The absorption by melanin at these wavelengths is much smaller than at the shorter visible wavelengths. Nevertheless, the absorption is non-zero, and the distinctly higher absorption in red compared to infra-red may affect light paths of red and infra-red differently. The respective modulation amplitudes might also be affected and consequently cause a pigmentation differential bias. Enrollment methodology Proposed pigmentation metrics Two pigmentation metrics are used in the new standard to ensure study enrollment spans the wide range of human skin tones: the Monk Skin Tone (MST) scale (7) and the individual typology angle (ITA) (8). Illustrations of MST and ITA are provided in Figure 2. The main difference between these two metrics is that the MST is subjective while ITA is objective (9). A second difference is that MST aims at expressing pigmentation as the appearance, complexion, or ‘gestalt’ of an individual as a whole, not a specific skin spot [10]. ITA expresses melanin pigmentation on a certain anatomical location, which can vary considerably within an individual [11,12]. The ITA is measured by placing a color-sensing device (spectrophotometer/colorimeter) on the skin, evaluating a circular spot of several mm in diameter [13], depending on the device. By measuring how much of the device’s illuminating light is reflected at different wavelengths, an objective skin color is determined and expressed in the CIELab standardized color space (14) using three parameters: L*, a*, and b*. The lightness value L* ranges from black (value 0) to white (value 100). Parameters a* and b* represent the chromaticity where a* represents the balance between magenta/red (positive values) and green (negative values). Similarly, b* represents the balance between yellow (positive values) and blue (negative values). Skin typically has positive values for a* and b* as it is reddish/yellowish rather than greenish/blueish. The ITA conveniently expresses melanin pigmentation in a one-dimensional parameter (an angle), noting the curved shape in the Figure 2 B graph of L* vs b* in individuals with differing melanin pigmentation [15]. A user wishing to measure ITA computes it from L* and b* (implicitly discarding a*) using the formula [8] shown above the graph. To minimize errors in the implementation of the formula [16] it is recommended to do this in a spreadsheet and perform a sanity check by comparing the results with known ITA values. Besides the subjective/objective difference, it became clear from discussions and research, that MST and ITA both have distinct advantages and disadvantages (Table 1). The JWG approach leverages their complementary qualities. Note that Table 1 contains only the key -advantages for the stated purpose of participant enrollment in PO verification studies. For many other purposes one can identify other key advantages such as cost and ease of use (MST) and disadvantages such as high cost for colorimeters (for ITA) with traceable calibration. The strong complementary nature of MST and ITA led the JWG to define enrollment criteria as visualized in Figure 3, using both metrics to place a recruited participant in one of three bins: Light, Medium, or Dark (L, M and D, respectively). Each of the L, M and D bins should have at least 25% of all participants in a study, providing some flexibility to recruiters.The rationale for the criteria is given in the next sub-section. Rationale behind using the MST and ITA on the forehead for enrolling participants in pulse oximeter verification studies A point-by-point rationale for the choices is provided below. To assess a potential skin-color dependent SpO2 bias in PO devices, a balanced and representative enrollment in verification studies is required, where ‘representative’ is understood to express variations in participant appearance, complexion related to melanin pigmentation (light to dark skin). MST scale directly parameterizes this diversity and would be ideal as a participant enrollment tool if it were purely objective, e.g. with the use of an automated algorithm (23). Rater bias (24) are issues of some relevance considering that human assessments are subjective in nature. The subjectivity offers ‘wiggle room’ which might lead to bias, knowingly or unknowingly, to fulfill the requirements for specific pigmentation bins. ITA measured on the forehead is a plausible surrogate for MST as it estimates melanin content in skin on a location strongly related to appearance and complexion (12). Leveraging the complementary advantages of MST and ITA on the forehead, they will simultaneously be used as enrollment metrics to assign participants in bins L, M and D. The precision and accuracy characteristics of ITA (25) justifies using ITA as a “tie-breaker" if MST and ITA do not assign the same L, M and D enrollment bin. MST, simultaneously, is integral to the recruitment, enrollment, analysis process guaranteeing that people can relate to the ITA value (recruitment, enrollment, error-checking, communication, reporting-out to lay-people). There is an implicit suggestion in the first bullet point that appearance/complexion is a primary correlator with differential bias related to pigmentation; this is a conscious choice. (See comments in Discussion regarding device performance analysis) ITA can be prone to errors (16) and is visually color-blind; MST is helpful for double checking during all stages of the study (recruitment, enrollment, analysis). Justification of bin sizes, and targeted participant enrollment The cut-off values (degrees ITA) and associated MST grades for the bins were inspired by a RWS of PO (12) in a public safety net hospital in San Francisco where a significant proportion of the study population had ITA<-50 ◦ (12,26). This justified the additional condition for the dark bin (half of data should have ITA <-50 ◦ ) as a feasible recruitment condition in Western geographies. The range of ITA found in this study (12), is consistent with a recently published database of skin reflectance spectra and associated color values (17) (herein referred to as ISSA). It provides 2093 data points for the forehead, representing eight different ethnicities: Caucasian, Chinese, South Asian (Pakistani), African, Middle Eastern (Iraqi), Southeast Asian (Thai), Japanese and Middle Eastern (Arabian). Plotting all the 2093 data would appear as if the spread in the L*b* projection is much larger for medium skin than for dark and light skin, simply because more medium skin was measured. For the graph shown in Figure 3A, randomly selected data from ISSA with an equal density of points is shown for all ITA from -75° to 70°. With the approximate range of human skin color indicated by the blue dashed lines (ITA= -80° and 66°), the L, M, and D bins create an approximately equal sized representation across the span of human skin color ( i.e. , with a goal of balanced distribution of data). Note that the shape of the L*b* data point cloud is mostly encompassed by the yellow circle segments, consistent with the formula of the ITA and its ability to describe skin melanin pigmentation in a single convenient parameter (an angle). Discussion The participant enrollment criteria shown is the result of carefully weighing many pros and cons of many metrics which was only possible after the JWG had become familiar with MST and ITA, and also many other methods of skin tone characterization (25), each having their own specific properties. The next few sections are intended to share some of the aspects that were considered in the choices for MST and ITA. Discrepancy between L* values for MST and L* values from colorimetry on skin. Please note that the lightness L* of the MST grades for the three examples in Figure 4B (L* values for MST B, F and I are 92, 55 and 21, respectively, see Figure 2A) are considerably different from the L* as measured with a colorimeter: 68, 52 and 31, respectively. This is consistent with the phenomenon described recently (10): raters choose much lighter swatches when asked to match the forehead for light skin, and much darker swatches for dark skin. This phenomenon was consistently observed in four raters across two studies under different lighting conditions. Note that the mentioned L* values are not measured, and are for illustrative purposes only in describing the mentioned phenomenon (10) which can be described, in a simplified manner, as: light skin is perceived as lighter, and dark skin as darker. While it can be confusing to have two ranges for skin color (a narrower L* range for objective skin color, and a much larger L* range for the MST scale), it illustrates why including a scale like MST is important to use contextually in its relationship to ITA. When presenting the distribution of enrolled participants in clinical verification studies, MST colors are much more effective than ITA, and more effective than other skin-tone scales, as discussed in the next sections. Placing the MST into perspective Related to the discrepancy described in the previous section, recent studies suggest that a specific skin tone scales has one of two intended goals. The first goal is to determine appearance, or complexion. The MST and presumably also other scales (Figure 4) such as PERLA and Von Luschan (VL) are used by holding them next (e.g. at one meter distance) to a person and considering the entire persona, including contextual features such as hair color, but not on top of the skin. Other scales, like the Pantone® SkinTone™ (PST) Guide (27) and L'Oréal™ (13) are presumably intended to make a match with skin color by placing them on top of a certain anatomical location like the forehead or arm, and find the best matching swatch by looking through (a factory made) perforation in the swatch which presumably helps to minimize the context. The JWG recognized that human perception of skin color and objectively-measured skin color can be very different: light skin is perceived as much lighter, and dark skin is perceived as much darker (10). The MST scale, but also other scales like PERLA and VL, reflect this phenomenon by offering a much larger light-dark (L*) range than would correspond with a scale that represents the objective range of skin colors such as the PST (10,27). As a consequence, the instructions for use of this scale were recently refined (10): punching holes in a printed MST and placing it on the skin to make a color match is no longer recommended. This is important because the instructions for use for most scales are often absent, difficult to find, or ambiguous (27–31) while the results are highly dependent on how they are used. MST is preferred in the current application for several reasons: it has well defined colors that give a fair representation (24), is relatively well researched compared to other scales, has a certain momentum (Google and Meta adopted it to be used in machine learning (32–34)) , and has a convenient ten degree granularity with instructions for use that are better defined than for other scales (10). New scales are still