The Eumelanin Human Skin Colour Scale: A Proof-of-Concept Study

From data to diagnosis: skin cancer image datasets for artificial intelligence

Artificial intelligence (AI) solutions for skin cancer diagnosis continue to gain momentum, edging closer towards broad clinical use. These AI models, particularly deep-learning architectures, require large digital image datasets for development. This review provides an overview of the datasets used to develop AI algorithms and highlights the importance of dataset transparency for the evaluation of algorithm generalizability across varying populations and settings. Current challenges for curation of clinically valuable datasets are detailed, which include dataset shifts arising from demographic variations and differences in data collection methodologies, along with inconsistencies in labelling. These shifts can lead to differential algorithm performance, compromise of clinical utility, and the propagation of discriminatory biases when developed algorithms are implemented in mismatched populations. Limited representation of rare skin cancers and minoritized groups in existing datasets are highlighted, which can further skew algorithm performance. Strategies to address these challenges are presented, which include improving transparency, representation and interoperability. Federated learning and generative methods, which may improve dataset size and diversity without compromising privacy, are also examined. Lastly, we discuss model-level techniques that may address biases entrained through the use of datasets derived from routine clinical care. As the role of AI in skin cancer diagnosis becomes more prominent, ensuring the robustness of underlying datasets is increasingly important.

The relevance and accuracy of an AI algorithm-based descriptor on 23 facial attributes in a diverse female US population

BACKGROUND

The response of AI in situations that mimic real life scenarios is poorly explored in populations of high diversity.

OBJECTIVE

To assess the accuracy and validate the relevance of an automated, algorithm-based analysis geared toward facial attributes devoted to the adornment routines of women.

METHODS

In a cross-sectional study, two diversified groups presenting similar distributions such as age, ancestry, skin phototype, and geographical location was created from the selfie images of 1041 female in a US population. 521 images were analyzed as part of a new training dataset aimed to improve the original algorithm and 520 were aimed to validate the performance of the AI. From a total 23 facial attributes (16 continuous and 7 categorical), all images were analyzed by 24 make-up experts and by the automated descriptor tool.

RESULTS

For all facial attributes, the new and the original automated tool both surpassed the grading of the experts on a diverse population of women. For the 16 continuous attributes, the gradings obtained by the new system strongly correlated with the assessment made by make-up experts (r ≥ 0.80; p < 0.0001) and supported by a low error rate. For the seven categorical attributes, the overall accuracy of the AI-facial descriptor was improved via enrichment of the training dataset. However, some weaker performance in spotting specific facial attributes were noted.

CONCLUSION

In conclusion, the AI-automatic facial descriptor tool was deemed accurate for analysis of facial attributes for diverse women although some skin complexion, eye color, and hair features required some further finetuning.

Global consensus process to establish a core dataset for hidradenitis suppurativa registries

BACKGROUND

Several registries for hidradenitis suppurativa (HS) already exist in Europe and the USA. There is currently no global consensus on a core dataset (CDS) for these registries. Creating a global HS registry is challenging, owing to logistical and regulatory constraints, which could limit opportunities for global collaboration as a result of differences in the dataset collected. The solution is to encourage all HS registries to collect the same CDS of information, allowing registries to collaborate.

OBJECTIVES

To establish a core set of items to be collected by all HS registries globally. The core set will cover demographic details, comorbidities, clinical examination findings, patient-reported outcome measures and treatments.

METHODS

Beginning in September 2022, 20 participants - including both clinicians with expertise in HS and patient advocates - from eight countries across three continents participated in a Delphi process consisting of four rounds of voting, with all participants completing each round. A list of potential items for inclusion in the core set was generated from the relevant published literature, including systematic reviews of comorbidities in HS, clinical and examination findings, and epidemiology. For disease severity and progression items, the Hidradenitis SuppuraTiva Core outcome set International Collaboration (HiSTORIC) core set and other relevant instruments were considered for inclusion. This resulted in 47 initial items. Participants were invited to suggest additional items to include during the first round. Anonymous feedback was provided to inform each subsequent round of voting to encourage consensus.

RESULTS

The eDelphi process established a CDS of 48 items recommended for inclusion in all HS registries globally.

CONCLUSIONS

The routine adoption of this CDS in current and future HS registries should allow registries in different parts of the world to collaborate, enabling research requiring large numbers of participants.