being proposed, however, accompanied by research into factors involved in assessment of skin tone. Such factors include rater’s race and background color of the paper on which swatches are presented (e.g., grey instead of white) (31). Moreover, the MST colors focus on a variation in melanin only, as opposed to other scales which also express variations in dermal blood (redness). This is illustrated in Figure 4. The MST swatches from light to dark in the L* vs b* chart shows a (banana) shape quite similar to the variations from melanin as defined by Del Bino et al (15), albeit in a stretched form to account for the differences between perceived and objectively measured skin color (10). Scales such as PERLA or VL also have one or more swatches that are reddish, presumably to express the impact of blood variations on skin color, however, they offer only one degree of freedom. The single-axis mixes melanin and blood variation and thus creates a non-monotonic relationship with objective melanin measurements (35). We emphasize that scales with one or more reddish swatches (e.g. PERLA) are perhaps less preferred for the PO purpose (to describe variations in melanin) but may be preferred in other fields where describing variations in redness as well as melanin is desired. Scales such as PST and L'Oréal™ implicitly acknowledge the two main skin chromophores (melanin and blood, see Figure 1A) by introducing two axes. Such scales, however, are cumbersome to be used for our purpose and a poor imprecise surrogate for colorimetry. More importantly, they would not serve the communication purpose offered by MST because they match objectively measured skin colors, which differ from those perceived by humans in everyday life: i.e. skin color is seen in context: face, hair, other exposed body parts. Humans perceive skin color quite differently when it is in context (e.g., the face) versus when it is printed on a sheet of paper with a number of swatches (10). And even under the same conditions, different people may see them differently (36). Many mechanisms (37,38) may play a role in this phenomenon, including memory effects, facial features, and expectations (39–41). In this context, it is important to realize that also objective measurements of skin can have many dimensions and that the approach of using ITA is a conscious simplification. ITA is precise, accurate and reproducible across the world, but many factors that can be objectively identified are implicitly discarded or minimized in the color measurement. This is illustrated in Figure 5 and leads us to the next section. Placing the ITA into perspective Skin color homogeneity, hair, sebum (enhancing specular reflection), surface roughness are examples of factors that are likely to impact human perception of the skin. It is not feasible, and unnecessary, to measure each of these parameters. Skin color measured on the forehead by a colorimeter and reduced to an ITA, for the purpose of the standard, is a suitable surrogate for general appearance or complexion as determined by MST. Moreover, ITA was shown to correlate well with melanin concentration (5,15), arguably the dominant parameter in the MST scale (see previous section) While Figure 5 expresses the complexity of objectively measurable parameters involved in skin color, there are still many more parameters that relate to how humans actually perceive skin color when context is not minimized and the skin is observed in the context of a face, including hair color, eye color, etc. These parameters are more difficult to determine as they involve psycho-physical experiments (10,41) and include factors such as regional background (24) , experience, socio-economic status (of both participant and rater?). The JWG aimed to balance rigor with pragmatism/feasibility and make choices with respect to hypotheses, which leads us to the next section. The rationale behind using MST, and ITA on the forehead was provided in one of the previous sections. For completeness and increased understanding of the JWG’s approach to the tasks, listed here are alternative hypotheses that were discussed but not adopted for participant enrollment. Alternative hypotheses We are aware that we chose to focus on pigmentation differential bias. There may be many other differential biases. Examples of alternative hypotheses that were considered but not selected for this draft are listed below. Ethnicity is the primary correlator with observed differential bias in RWS. Race is the primary correlator with the observed differential bias in RWS. A combination of low perfusion and skin pigmentation (possibly correlated through ethnicity) is the primary correlator with differential bias in RWS(43) Epidermal thickness is the primary correlator with observed differential bias in RWS. Nail thickness is the primary correlator with observed differential bias in RWS. While some of the parameters mentioned above are not used for enrolling study participants, they will be measured or recorded, if possible, and used in the analysis for characterizing differential bias to elucidate the root cause expediently, to produce devices that perform equally well for all skin colors, ethnicities and races. Of note, the JWG uses slightly different working hypotheses for the two identified tasks . For the participant enrollment criteria the JWG used MST and melanin pigmentation (ITA) at the forehead while for the second task, creating the pigmentation differential bias metric, melanin pigmentation at the probe site is used. This reflects the early stage of research into the root cause of the RWS observations, as well as the democratic and open character of the discussions that took place within the JWG. Moreover, the ITA on the forehead is quite strongly correlated with the ITA measured on the distal phalange (12), considered to be a good representation of the pigmentation on the sensor site for finger probes. Conclusions The draft revision of the ISO 80601-2-61 standard requires verifying pulse oximeter performance in a controlled desaturation study with enrolled participants falling into the three skin color bins, L, M and D, shown in Fig. 3 B in similar numbers. The use of a relatively small number of bins provides a good balance between participant recruitment/enrollment feasibility and scientific rigor. Experience has shown that it can be difficult to recruit people with darker skin. There is an additional requirement that at least half of participants in the ‘Dark’ bin have an ITA < -50° ( 21 , 22 ) to ensure this proportion of the skin tone category is not skewed to the lighter margin. The JWG carefully and exhaustively weighed factors involved in participant enrollment and considered the bins as defined in Fig. 3 to be such that inclusion in verification studies will be fair and representative of the global population. With the current state of knowledge, the JWG considers complexion, appearance, ‘gestalt’ to be the preferred metric to ensure enrolling participants across a wide distribution of skin tones. The MST directly correlates with this metric and describes it well in relatable values/colors and has high levels of inter-rater reliability ( 24 ). However, the subjective nature of annotation by humans also allows for various sources of bias: regional cultural contexts ( 24 ) may influence how people perceive skin tone. As a consequence, subjective methods to assess pigmentation are currently considered to have too much uncertainty to be adopted as the decisive metric for participant enrollment ( 25 ) in verification studies for PO. Simultaneously, the JWG acknowledges that effective communication to lay-audiences is critically important, and that ITA is not suitable for that purpose. Moreover, measuring ITA is itself prone to errors. Therefore, the updated ISO draft recommends using both ITA and MST characterized on the forehead as metrics for participant enrollment, where ITA will be used as the tie-breaker for the cases when MST and ITA binning are not consistent with one another. ITA should be measured with devices that have traceable calibration. This enrollment methodology is not limited to only clinical laboratory studies with healthy adult participants. The measurement protocol should be practical for critically ill patients at the bedside for the entire age spectrum from pre-term newborns to adults ( 44 ). In this context, the JWG is currently exploring how smaller, lighter, affordable, and user-friendlier colorimeters can be used. Typically, these less costly colorimeters are sufficiently accurate and precise for the purposes stated herein. Traceable calibration is often lacking for these less costly devices; it may be possible to create a verification/calibration method to ensure the measurements are reliable. Finally, at this stage of knowledge, it is premature to conclude what the root cause(s) for the RWS observations are due to, or even the perfect metric to ensure diverse participant enrollment. Scientific evidence will continue to drive the rationales behind future ISO device standards in an ongoing effort to improve this facet of pulse oximeter performance. Declarations Acknowledgements We thank Sandy Weininger, Mike Lipnick for their critical reading of the manuscript and valuable suggestions. We also thank all JWG members whose diligent work has helped shape the current draft standard as well as OpenOximetry (openoximetry.org) for gathering and sharing knowledge and study data. DISCLOSURES: Prof. Monk is the developer of the Monk SkinTone Scale. Drs. Verkruysse, Milkes, Mannheimer and Kopotic work for commercial enterprises: Philips, Medtronic, Apple and BD/EdwardsLifesciences, respectively. CONFLICTS OF INTEREST: NA CLINICAL TRIAL NUMBER AND REGISTRY URL. Not applicable ABBREVIATED TITLE: Not applicable AUTHOR CONTRIBUTIONS Wim Verkruysse initiated and helped write the manusript. All other authors helped write the manuscript. References ISO 80601-2-61 Medical Electrical Equipment —Part 2-61: Particular requirements for basic safety and essential performance of pulse oximeter equipment; expected to be formally published in early 2026. 2025. Sjoding MW, Dickson RP, Iwashyna TJ, Gay SE, Valley TS. Racial Bias in Pulse Oximetry Measurement. N Engl J Med. 2020 Dec 17;383(25):2477–8. Moran-Thomas A. How a Popular Medical Device Encodes Racial Bias. Boston Review [Internet]. 2020 Aug 5 [cited 2025 Apr 2]; Available from: http://bostonreview.net/science-nature-race/amy-moran-thomas-how-popular-medical-device-encodes-racial-bias. Jubran A, Tobin MJ. Reliability of Pulse Oximetry in Titrating Supplemental Oxygen Therapy in Ventilator-Dependent Patients. CHEST. 1990 Jun 1;97(6):1420–5. Alaluf S, Atkins D, Barrett K, Blount M, Carter N, Heath A. The Impact of Epidermal Melanin on Objective Measurements of Human Skin Colour. Pigment Cell Research. 2002 Apr;15(2):119–26. Verkruysse W, Lucassen GW, van Gemert MJC. Simulation of color of port wine stain skin and its dependence on skin variables. Lasers in Surgery and Medicine. 1999;25(2):131–9. Monk EP. The Monk Skin Tone Scale. 2019 [cited 2024 Apr 19]; Available from: https://osf.io/preprints/socarxiv/pdf4c/ Chardon A, Cretois I, Hourseau C. Skin colour typology and suntanning pathways. International Journal of Cosmetic Science. 1991;13(4):191–208. Vasudevan S, Vogt WC, Weininger S, Pfefer TJ. Melanometry for objective evaluation of skin pigmentation in pulse oximetry studies. Commun Med. 2024 Jul 11;4(1):138. Verkruysse W, Monk EPJ, Jaffe MB. Objective and Perceived Skin Color: Consequences for the Use of Skintone Scales. Anesthesia & Analgesia. 2025 Jan;140(1):e2. Groen S, Verkruysse W, Jaffe M, Muehlsteff J. Skin Color at Pulse Oximeter Measurement Sites Varies Considerably Between and Within Individuals. Poster presented at: AACN NTI National Teaching Institute 2024; 2024 May 29; Denver, CO. Lipnick MS, Chen D, Law T, Moore Jr K, Lester JC, Monk EP, et al. Comparison of methods for characterizing skin pigment diversity in research cohorts. medRxiv. 2025;2025–02. Gordon RA, Branigan AR, Khan MA, Nunez JG. Best Practices for Measuring Skin Color in Surveys. Survey Practice [Internet]. 2024 May 23 [cited 2025 Jan 19];17. Available from: https://www.surveypractice.org/article/94628-best-practices-for-measuring-skin-color-in-surveys Standard, C. I. E. (2007). Colorimetry-part 4: CIE 1976 L* a* b* colour space. International Standard, 2019-06. Del Bino S, Bernerd F. Variations in skin colour and the biological consequences of ultraviolet radiation exposure. Br J Dermatol. 2013;169(s3):33–40. Verkruysse W, Jaffe MB. Letter Regarding the Potential Improper Use of the Individual Typology Angle in the Context of Disparate Bias in Pulse Oximetry. Anesthesia & Analgesia. 2024 May;138(5):e31. Lu Y, Xiao K, Pointer M, He R, Zhou S, Nasseraldin A, et al. The International Skin Spectra Archive (ISSA): a multicultural human skin phenotype and colour spectra collection. Sci Data [Internet]. 2025 Mar 23 [cited 2025 Mar 29];12(1). Available from: https://www.nature.com/articles/s41597-025-04857-5 Pexels [Internet]. [cited 2025 Apr 3]. Photo by Polina Tankilevitch on Pexels. Available from: https://www.pexels.com/photo/woman-in-white-top-wearing-blue-knit-cap-4723530/ Pexels [Internet]. [cited 2025 Apr 3]. Photo by Centre for Ageing Better on Pexels. Available from: https://www.pexels.com/photo/woman-leaning-head-on-hand-17153160/ Pexels [Internet]. [cited 2025 Apr 3]. Photo by António Ribeiro on Pexels. Available from: https://www.pexels.com/photo/portrait-of-woman-with-crossed-arms-10619795/ Fong N, Lipnick M, Bickler P, Feiner J, Law T. OpenOximetry Repository [Internet]. PhysioNet; [cited 2025 May 6]. Available from: https://physionet.org/content/openox-repo/1.0.1/ Leeb G, Auchus I, Law T, Bickler P, Feiner J, Hashi S, et al. The performance of 11 fingertip pulse oximeters during hypoxemia in healthy human participants with varied, quantified skin pigment. eBioMedicine [Internet]. 2024 Apr 1 [cited 2025 May 6];102. Available from: https://www.thelancet.com/journals/ebiom/article/PIIS2352-3964(24)00086-0/fulltext Rejón Piña RA, Ma C. Classification Algorithm for Skin Color (CASCo): A new tool to measure skin color in social science research. Social Science Quarterly. 2023;104(2):168–79. Schumann C, Olanubi GO, Wright A, Monk Jr. E, Heldreth C, Ricco S. Consensus and Subjectivity of Skin Tone Annotation for ML Fairness [Internet]. arXiv; 2023 [cited 2023 Dec 19]. Available from: http://arxiv.org/abs/2305.09073 Verkruysse W, Jaffe MB, Lipnick M, Zemouri C. Challenges of Subjective Skin Color Scales: The Case for the Use of Objective Pigmentation Measurement Methods in Regulatory Pulse Oximetry Studies. Anesthesia & Analgesia. 2024;10–1213. The Open Oximetry March 13, 2024 – Skin Color Quantification Subgroup Meeting 3 accessed via https://youtu.be/6ELI__1XcbI. Pantone. Pantone. [cited 2024 Feb 6]. PANTONE SkinTone Guide. Available from: https://www.pantone.com/skintone Cohen PR, DiMarco MA, Geller RL, Darrisaw LA. Colorimetric Scale for Skin of Color: A Practical Classification Scale for the Clinical Assessment, Dermatology Management, and Forensic Evaluation of Individuals With Skin of Color. Cureus [Internet]. 2023 Nov 1 [cited 2024 Mar 4]; Available from: https://www.cureus.com/articles/201038-colorimetric-scale-for-skin-of-color-a-practical-classification-scale-for-the-clinical-assessment-dermatology-management-and-forensic-evaluation-of-individuals-with-skin-of-color Telles EE. Multiple measures of ethnoracial classification in Latin America. Ethnic and Racial Studies. 2017 Oct 21;40(13):2340–6. Felix von Luschan. In: Wikipedia [Internet]. 2024 [cited 2025 Jan 19]. Available from: https://en.wikipedia.org/w/index.php?title=Felix_von_Luschan&oldid=1241316357 Cook C, Howard J, Rabbitt L, Shuggi I, Sirotin Y, Tipton J, et al. Colorimetric skin tone scale for improved accuracy of human skin tone annotations. ACM J Responsib Comput [Internet]. 2025 May 3 [cited 2025 May 15]; Available from: https://dl.acm.org/doi/10.1145/3730409 Google researchers, Ellis Monk. Skin Tone Research @ Google AI | Start using the Monk Skin Tone Scale [Internet]. [cited 2023 Dec 7]. Available from: https://skintone.google Gustafson L, Rolland C, Ravi N, Duval Q, Adcock A, Fu CY, et al. FACET: Fairness in Computer Vision Evaluation Benchmark. In 2023 [cited 2025 May 15]. p. 20370–82. Available from: https://openaccess.thecvf.com/content/ICCV2023/html/Gustafson_FACET_Fairness_in_Computer_Vision_Evaluation_Benchmark_ICCV_2023_paper.html Porgali B, Albiero V, Ryda J, Ferrer CC, Hazirbas C. The casual conversations v2 dataset. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition [Internet]. 2023 [cited 2025 May 15]. p. 10–7. Available from: https://openaccess.thecvf.com/content/CVPR2023W/TCV/html/Porgali_The_Casual_Conversations_v2_Dataset_CVPRW_2023_paper.html Swiatoniowski AK, Quillen EE, Shriver MD, Jablonski NG. Technical Note: Comparing von Luschan skin color tiles and modern spectrophotometry for measuring human skin pigmentation. American Journal of Physical Anthropology. 2013;151(2):325–30. Bosten JM. Do You See What I See? Diversity in Human Color Perception. Annu Rev Vis Sci. 2022 Sep 15;8(1):101–33. Baker LJ, Levin DT. The Face-Race Lightness Illusion Is Not Driven by Low-level Stimulus Properties: An Empirical Reply to Firestone and Scholl (2014). Psychon Bull Rev. 2016 Dec;23(6):1989–95. Firestone C, Scholl BJ. Cognition does not affect perception: Evaluating the evidence for “top-down” effects. Behav Brain Sci. 2016;39:e229. Lyngs U, Cohen E, Hattori WT, Newson M, Levin DT. Hearing in color: How expectations distort perception of skin tone. Journal of Experimental Psychology: Human Perception and Performance. 2016 Dec;42(12):2068–76. Hansen T, Olkkonen M, Walter S, Gegenfurtner KR. Memory modulates color appearance. Nat Neurosci. 2006 Nov;9(11):1367–8. Levin DT, Banaji MR. Distortions in the perceived lightness of faces: The role of race categories. J Exp Psychol. 2006;135(4):501–12. Dlugos JF, Taylor JL. Materials Characterization: UV/Vis/NIR Spectroscopy [Internet]. Perkin Elmer; 2019 [cited 2024 Apr 15]. Available from: https://resources.perkinelmer.com/lab-solutions/resources/docs/App_Visible-Reflectance-Spectroscopy-Human-Skin.pdf Gudelunas MK, Lipnick M, Hendrickson C, Vanderburg S, Okunlola B, Auchus I, et al. Low Perfusion and Missed Diagnosis of Hypoxemia by Pulse Oximetry in Darkly Pigmented Skin: A Prospective Study. Anesthesia & Analgesia. 2022 Mar 18;10.1213/ANE.0000000000006755. Sharma M, Pickhardt AJ, Madison, Tong L, Evans MS. Objective Assessment and Quantification of Skin Color and Melanin in Neonates and Infants: A State-Of-The-Art Review. Pediatric Dermatology [Internet]. [cited 2025 Apr 2];n/a(n/a). Available from: https://onlinelibrary.wiley.com/doi/abs/10.1111/pde.15902 Table Table 1 Key advantages and disadvantages of MST and ITA for assessing skin color of participants in pulse oximeter verification studies MST ITA Advantages Communication to interested parties on enrollment characteristics as the colors relate directly to how humans perceive skin color/complexion. In line with good laboratory practice of using a calibrated device to record an objective, precise, accurate, repeatable measurement with negligible rater bias (25). Focuses on melanin, which is the variable of interest of the working hypothesis. Disadvantages Subjective nature: metric collected from different human raters can add variability in observed MST grades. Improving precision is cumbersome. Rater bias, willful or not, is difficult to eliminate. Standardized printing is cumbersome. Cumbersome to be used in communication because objectively measured skin colors, and ITA even more so, poorly relate to how people experience skin color and the range of skin color. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 09 Aug, 2025 Reviewers invited by journal 17 Jul, 2025 Editor assigned by journal 15 Jul, 2025 Submission checks completed at journal 15 Jul, 2025 First submitted to journal 13 Jul, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7112852","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Short Report","associatedPublications":[],"authors":[{"id":498022269,"identity":"5c11e40f-3167-400f-9fd3-76299d8d9fdf","order_by":0,"name":"Wim Verkruysse","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEElEQVRIiWNgGAWjYDACZghlwMDA2AAkGOQYJBgbDySQosUYqKUBvxYGuBYISGyQYGA4gFfpce4EZp6Ke8bmM5IbCn/8sUnfLt3ccODhjjoG/vZurJYZHObdwMxzpthM5kZigzFvW1ruzjkHGw4knjnMIHHm7AZsWsyAWhhntiXYSEgAtTA2HM7dANR7ILHtAIOBRC4eLf8gWgx//PmfbgDRUodXC8PHhgQzkBYDHrYDCVAtzDi12AO1HPhwLMFYguchyC/JhjtnJIL9woPLL5L9Zzc+SKhJMJzBnv4M6DA7eXOJ9IcPf+6ok+Nv78WqBQQOQGk2cOSACcYGBh5cypEB8wNkLaNgFIyCUTAKYAAAvvNrE/Y63hwAAAAASUVORK5CYII=","orcid":"","institution":"Philips (Netherlands)","correspondingAuthor":true,"prefix":"","firstName":"Wim","middleName":"","lastName":"Verkruysse","suffix":""},{"id":498022271,"identity":"105e0aa8-2942-441e-bce1-89a275c09aae","order_by":1,"name":"Bernice E. Rogowitz","email":"","orcid":"","institution":"Visual Perspective, Editor In Chief, Journal of Perceptual Imaging (JPI)","correspondingAuthor":false,"prefix":"","firstName":"Bernice","middleName":"E.","lastName":"Rogowitz","suffix":""},{"id":498022274,"identity":"5bb63d46-6225-44dd-8279-b51756cdb234","order_by":2,"name":"David H. Milkes","email":"","orcid":"","institution":"Medtronic","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"H.","lastName":"Milkes","suffix":""},{"id":498022276,"identity":"7692117e-9e1b-4e1f-9c3f-61a06ca6c346","order_by":3,"name":"Paul Mannheimer","email":"","orcid":"","institution":"Apple Inc. Cupertino CA","correspondingAuthor":false,"prefix":"","firstName":"Paul","middleName":"","lastName":"Mannheimer","suffix":""},{"id":498022277,"identity":"b8948e55-a55d-45d3-a2ad-bfd4a71a5c72","order_by":4,"name":"Robert J. Kopotic","email":"","orcid":"","institution":"Edwards Lifesciences (United States)","correspondingAuthor":false,"prefix":"","firstName":"Robert","middleName":"J.","lastName":"Kopotic","suffix":""},{"id":498022278,"identity":"6ea00142-93f5-4a64-b42a-98fa76a47e36","order_by":5,"name":"Ellis