Integrating skin color assessments into clinical practice and research: A review of current approaches

Skin color classification can have importance in skin health, pigmentary disorders, and oncologic condition assessments. It is also critical for evaluating disease course and response to a variety of therapeutic interventions and aids in accurate classification of participants in clinical research studies. A panel of dermatologists conducted a literature review to assess the strengths and limitations of existing classification scales, as well as to compare their preferences and utilities. We identified 17 skin classification systems utilized in dermatologic settings. These systems include a range of parameters such as UV light reactivity, race, ethnicity, and degree of pigmentation. The Fitzpatrick skin type classification is most widely used and validated. However it has numerous limitations including its conflation with race, ethnicity, and skin color. There is a lack of validation data available for the remaining scales. There are significant deficiencies in current skin classification instruments. Consensus-based initiatives to drive the development of validated and reliable tools are critically needed.

Impact of Skin Pigmentation on Cerebral Regional Saturation of Oxygen Using Near-Infrared Spectroscopy: A Systematic Review

OBJECTIVES

Near-infrared spectroscopy (NIRS) is used in critical care settings to measure regional cerebral tissue oxygenation (rSo2). However, the accuracy of such measurements has been questioned in darker-skinned individuals due to the confounding effects of light absorption by melanin. In this systematic review, we aim to synthesize the available evidence on the effect of skin pigmentation on rSo2 readings.

DATA SOURCES

We systematically searched MEDLINE, Cochrane Database of Systematic Reviews, Embase, and Google Scholar from inception to July 1, 2023.

STUDY SELECTION

In compliance with our PROSPERO registration (CRD42022347548), we selected articles comparing rSo2 measurements in adults either between racial groups or at different levels of skin pigmentation. Two independent reviewers conducted full-text reviews of all potentially relevant articles.

DATA EXTRACTION

We extracted data on self-reported race or level of skin pigmentation and mean rSo2 values.

DATA SYNTHESIS

Of the 11,495 unique records screened, two studies (n = 7,549) met our inclusion criteria for systematic review. Sun et al (2015) yielded significantly lower rSo2 values for African Americans compared with Caucasians, whereas Stannard et al (2021) found little difference between self-reported racial groups. This discrepancy is likely because Stannard et al (2021) used a NIRS platform which specifically purports to control for the effects of melanin. Several other studies that did not meet our inclusion criteria corroborated the notion that skin pigmentation results in lower rSo2 readings.

CONCLUSIONS

Skin pigmentation likely results in attenuated rSo2 readings. However, the magnitude of this effect may depend on the specific NIRS platform used.

Skin Protection by Carotenoid Pigments

Sunlight, despite its benefits, can pose a threat to the skin, which is a natural protective barrier. Phototoxicity caused by overexposure, especially to ultraviolet radiation (UVR), results in burns, accelerates photoaging, and causes skin cancer formation. Natural substances of plant origin, i.e., polyphenols, flavonoids, and photosynthetic pigments, can protect the skin against the effects of radiation, acting not only as photoprotectors like natural filters but as antioxidant and anti-inflammatory remedies, alleviating the effects of photodamage to the skin. Plant-based formulations are gaining popularity as an attractive alternative to synthetic filters. Over the past 20 years, a large number of studies have been published to assess the photoprotective effects of natural plant products, primarily through their antioxidant, antimutagenic, and anti-immunosuppressive activities. This review selects the most important data on skin photodamage and photoprotective efficacy of selected plant carotenoid representatives from in vivo studies on animal models and humans, as well as in vitro experiments performed on fibroblast and keratinocyte cell lines. Recent research on carotenoids associated with lipid nanoparticles, nanoemulsions, liposomes, and micelles is reviewed. The focus was on collecting those nanomaterials that serve to improve the bioavailability and stability of carotenoids as natural antioxidants with photoprotective activity.

Optimisation of skin phototype classification

Understanding individuals' skin pigmentation and photosensitivity is important in judging risk of skin cancer and response to certain treatment modalities. However, individuals with darkly pigmented skin are poorly represented in the widely used Fitzpatrick skin phototype (FST) system. Moreover, the FST system is prone to misuse, as it relies on subjective patient and clinician assessment of skin type, and does not clearly differentiate pigmentation from photosensitivity. By evaluating the key literature surrounding the FST system, its criticisms and proposed alternatives, this review serves to understand how skin phototype classification can be optimised.