P. Monk","email":"","orcid":"","institution":"Harvard University","correspondingAuthor":false,"prefix":"","firstName":"Ellis","middleName":"P.","lastName":"Monk","suffix":""}],"badges":[],"createdAt":"2025-07-13 11:08:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7112852/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7112852/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93074529,"identity":"06128391-2dd7-4c63-8d5d-961b0442456d","added_by":"auto","created_at":"2025-10-08 18:41:03","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2525147,"visible":true,"origin":"","legend":"","description":"","filename":"ShortPaperonISOrationaleJClinCompv1nofields.docx","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/53e2f86b2122bab7ae49eea5.docx"},{"id":93074531,"identity":"4aff6563-f01b-4168-927e-40deb876802c","added_by":"auto","created_at":"2025-10-08 18:41:03","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7577,"visible":true,"origin":"","legend":"","description":"","filename":"84bab40ac2574244af5271d0b16f6326.json","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/57fe5523cbc948dba1b06185.json"},{"id":93074532,"identity":"3f63413f-0e9f-4405-ab88-945c43e50c6f","added_by":"auto","created_at":"2025-10-08 18:41:04","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":104041,"visible":true,"origin":"","legend":"","description":"","filename":"84bab40ac2574244af5271d0b16f63261enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/67946deaa64094457b76c1bc.xml"},{"id":93074537,"identity":"e99eec68-eb03-40c6-b430-d3b94dd9512e","added_by":"auto","created_at":"2025-10-08 18:41:05","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":961302,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/c21bffff232b60558c18bec0.png"},{"id":93074536,"identity":"318720fd-9bad-4e77-9344-6c4398864b2b","added_by":"auto","created_at":"2025-10-08 18:41:04","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":641688,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/ff99d253d0a3d2ad6e7873a9.png"},{"id":93074516,"identity":"d041c143-53db-4bd2-b9a1-22496029e247","added_by":"auto","created_at":"2025-10-08 18:41:01","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":422459,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/a2d9cb2c2f35a4998a220079.png"},{"id":93074512,"identity":"8cb6d720-4801-4006-974f-5de9bea92add","added_by":"auto","created_at":"2025-10-08 18:40:59","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":294979,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/b5452641da6e19468c38f108.png"},{"id":93074513,"identity":"50567c2e-8b09-40b5-8682-b35d1b38b160","added_by":"auto","created_at":"2025-10-08 18:40:59","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":120988,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/19ae15ffde8467135b1f61c8.png"},{"id":93074546,"identity":"67e86c0b-20f3-4a22-94a9-66beeb556b21","added_by":"auto","created_at":"2025-10-08 18:41:07","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":160829,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/f98b5e1682cc90495b8ea393.png"},{"id":93074519,"identity":"0451eebe-42bb-4a05-86e8-d1622e0f5920","added_by":"auto","created_at":"2025-10-08 18:41:01","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":108412,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/eb7213dab93fa190595cbe1c.png"},{"id":93074556,"identity":"66eb3871-dba3-4a96-8ce3-167f311c6057","added_by":"auto","created_at":"2025-10-08 18:41:11","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93786,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/c8138d4b011c957db07fb6b6.png"},{"id":93075097,"identity":"2219f7db-69ec-4a1a-be22-dd5b1ca9346e","added_by":"auto","created_at":"2025-10-08 19:01:43","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":71541,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/6ea6c482b71a1f2a16ff0f59.png"},{"id":93074541,"identity":"f15df2d5-218d-48c6-bbf3-d66ab6ad07ab","added_by":"auto","created_at":"2025-10-08 18:41:05","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42966,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/1aaa9e62d81145a519d2158d.png"},{"id":93074555,"identity":"aa68b1bc-c62d-4f25-ab1d-70dc1ff53361","added_by":"auto","created_at":"2025-10-08 18:41:10","extension":"xml","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":100529,"visible":true,"origin":"","legend":"","description":"","filename":"84bab40ac2574244af5271d0b16f63261structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/7c2b692d5d7dc5a37f044af6.xml"},{"id":93074543,"identity":"8818918a-5c87-454d-b48c-1196467c25c7","added_by":"auto","created_at":"2025-10-08 18:41:06","extension":"html","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":114720,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/b443f6a23758f7a8bb808852.html"},{"id":93075092,"identity":"992c3ad9-cd1d-4349-848c-b0cbb42d92e1","added_by":"auto","created_at":"2025-10-08 19:01:43","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":961302,"visible":true,"origin":"","legend":"\u003cp\u003eIllustrations of A) dominant skin chromophores, B) absorption spectra, light penetration depths and resulting reflectance spectra and C) the pulse oximeter (PO) working principle (simplified). Red and orange arrows in B indicate wavelengths used in PO.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/7af12d4b2de6e94a2700663c.png"},{"id":93074558,"identity":"75c95153-a405-4dc9-82d4-782da7ce1d39","added_by":"auto","created_at":"2025-10-08 18:41:11","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":641688,"visible":true,"origin":"","legend":"\u003cp\u003eAn illustration of the two pigmentation measurement methods and how they are used in actual practice. The Monk SkinTone (MST) scale, shown in A, should be held somewhat near the face of a participant, after which the rater makes a choice. Pertinently, the scale should not be held on top of the skin after which the skin might be assessed through a hole in the scale; this leads to quite different assessments. In B, an illustration of colorimetry and Individual Typology Angle (ITA). The illustration of measurement on skin is shown on an arm, for privacy purposes only. The draft standard describes that ITA should be measured on the forehead, to match with the MST assessment on the face/complexion. Note – MST L*a*b* values shown in panel A refer to the A-J colors needed to replicate the scale, not the associated colorimetric values measured on skin; see also discussion.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/4702c0a02292551d67663c60.png"},{"id":93074514,"identity":"ca094647-f807-474a-83b7-d21f2569614e","added_by":"auto","created_at":"2025-10-08 18:41:01","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":422459,"visible":true,"origin":"","legend":"\u003cp\u003eA) An illustration of how L* and b* define an Individual Typology Angle (ITA), shown as a look-up table. Green dots are data from a skin color data-base17for the forehead, illustrating the range of ITA for human skin: approx. between -82 and 66 degrees. Yellow circle-segments, centered at L* = 50, and b* = 0, consistent with the ITA formula, encompass most data points. B) Three ITA bins are defined for participant enrollment in pulse oximeter verification: Light, Medium and Dark, acknowledging the range and distribution of human skin color on the forehead. \u0026nbsp;Each of the L, M and D bins should have at least 25% of all participants in a study, providing some flexibility to recruiters. Monk Skintone (MST) values are associated with the bins to ensure effective recruitment and communication (ITA is less familiar and comprehensible to most interested parties). Pictures of persons shown are Creative Commons18–20. The subdivision of the ‘D bin‘ in dark and very dark at ITA\u0026lt;-50 degrees21,22 is to ensure that very dark persons are included.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/ba4ebac5755508b372b4a101.png"},{"id":93074553,"identity":"e3a8d5f8-6a8f-40bb-93ba-a50903eff35e","added_by":"auto","created_at":"2025-10-08 18:41:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":294979,"visible":true,"origin":"","legend":"\u003cp\u003eA) Illustration of various skin tone scales. Abbreviations used: Pantone® SkinTone™ guide (PST), Fitzpatrick (FP), Von Luschan (VL), Project on Ethnicity and Race in Latin America (PERLA) and Monk SkinTone (MST). The PST and L'Oréal scales are presumably used in ‘color matching mode’, i.e., placing them on skin and finding the best match. The MST scale is used as an ‘appearance’ scale, i.e., it is held next to the face, but not on top of the skin. In B) and C) the MST scale stands out from FP, VL and PERLA by its monotonic behavior, and having a shape to the skin range. The skin data shown is from (17) and shows only forehead data (n = 2093). For L'Oréal scale, albino skin data we used references (13) and (42), respectively. PERLA colors are from personal communication with Dr. Telles and for FP we used an arbitrary version of the internet which we printed on a color printer and measured the colors.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/e269a6b02e2e1188ceadba2a.png"},{"id":93074571,"identity":"e1a1e442-3a72-44ac-aab1-8937871c275b","added_by":"auto","created_at":"2025-10-08 18:41:14","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":120988,"visible":true,"origin":"","legend":"\u003cp\u003eDescribing the appearance of skin diversity can have many dimensions. This overview only shows physical characteristics. For the purpose of characterizing skin tone due to melanin pigment, the one-dimensional Individual Typology Angle (ITA) is a suitable metric.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/19635dcbbd10cd01e4e758e5.png"},{"id":93077921,"identity":"c55fb8f3-97e5-4751-9b1a-2237d832a4d0","added_by":"auto","created_at":"2025-10-08 21:29:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2852358,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7112852/v1/c560aa17-cbb9-4662-bd76-cbdd170161bc.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Towards achieving even distributions of participant skin tones when verifying pulse oximeter performance","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAs an organization, \u003cem\u003eInternational Organization for Standardization\u003c/em\u003e (ISO) aims to improve the performance of pulse oximeters, a device that non-invasively estimates the hemoglobin oxygen saturation of arterial blood. Stakeholders in the ISO Oximeters joint work groups (JWG) , include clinicians, manufacturers, regulators and other experts who collaboratively develop and maintain the standards. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA recently published draft\u0026nbsp;(1)\u0026nbsp;on standardization of pulse oximetry (PO) includes significant changes to address the potential health inequity related to differences in SpO\u003csub\u003e2\u0026nbsp;\u003c/sub\u003ereadings between light and dark skin groups, that have been observed in real world studies (RWS)\u0026nbsp;(2\u0026ndash;4). In the process of updating the ISO standard to its third edition, the JWG agreed that the underlying root cause(s) for the observed bias differences remain(s) unknown. More specifically, whether the degree of melanin pigmentation is the direct root cause, or merely a strong correlator with another confounding variable, is not known. The JWG identified two main objectives, each with a slightly different working hypothesis on the root cause. In both hypotheses it is acknowledged that SpO\u003csub\u003e2\u003c/sub\u003e bias may have been missed in earlier device verification studies, where dark skin individuals were under-represented.\u003c/p\u003e\n\u003cp\u003eThe first objective was to ensure a balanced and wider distribution of skin tone among the participants enrolled in pulse oximeter verification studies. The second objective was to create the methodology to characterize the degree of difference in SpO\u003csub\u003e2\u003c/sub\u003e readings between dark and light skin participants, defined as \u003cem\u003epigmentation\u003c/em\u003e \u003cem\u003edifferential bias\u003c/em\u003e, observed in these verification studies. While these two objectives complement one another, they address different aspects of the task \u0026ndash; study participant enrollment and device characterization.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHere, we focus on the first objective: study participant enrollment. The main goal of this paper is to share the rationale, and a description of the pigmentation metrics used as they are likely unfamiliar to many. In the section Enrollment Methodology, these metrics are presented in a concise manner. They will be placed in a broader context in the Discussion, for a more thorough understanding of the metrics. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSince PO and the metrics to quantify pigmentation are both rooted in tissue/skin optics, we briefly address a few basic elements, shown in Figure 1. Skin has two dominant chromophores: hemoglobin (Hb) in blood, and melanin. Deoxy-hemoglobin and oxy-hemoglobin \u0026nbsp;(HHb and O2Hb, respectively) are the most relevant Hb species for the purpose of PO. Melanin, located in the epidermis, varies strongly in concentration between people, resulting in skin tones that span from very light to very dark (5). Absorption of light by blood and melanin give skin its characteristic range of colors. Besides absorption, light scattering also influences the penetration depths of each wavelength; as a general principle, the longer wavelengths penetrate deeper in skin. As a result, both the concentration and the distribution of chromophores have an impact on the skin color \u0026nbsp;(6). The reflectance spectra shown in Figure 1B illustrate that the impact of blood on the skin color becomes gradually smaller with increasing levels of melanin: the oxy-hemoglobin absorption peaks at 545nm and 577nm have a less pronounced effect for darker skin. The concept of PO is shown in Figure 1C. Red and infra-red light emitted by light emitting diodes (LEDs) and detected in transmission (or reflectance, in devices with different form factor) is modulated in time by the pulsatile volume of arterioles. The absorption by melanin at these wavelengths is much smaller than at the shorter visible wavelengths. Nevertheless, the absorption is non-zero, and the distinctly higher absorption in red compared to infra-red may affect light paths of red and infra-red differently. The respective modulation amplitudes might also be affected and consequently cause a pigmentation differential bias.\u0026nbsp;\u003c/p\u003e"},{"header":"Enrollment methodology ","content":"\u003ch2\u003eProposed pigmentation metrics\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eTwo pigmentation metrics are used in the new standard to ensure study enrollment spans the wide range of human skin tones: the Monk Skin Tone (MST) scale\u0026nbsp;(7)\u0026nbsp;and the individual typology angle (ITA)\u0026nbsp;(8). Illustrations of MST and ITA are provided in\u0026nbsp;Figure 2. The main difference between these two metrics is that the MST is \u003cem\u003esubjective\u003c/em\u003e while ITA is \u003cem\u003eobjective\u0026nbsp;\u003c/em\u003e(9). A second difference is that MST aims at expressing pigmentation as the appearance, complexion, or \u0026lsquo;gestalt\u0026rsquo; of an individual as a whole, not a specific skin spot\u0026nbsp;[10].\u003c/p\u003e\n\u003cp\u003eITA expresses melanin pigmentation on a certain anatomical location, which can vary considerably within an individual [11,12]. The ITA is measured by placing a color-sensing device (spectrophotometer/colorimeter) on the skin, evaluating a circular spot of several mm in diameter [13], depending on the device. By measuring how much of the device\u0026rsquo;s illuminating light is reflected at different wavelengths, an objective skin color is determined and expressed in the CIELab standardized color space (14) using three parameters: L*, a*, and b*. The lightness value L* ranges from black (value 0) to white (value 100). Parameters a* and b* represent the chromaticity where a* represents the balance between magenta/red (positive values) and green (negative values). Similarly, b* represents the balance between yellow (positive values) and blue (negative values). Skin typically has positive values for a* and b* as it is reddish/yellowish rather than greenish/blueish. The ITA conveniently expresses melanin pigmentation in a one-dimensional parameter (an angle), noting the curved shape in the Figure 2 B graph of L* vs b* in individuals with differing melanin pigmentation [15]. A user wishing to measure ITA computes it from L* and b* (implicitly discarding a*) using the formula [8] shown above the graph. To minimize errors in the implementation of the formula [16] it is recommended to do this in a spreadsheet and perform a sanity check by comparing the results with known ITA values.\u003c/p\u003e\n\u003cp\u003eBesides the subjective/objective difference, it became clear from discussions and research, that MST and ITA both have distinct advantages and disadvantages (Table 1). The JWG approach leverages their complementary qualities.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNote that\u0026nbsp;Table 1\u0026nbsp;contains only the \u003cem\u003ekey\u003c/em\u003e-advantages for the stated\u003cem\u003e\u0026nbsp;\u003c/em\u003epurpose of participant enrollment in PO verification studies. For many other purposes one can identify other key advantages such as cost and ease of use (MST) and disadvantages such as high cost for colorimeters (for ITA) with traceable calibration.\u003c/p\u003e\n\u003cp\u003eThe strong complementary nature of MST and ITA led the JWG to define enrollment criteria as visualized in Figure 3, using \u003cem\u003eboth\u003c/em\u003e metrics to place a recruited participant in one of three bins: Light, Medium, or Dark (L, M and D, respectively). Each of the L, M and D bins should have at least 25% of all participants in a study, providing some flexibility to recruiters.The rationale for the criteria is given in the next sub-section.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eRationale behind using the MST and ITA on the forehead for enrolling participants in pulse oximeter verification studies\u003c/h2\u003e\n\u003cp\u003eA point-by-point rationale for the choices is provided below.\u0026nbsp;\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eTo assess a potential skin-color dependent SpO2 bias in PO devices, a balanced and representative enrollment in verification studies is required, where \u0026lsquo;representative\u0026rsquo; is understood to express variations in participant appearance, complexion related to melanin pigmentation (light to dark skin).\u003c/li\u003e\n \u003cli\u003eMST scale directly parameterizes this diversity and would be ideal as a participant enrollment tool if it were purely objective, e.g. with the use of an automated algorithm\u0026nbsp;(23). Rater bias\u0026nbsp;(24)\u0026nbsp;are issues of some relevance considering that human assessments are subjective in nature. The subjectivity offers \u0026lsquo;wiggle room\u0026rsquo; which might lead to bias, knowingly or unknowingly, to fulfill the requirements for specific pigmentation bins.\u003c/li\u003e\n \u003cli\u003eITA measured on the forehead is a plausible surrogate for MST as it estimates melanin content in skin on a location strongly related to appearance and complexion\u0026nbsp;(12).\u003c/li\u003e\n \u003cli\u003eLeveraging the complementary advantages of MST and ITA on the forehead, they will simultaneously be used as enrollment metrics to assign participants in bins L, M and D.\u003c/li\u003e\n \u003cli\u003eThe precision and accuracy characteristics of ITA\u0026nbsp;(25)\u0026nbsp;justifies using ITA as a \u0026ldquo;tie-breaker\u0026quot; if MST and ITA do not assign the same L, M and D enrollment bin. MST, simultaneously, is integral to the recruitment, enrollment, analysis process guaranteeing that people can relate to the ITA value (recruitment, enrollment, error-checking, communication, reporting-out to lay-people).\u003c/li\u003e\n \u003cli\u003eThere is an implicit suggestion in the first bullet point that appearance/complexion is a primary correlator with differential bias related to pigmentation; this is a conscious choice. (See comments in Discussion regarding device performance analysis)\u003c/li\u003e\n \u003cli\u003eITA can be prone to errors\u0026nbsp;(16)\u0026nbsp;and is visually color-blind; MST is helpful for double checking during all stages of the study (recruitment, enrollment, analysis).\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eJustification of bin sizes, and targeted participant enrollment\u003c/h2\u003e\n\u003cp\u003eThe cut-off values (degrees ITA) and associated MST grades for the bins were inspired by a RWS of PO\u0026nbsp;(12)\u0026nbsp;in a public safety net hospital in San Francisco where a significant proportion of the study population had ITA\u0026lt;-50\u003csup\u003e◦\u003c/sup\u003e (12,26). This justified the additional condition for the dark bin (half of data should have ITA \u0026lt;-50\u003csup\u003e◦\u003c/sup\u003e) as a feasible recruitment condition in Western geographies. \u0026nbsp;The range of ITA found in this study\u0026nbsp;(12), is consistent with a recently published database of skin reflectance spectra and associated color values\u0026nbsp;(17)\u0026nbsp;(herein referred to as ISSA). It provides 2093 data points for the forehead, representing eight different ethnicities: Caucasian, Chinese, South Asian (Pakistani), African, Middle Eastern (Iraqi), Southeast Asian (Thai), Japanese and Middle Eastern (Arabian). Plotting all the 2093 data would appear as if the spread in the L*b* projection is much larger for medium skin than for dark and light skin, simply because more medium skin was measured. For the graph shown in\u0026nbsp;Figure 3A, randomly selected data from ISSA with an equal density of points is shown for all ITA from -75\u0026deg; to 70\u0026deg;. With the approximate range of human skin color indicated by the blue dashed lines (ITA= -80\u0026deg; and 66\u0026deg;), the L, M, and D bins create an approximately equal sized representation across the span of human skin color (\u003cem\u003ei.e.\u003c/em\u003e, with a goal of balanced distribution of data).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNote that the shape of the L*b* data point cloud is mostly encompassed by the yellow circle segments, consistent with the formula of the ITA and its ability to describe skin melanin pigmentation in a single convenient parameter (an angle).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe participant enrollment criteria shown is the result of carefully weighing many pros and cons of many metrics which was only possible after the JWG had become familiar with MST and ITA, and also many other methods of skin tone characterization\u0026nbsp;(25), each having their own specific properties. The next few sections are intended to share some of the aspects that were considered in the choices for MST and ITA.\u003c/p\u003e\n\u003ch2\u003eDiscrepancy between L* values for MST and L* values from colorimetry on skin.\u003c/h2\u003e\n\u003cp\u003ePlease note that the lightness L* of the MST grades for the three examples in Figure 4B (L* values for MST B, F and I are 92, 55 and 21, respectively, see\u0026nbsp;Figure 2A) are considerably different from the L* as measured with a colorimeter: 68, 52 and 31, respectively. This is consistent with the phenomenon described recently\u0026nbsp;(10): raters choose much lighter swatches when asked to match the forehead for light skin, and much darker swatches for dark skin. This phenomenon was consistently observed in four raters across two studies under different lighting conditions. Note that the mentioned L* values are not measured, and are for illustrative purposes only in describing the mentioned phenomenon\u0026nbsp;(10)\u0026nbsp;which can be described, in a simplified manner, as: light skin is perceived as lighter, and dark skin as darker.\u003c/p\u003e\n\u003cp\u003eWhile it can be confusing to have two ranges for skin color (a narrower L* range for objective skin color, and a much larger L* range for the MST scale), it illustrates why including a scale like MST is important to use contextually in its relationship to ITA. When presenting the distribution of enrolled participants in clinical verification studies, MST colors are much more effective than ITA, and more effective than other skin-tone scales, as discussed in the next sections.\u003c/p\u003e\n\u003ch2\u003ePlacing the MST into perspective\u003c/h2\u003e\n\u003cp\u003eRelated to the discrepancy described in the previous section, recent studies suggest that a specific skin tone scales has one of two intended goals. The first goal is to determine appearance, or complexion. The MST and presumably also other scales (Figure 4) such as PERLA and Von Luschan (VL) are used by holding them next (e.g. at one meter distance) to a person and considering the entire persona, including contextual features such as hair color, but \u003cem\u003enot\u003c/em\u003e on top of the skin. Other scales, like the Pantone\u0026reg;\u0026nbsp;SkinTone\u0026trade;\u0026nbsp;(PST) Guide\u0026nbsp;(27)\u0026nbsp;and L\u0026apos;Or\u0026eacute;al\u0026trade;\u0026nbsp;(13)\u0026nbsp;are presumably intended to make a match with skin color by placing them on top of a certain anatomical location like the forehead or arm, and find the best matching swatch by looking through (a factory made) perforation in the swatch which presumably helps to minimize the context. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe JWG recognized that human perception of skin color and objectively-measured skin color can be very different: light skin is perceived as much lighter, and dark skin is perceived as much darker\u0026nbsp;(10). The MST scale, but also other scales like PERLA and VL, reflect this phenomenon by offering a much larger light-dark (L*) range than would correspond with a scale that represents the objective range of skin colors such as the PST\u0026nbsp;(10,27). As a consequence, the instructions for use of this scale were recently refined\u0026nbsp;(10): punching holes in a printed MST and placing it on the skin to make a color match is no longer recommended. This is important because the instructions for use for most scales are often absent, difficult to find, or ambiguous\u0026nbsp;(27\u0026ndash;31)\u0026nbsp;while the results are highly dependent on how they are used. MST is preferred in the current application for several reasons: it has well defined colors that give a fair representation\u0026nbsp;(24), is relatively well researched compared to other scales, has a certain momentum (Google and Meta adopted it to be used in machine learning\u0026nbsp;(32\u0026ndash;34)) , and has a convenient ten degree granularity with instructions for use that are better defined than for other scales\u0026nbsp;(10). New scales are still being proposed, however, accompanied by research into factors involved in assessment of skin tone. Such factors include rater\u0026rsquo;s race and background color of the paper on which swatches are presented (e.g., grey instead of white)\u0026nbsp;(31).\u003c/p\u003e\n\u003cp\u003eMoreover, the MST colors focus on a variation in melanin only, as opposed to other scales which also express variations in dermal blood (redness). This is illustrated in\u0026nbsp;Figure 4. The MST swatches from light to dark in the L* vs b* chart shows a (banana) shape quite similar to the variations from melanin as defined by Del Bino et al\u0026nbsp;(15), albeit in a stretched form to account for the differences between perceived and objectively measured skin color\u0026nbsp;(10). Scales such as PERLA or VL also have one or more swatches that are reddish, presumably to express the impact of blood variations on skin color, however, they offer only one degree of freedom. The single-axis mixes melanin and blood variation and thus creates a non-monotonic relationship with objective melanin measurements\u0026nbsp;(35). We emphasize that scales with one or more reddish swatches (e.g. PERLA) are perhaps less preferred for the PO purpose (to describe variations in melanin) but may be preferred in other fields where describing variations in redness as well as melanin is desired.\u003c/p\u003e\n\u003cp\u003eScales such as PST and L\u0026apos;Or\u0026eacute;al\u0026trade; implicitly acknowledge the two main skin chromophores (melanin and blood, see Figure 1A) by introducing two axes. Such scales, however, are cumbersome to be used for our purpose and a poor imprecise surrogate for colorimetry. More importantly, they would not serve the communication purpose offered by MST because they match objectively measured skin colors, which differ from those perceived by humans in everyday life: \u003cem\u003ei.e.\u003c/em\u003e skin color is seen in context: face, hair, other exposed body parts. \u0026nbsp;Humans perceive skin color quite differently when it is in context (e.g., the face) versus when it is printed on a sheet of paper with a number of swatches\u0026nbsp;(10). And even under the same conditions, different people may see them differently\u0026nbsp;(36). Many mechanisms\u0026nbsp;(37,38)\u0026nbsp;may play a role in this phenomenon, including memory effects, facial features, and expectations\u0026nbsp;(39\u0026ndash;41).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn this context, it is important to realize that also objective measurements of skin can have many dimensions and that the approach of using ITA is a conscious simplification. ITA is precise, accurate and reproducible across the world, but many factors that can be objectively identified are implicitly discarded or minimized in the color measurement. This is illustrated in Figure 5 and leads us to the next section.\u003c/p\u003e\n\u003ch2\u003ePlacing the ITA into perspective\u003c/h2\u003e\n\u003cp\u003eSkin color homogeneity, hair, sebum (enhancing specular reflection), surface roughness are examples of factors that are likely to impact human perception of the skin. It is not feasible, and unnecessary, to measure each of these parameters. Skin color measured on the forehead by a colorimeter and reduced to an ITA, for the purpose of the standard, is a suitable surrogate for general appearance or complexion as determined by MST. Moreover, ITA was shown to correlate well with melanin concentration (5,15), arguably the dominant parameter in the MST scale (see previous section)\u003c/p\u003e\n\u003cp\u003eWhile Figure 5 expresses the complexity of \u003cem\u003eobjectively\u003c/em\u003e measurable parameters involved in skin color, there are still many more parameters that relate to how humans actually \u003cem\u003eperceive\u003c/em\u003e skin color when context is not minimized and the skin is observed in the context of a face, including hair color, eye color, etc. These parameters are more difficult to determine as they involve psycho-physical experiments (10,41) and include factors such as regional background (24) , experience, socio-economic status (of both participant and rater?). The JWG aimed to balance rigor with pragmatism/feasibility and make choices with respect to hypotheses, which leads us to the next section.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe rationale behind using MST, and ITA on the forehead was provided in one of the previous sections. For completeness and increased understanding of the JWG\u0026rsquo;s approach to the tasks, listed here are alternative hypotheses that were discussed but \u003cem\u003enot\u003c/em\u003e adopted for participant enrollment.\u003c/p\u003e\n\u003ch2\u003eAlternative hypotheses\u003c/h2\u003e\n\u003cp\u003eWe are aware that we chose to focus on pigmentation differential bias. There may be many other differential biases. Examples of alternative hypotheses that were considered but not selected for this draft are listed below.\u0026nbsp;\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eEthnicity is the primary correlator with observed differential bias in RWS.\u003c/li\u003e\n \u003cli\u003eRace is the primary correlator with the observed differential bias in RWS.\u003c/li\u003e\n \u003cli\u003eA combination of low perfusion and skin pigmentation (possibly correlated through ethnicity) is the primary correlator with differential bias in RWS(43)\u003c/li\u003e\n \u003cli\u003eEpidermal thickness is the primary correlator with observed differential bias in RWS.\u003c/li\u003e\n \u003cli\u003eNail thickness is the primary correlator with\u0026nbsp;observed differential bias in RWS.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eWhile some of the parameters mentioned above are not used for enrolling study participants, they will be measured or recorded, if possible, and used in the analysis for characterizing differential bias to elucidate the root cause expediently, to produce devices that perform equally well for all skin colors, ethnicities and races.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOf note, the JWG uses slightly different working hypotheses for the two identified tasks . \u0026nbsp;For the participant enrollment criteria the JWG used MST and melanin pigmentation (ITA) \u003cem\u003eat the forehead\u003c/em\u003e while for the second task, creating the pigmentation differential bias metric, melanin pigmentation \u003cem\u003eat the probe site\u003c/em\u003e is used. This reflects the early stage of research into the root cause of the RWS observations, as well as the democratic and open character of the discussions that took place within the JWG. Moreover, the ITA on the forehead is quite strongly correlated with the ITA measured on the distal phalange (12), considered to be a good representation of the pigmentation on the sensor site for finger probes.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThe draft revision of the ISO 80601-2-61 standard requires verifying pulse oximeter performance in a controlled desaturation study with enrolled participants falling into the three skin color bins, L, M and D, shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003eB in similar numbers. The use of a relatively small number of bins provides a good balance between participant recruitment/enrollment feasibility and scientific rigor. Experience has shown that it can be difficult to recruit people with darker skin. There is an additional requirement that at least half of participants in the \u0026lsquo;Dark\u0026rsquo; bin have an ITA \u0026lt; -50\u0026deg; (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e) to ensure this proportion of the skin tone category is not skewed to the lighter margin. The JWG carefully and exhaustively weighed factors involved in participant enrollment and considered the bins as defined in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003e to be such that inclusion in verification studies will be fair and representative of the global population.\u003c/p\u003e\u003cp\u003eWith the current state of knowledge, the JWG considers complexion, appearance, \u0026lsquo;gestalt\u0026rsquo; to be the preferred metric to ensure enrolling participants across a wide distribution of skin tones. The MST directly correlates with this metric and describes it well in relatable values/colors and has high levels of inter-rater reliability (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). However, the subjective nature of annotation by humans also allows for various sources of bias: regional cultural contexts (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e) may influence how people perceive skin tone. As a consequence, subjective methods to assess pigmentation are currently considered to have too much uncertainty to be adopted as the decisive metric for participant enrollment (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e) in verification studies for PO. Simultaneously, the JWG acknowledges that effective communication to lay-audiences is critically important, and that ITA is not suitable for that purpose. Moreover, measuring ITA is itself prone to errors. Therefore, the updated ISO draft recommends using both ITA and MST characterized on the forehead as metrics for participant enrollment, where ITA will be used as the tie-breaker for the cases when MST and ITA binning are not consistent with one another. ITA should be measured with devices that have traceable calibration.\u003c/p\u003e\u003cp\u003eThis enrollment methodology is not limited to only clinical laboratory studies with healthy adult participants. The measurement protocol should be practical for critically ill patients at the bedside for the entire age spectrum from pre-term newborns to adults (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). In this context, the JWG is currently exploring how smaller, lighter, affordable, and user-friendlier colorimeters can be used. Typically, these less costly colorimeters are sufficiently accurate and precise for the purposes stated herein. Traceable calibration is often lacking for these less costly devices; it may be possible to create a verification/calibration method to ensure the measurements are reliable.\u003c/p\u003e\u003cp\u003eFinally, at this stage of knowledge, it is premature to conclude what the root cause(s) for the RWS observations are due to, or even the perfect metric to ensure diverse participant enrollment. Scientific evidence will continue to drive the rationales behind future ISO device standards in an ongoing effort to improve this facet of pulse oximeter performance.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAcknowledgements\u003c/h2\u003e\n\u003cp\u003eWe thank Sandy Weininger, Mike Lipnick for their critical reading of the manuscript and valuable suggestions. We also thank all JWG members whose diligent work has helped shape the current draft standard as well as OpenOximetry (openoximetry.org) for gathering and sharing knowledge and study data.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDISCLOSURES:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eProf. Monk is the developer of the Monk SkinTone Scale. Drs. Verkruysse, Milkes, Mannheimer and Kopotic work for commercial enterprises: Philips, Medtronic, Apple and BD/EdwardsLifesciences, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCONFLICTS OF INTEREST:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNA\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCLINICAL TRIAL NUMBER AND REGISTRY URL.\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eABBREVIATED TITLE:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAUTHOR CONTRIBUTIONS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWim Verkruysse initiated and helped write the manusript.\u003c/p\u003e\n\u003cp\u003eAll other authors helped write the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eISO 80601-2-61 Medical Electrical Equipment \u0026mdash;Part 2-61: Particular requirements for basic safety and essential performance of pulse oximeter equipment; expected to be formally published in early 2026. 2025. \u003c/li\u003e\n\u003cli\u003eSjoding MW, Dickson RP, Iwashyna TJ, Gay SE, Valley TS. Racial Bias in Pulse Oximetry Measurement. N Engl J Med. 2020 Dec 17;383(25):2477\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eMoran-Thomas A. How a Popular Medical Device Encodes Racial Bias. Boston Review [Internet]. 2020 Aug 5 [cited 2025 Apr 2]; Available from: http://bostonreview.net/science-nature-race/amy-moran-thomas-how-popular-medical-device-encodes-racial-bias.\u003c/li\u003e\n\u003cli\u003eJubran A, Tobin MJ. Reliability of Pulse Oximetry in Titrating Supplemental Oxygen Therapy in Ventilator-Dependent Patients. CHEST. 1990 Jun 1;97(6):1420\u0026ndash;5. \u003c/li\u003e\n\u003cli\u003eAlaluf S, Atkins D, Barrett K, Blount M, Carter N, Heath A. The Impact of Epidermal Melanin on Objective Measurements of Human Skin Colour. Pigment Cell Research. 2002 Apr;15(2):119\u0026ndash;26. \u003c/li\u003e\n\u003cli\u003eVerkruysse W, Lucassen GW, van Gemert MJC. Simulation of color of port wine stain skin and its dependence on skin variables. Lasers in Surgery and Medicine. 1999;25(2):131\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eMonk EP. The Monk Skin Tone Scale. 2019 [cited 2024 Apr 19]; Available from: https://osf.io/preprints/socarxiv/pdf4c/\u003c/li\u003e\n\u003cli\u003eChardon A, Cretois I, Hourseau C. Skin colour typology and suntanning pathways. International Journal of Cosmetic Science. 1991;13(4):191\u0026ndash;208. \u003c/li\u003e\n\u003cli\u003eVasudevan S, Vogt WC, Weininger S, Pfefer TJ. Melanometry for objective evaluation of skin pigmentation in pulse oximetry studies. Commun Med. 2024 Jul 11;4(1):138. \u003c/li\u003e\n\u003cli\u003eVerkruysse W, Monk EPJ, Jaffe MB. Objective and Perceived Skin Color: Consequences for the Use of Skintone Scales. Anesthesia \u0026amp; Analgesia. 2025 Jan;140(1):e2. \u003c/li\u003e\n\u003cli\u003eGroen S, Verkruysse W, Jaffe M, Muehlsteff J. Skin Color at Pulse Oximeter Measurement Sites Varies Considerably Between and Within Individuals. Poster presented at: AACN NTI National Teaching Institute 2024; 2024 May 29; Denver, CO. \u003c/li\u003e\n\u003cli\u003eLipnick MS, Chen D, Law T, Moore Jr K, Lester JC, Monk EP, et al. Comparison of methods for characterizing skin pigment diversity in research cohorts. medRxiv. 2025;2025\u0026ndash;02. \u003c/li\u003e\n\u003cli\u003eGordon RA, Branigan AR, Khan MA, Nunez JG. Best Practices for Measuring Skin Color in Surveys. Survey Practice [Internet]. 2024 May 23 [cited 2025 Jan 19];17. Available from: https://www.surveypractice.org/article/94628-best-practices-for-measuring-skin-color-in-surveys\u003c/li\u003e\n\u003cli\u003eStandard, C. I. E. (2007). Colorimetry-part 4: CIE 1976 L* a* b* colour space. International Standard, 2019-06. \u003c/li\u003e\n\u003cli\u003eDel Bino S, Bernerd F. Variations in skin colour and the biological consequences of ultraviolet radiation exposure. Br J Dermatol. 2013;169(s3):33\u0026ndash;40. \u003c/li\u003e\n\u003cli\u003eVerkruysse W, Jaffe MB. Letter Regarding the Potential Improper Use of the Individual Typology Angle in the Context of Disparate Bias in Pulse Oximetry. Anesthesia \u0026amp; Analgesia. 2024 May;138(5):e31. \u003c/li\u003e\n\u003cli\u003eLu Y, Xiao K, Pointer M, He R, Zhou S, Nasseraldin A, et al. The International Skin Spectra Archive (ISSA): a multicultural human skin phenotype and colour spectra collection. Sci Data [Internet]. 2025 Mar 23 [cited 2025 Mar 29];12(1). Available from: https://www.nature.com/articles/s41597-025-04857-5\u003c/li\u003e\n\u003cli\u003ePexels [Internet]. [cited 2025 Apr 3]. Photo by Polina Tankilevitch on Pexels. Available from: https://www.pexels.com/photo/woman-in-white-top-wearing-blue-knit-cap-4723530/\u003c/li\u003e\n\u003cli\u003ePexels [Internet]. [cited 2025 Apr 3]. Photo by Centre for Ageing Better on Pexels. Available from: https://www.pexels.com/photo/woman-leaning-head-on-hand-17153160/\u003c/li\u003e\n\u003cli\u003ePexels [Internet]. [cited 2025 Apr 3]. Photo by Ant\u0026oacute;nio Ribeiro on Pexels. Available from: https://www.pexels.com/photo/portrait-of-woman-with-crossed-arms-10619795/\u003c/li\u003e\n\u003cli\u003eFong N, Lipnick M, Bickler P, Feiner J, Law T. OpenOximetry Repository [Internet]. PhysioNet; [cited 2025 May 6]. Available from: https://physionet.org/content/openox-repo/1.0.1/\u003c/li\u003e\n\u003cli\u003eLeeb G, Auchus I, Law T, Bickler P, Feiner J, Hashi S, et al. The performance of 11 fingertip pulse oximeters during hypoxemia in healthy human participants with varied, quantified skin pigment. eBioMedicine [Internet]. 2024 Apr 1 [cited 2025 May 6];102. Available from: https://www.thelancet.com/journals/ebiom/article/PIIS2352-3964(24)00086-0/fulltext\u003c/li\u003e\n\u003cli\u003eRej\u0026oacute;n Pi\u0026ntilde;a RA, Ma C. Classification Algorithm for Skin Color (CASCo): A new tool to measure skin color in social science research. Social Science Quarterly. 2023;104(2):168\u0026ndash;79. \u003c/li\u003e\n\u003cli\u003eSchumann C, Olanubi GO, Wright A, Monk Jr. E, Heldreth C, Ricco S. Consensus and Subjectivity of Skin Tone Annotation for ML Fairness [Internet]. arXiv; 2023 [cited 2023 Dec 19]. Available from: http://arxiv.org/abs/2305.09073\u003c/li\u003e\n\u003cli\u003eVerkruysse W, Jaffe MB, Lipnick M, Zemouri C. Challenges of Subjective Skin Color Scales: The Case for the Use of Objective Pigmentation Measurement Methods in Regulatory Pulse Oximetry Studies. Anesthesia \u0026amp; Analgesia. 2024;10\u0026ndash;1213. \u003c/li\u003e\n\u003cli\u003eThe Open Oximetry March 13, 2024 \u0026ndash; Skin Color Quantification Subgroup Meeting 3 accessed via https://youtu.be/6ELI__1XcbI. \u003c/li\u003e\n\u003cli\u003ePantone. Pantone. [cited 2024 Feb 6]. PANTONE SkinTone Guide. Available from: https://www.pantone.com/skintone\u003c/li\u003e\n\u003cli\u003eCohen PR, DiMarco MA, Geller RL, Darrisaw LA. Colorimetric Scale for Skin of Color: A Practical Classification Scale for the Clinical Assessment, Dermatology Management, and Forensic Evaluation of Individuals With Skin of Color. Cureus [Internet]. 2023 Nov 1 [cited 2024 Mar 4]; Available from: https://www.cureus.com/articles/201038-colorimetric-scale-for-skin-of-color-a-practical-classification-scale-for-the-clinical-assessment-dermatology-management-and-forensic-evaluation-of-individuals-with-skin-of-color\u003c/li\u003e\n\u003cli\u003eTelles EE. Multiple measures of ethnoracial classification in Latin America. Ethnic and Racial Studies. 2017 Oct 21;40(13):2340\u0026ndash;6. \u003c/li\u003e\n\u003cli\u003eFelix von Luschan. In: Wikipedia [Internet]. 2024 [cited 2025 Jan 19]. Available from: https://en.wikipedia.org/w/index.php?title=Felix_von_Luschan\u0026amp;oldid=1241316357\u003c/li\u003e\n\u003cli\u003eCook C, Howard J, Rabbitt L, Shuggi I, Sirotin Y, Tipton J, et al. Colorimetric skin tone scale for improved accuracy of human skin tone annotations. ACM J Responsib Comput [Internet]. 2025 May 3 [cited 2025 May 15]; Available from: https://dl.acm.org/doi/10.1145/3730409\u003c/li\u003e\n\u003cli\u003eGoogle researchers, Ellis Monk. Skin Tone Research @ Google AI | Start using the Monk Skin Tone Scale [Internet]. [cited 2023 Dec 7]. Available from: https://skintone.google\u003c/li\u003e\n\u003cli\u003eGustafson L, Rolland C, Ravi N, Duval Q, Adcock A, Fu CY, et al. FACET: Fairness in Computer Vision Evaluation Benchmark. In 2023 [cited 2025 May 15]. p. 20370\u0026ndash;82. Available from: https://openaccess.thecvf.com/content/ICCV2023/html/Gustafson_FACET_Fairness_in_Computer_Vision_Evaluation_Benchmark_ICCV_2023_paper.html\u003c/li\u003e\n\u003cli\u003ePorgali B, Albiero V, Ryda J, Ferrer CC, Hazirbas C. The casual conversations v2 dataset. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition [Internet]. 2023 [cited 2025 May 15]. p. 10\u0026ndash;7. Available from: https://openaccess.thecvf.com/content/CVPR2023W/TCV/html/Porgali_The_Casual_Conversations_v2_Dataset_CVPRW_2023_paper.html\u003c/li\u003e\n\u003cli\u003eSwiatoniowski AK, Quillen EE, Shriver MD, Jablonski NG. Technical Note: Comparing von Luschan skin color tiles and modern spectrophotometry for measuring human skin pigmentation. American Journal of Physical Anthropology. 2013;151(2):325\u0026ndash;30. \u003c/li\u003e\n\u003cli\u003eBosten JM. Do You See What I See? Diversity in Human Color Perception. Annu Rev Vis Sci. 2022 Sep 15;8(1):101\u0026ndash;33. \u003c/li\u003e\n\u003cli\u003eBaker LJ, Levin DT. The Face-Race Lightness Illusion Is Not Driven by Low-level Stimulus Properties: An Empirical Reply to Firestone and Scholl (2014). Psychon Bull Rev. 2016 Dec;23(6):1989\u0026ndash;95. \u003c/li\u003e\n\u003cli\u003eFirestone C, Scholl BJ. Cognition does not affect perception: Evaluating the evidence for \u0026ldquo;top-down\u0026rdquo; effects. Behav Brain Sci. 2016;39:e229. \u003c/li\u003e\n\u003cli\u003eLyngs U, Cohen E, Hattori WT, Newson M, Levin DT. Hearing in color: How expectations distort perception of skin tone. Journal of Experimental Psychology: Human Perception and Performance. 2016 Dec;42(12):2068\u0026ndash;76. \u003c/li\u003e\n\u003cli\u003eHansen T, Olkkonen M, Walter S, Gegenfurtner KR. Memory modulates color appearance. Nat Neurosci. 2006 Nov;9(11):1367\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eLevin DT, Banaji MR. Distortions in the perceived lightness of faces: The role of race categories. J Exp Psychol. 2006;135(4):501\u0026ndash;12. \u003c/li\u003e\n\u003cli\u003eDlugos JF, Taylor JL. Materials Characterization: UV/Vis/NIR Spectroscopy [Internet]. Perkin Elmer; 2019 [cited 2024 Apr 15]. Available from: https://resources.perkinelmer.com/lab-solutions/resources/docs/App_Visible-Reflectance-Spectroscopy-Human-Skin.pdf\u003c/li\u003e\n\u003cli\u003eGudelunas MK, Lipnick M, Hendrickson C, Vanderburg S, Okunlola B, Auchus I, et al. Low Perfusion and Missed Diagnosis of Hypoxemia by Pulse Oximetry in Darkly Pigmented Skin: A Prospective Study. Anesthesia \u0026amp; Analgesia. 2022 Mar 18;10.1213/ANE.0000000000006755. \u003c/li\u003e\n\u003cli\u003eSharma M, Pickhardt AJ, Madison, Tong L, Evans MS. Objective Assessment and Quantification of Skin Color and Melanin in Neonates and Infants: A State-Of-The-Art Review. Pediatric Dermatology [Internet]. [cited 2025 Apr 2];n/a(n/a). Available from: https://onlinelibrary.wiley.com/doi/abs/10.1111/pde.15902\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"648\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 648px;\"\u003e\n \u003cp\u003eTable 1 Key advantages and disadvantages of MST and ITA for assessing skin color of participants in pulse oximeter verification studies\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 257px;\"\u003e\n \u003cp\u003eMST\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 270px;\"\u003e\n \u003cp\u003eITA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003eAdvantages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 257px;\"\u003e\n \u003cp\u003eCommunication to interested parties on enrollment characteristics as the colors relate directly to how humans perceive skin color/complexion.\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 270px;\"\u003e\n \u003cp\u003eIn line with good laboratory practice of using a calibrated device to record an objective, precise, accurate, repeatable measurement with negligible rater bias\u0026nbsp;(25). Focuses on melanin, which is the variable of interest of the working hypothesis.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003eDisadvantages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 257px;\"\u003e\n \u003cp\u003eSubjective nature: metric collected from different human raters can add variability in observed MST grades. Improving precision is cumbersome. Rater bias, willful or not, is difficult to eliminate. Standardized printing is cumbersome.\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 270px;\"\u003e\n \u003cp\u003eCumbersome to be used in communication because objectively measured skin colors, and ITA even more so, poorly relate to how people experience skin color and the range of skin color.\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-clinical-monitoring-and-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Journal of Clinical Monitoring and Computing](https://www.springer.com/journal/10877)","snPcode":"10877","submissionUrl":"https://submission.nature.com/new-submission/10877/3","title":"Journal of Clinical Monitoring and Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7112852/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7112852/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003ePurpose.\u003c/h2\u003e\u003cp\u003eThis communication describes a a skin tone characterization methodology for the international standard ISO 80601-2-61:2026 (Ed 3) on pulse oximeter basic safety and essential performance. The methodology\u0026rsquo;s goal is to create the appropriate proportional distributions of skin tones for participants in pulse oximeters performance verification studies. We elucidate the techniques and associated metrics used for characterizing skin tone. This standard is currently in its final draft and voting stage before becoming formally adopted.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eVarious methods to quantify pigmentation were evaluated for their suitability for the purpose. Evaluation criteria included precision, accuracy, standardizability and efficacy for communication. The evaluations and selections are the result of ISO Joint Working Group activity over a period of about two years.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eThe individual typology angle (ITA), was identified as a preferred objective measurement method of pigmentation. It is colorimetry based, is consequently highly standardized and has high precision and accuracy. The Monk Skin Tone scale was selected as a subjective method to complement the objective method as it is much more relatable than the ITA.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003eA standardized method for assessing skin tone and standardized expectation for balanced participant enrollment across the world\u0026rsquo;s range of skin tones was defined. A colorimetry based metric is used for precision and accuracy in combination with a color scale for optimal communication and recruitment.\u003c/p\u003e","manuscriptTitle":"Towards achieving even distributions of participant skin tones when verifying pulse oximeter performance","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-08 18:37:55","doi":"10.21203/rs.3.rs-7112852/v1","editorialEvents":[{"type":"communityComments","content":1},{"type":"decision","content":"Revision requested","date":"2025-08-09T09:25:30+00:00","index":"","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-07-17T19:13:05+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-15T04:11:58+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-15T04:10:47+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Clinical Monitoring and Computing","date":"2025-07-13T11:00:44+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-clinical-monitoring-and-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Journal of Clinical Monitoring and Computing](https://www.springer.com/journal/10877)","snPcode":"10877","submissionUrl":"https://submission.nature.com/new-submission/10877/3","title":"Journal of Clinical Monitoring and Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"6ff474a7-767b-46cf-8918-e10f35aacafa","owner":[],"postedDate":"October 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-05-21T14:08:45+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-08 18:37:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7112852","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7112852","identity":"rs-7112852","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}