Public views are needed for skin colour scales

Our public engagement exercise showed that the Fitzpatrick Scale does not adequately represent how members of the public consider their own skin. Descriptions offered by respondents regarding their skin tone were richly varied; some also indicated that structured and embedded inequalities affected the language they used. The patient voice is key to developing comprehensive, inclusive and widely understood tools.

Using Technology to Detect Erythema Across Skin Tones

OBJECTIVE

To examine the effectiveness of the ColorMeter DSM III (ColorMeter; Cortex Technology) at grouping individuals by skin tone and measuring erythema/skin discoloration after erythema induction across skin tones.

METHODS

This pre/post experimental study induced erythema on a convenience sample of 61 healthy adults. Skin tone at baseline was measured using the ColorMeter, Munsell Soil Color Chart 5YR (Munsell), and Pantone SkinTone Guide (Pantone) and compared with the Eumelanin Human Skin Colour Scale (Eumelanin Scale) groupings. Erythema and melanin values on the arm immediately and after recovery time were compared with baseline values. Melanin was measured at five body regions on the face and arm.

RESULTS

Participants were predominantly women (64% [n = 39] women, 36% [n = 22] men) and young (mean, 28.8 ± 14.3 years); 5% (n = 3) were Hispanic, 26% (n = 16) Asian, 29% (n = 18) Black, 38% (n = 23) White, and 7% (n = 4) identified with more than one race. ColorMeter lightness (L*) and melanin measures were strongly correlated with both Munsell and Pantone values. Munsell skin tone groups were not aligned with Eumelanin Scale groupings. Most participants were in the Eumelanin intermediate-low group, and this changed depending on which body location melanin value was used. The change in erythema from baseline did not differ significantly across skin tone groups at the ulnar head, but on the forearm at the delayed time point, significant differences existed between light and both medium and dark skin tone groups (P = .001; 95% CI, 0.04-0.37).

CONCLUSIONS

The ColorMeter provides an effective objective measure of skin tone and erythema/discoloration across various skin tones and may improve on current standards for detection. The proposed Eumelanin Scale-Modified provides additional sensitivity for persons with medium skin tones.

Current Perspectives on Pressure Injuries in Persons with Dark Skin Tones from the National Pressure Injury Advisory Panel

BACKGROUND

Pressure injury (PI) development is multifactorial. In patients with dark skin tones, identifying impending PIs by visual skin assessment can be especially challenging. The need for improved skin assessment techniques, especially for persons with dark skin tones, continues to increase. Similarly, greater awareness of the need for inclusivity with regard to representation of diverse skin colors/tones in education materials is apparent.

OBJECTIVE

To provide current perspectives from the literature surrounding skin assessment and PI development in patients with dark skin tones.

METHODS

The following elements will be discussed through the lens of skin tone: (1) historical perspectives of PI staging from the National Pressure Injury Advisory Panel, (2) epidemiology of PI, (3) anatomy and physiology of the skin, (3) skin tone assessment and measurement, (4) augmented visual assessment modalities, (5) PI prevention, (6) PI healing, (7) social determinants of health, and (8) gaps in clinician education.

CONCLUSIONS

This article highlights the gap in our clinical knowledge regarding PIs in patients with dark skin tones. Racial disparities with regard to PI development and healing are especially clear among patients with dark skin tones. Skin tone color assessment must be standardized and quantifiable in clinical education, practice, and research. This work is urgently needed, and support from private and governmental agencies is essential.

Deconstructing Eurocentrism in skin pigmentation research via the incorporation of diverse populations and theoretical perspectives

The evolution of skin pigmentation has been shaped by numerous biological and cultural shifts throughout human history. Vitamin D is considered a driver of depigmentation evolution in humans, given the deleterious health effects associated with vitamin D deficiency, which is often shaped by cultural factors. New advancements in genomics and epigenomics have opened the door to a deeper exploration of skin pigmentation evolution in both contemporary and ancient populations. Data from ancient Europeans has offered great context to the spread of depigmentation alleles via the evaluation of migration events and cultural shifts that occurred during the Neolithic. However, novel insights can further be gained via the inclusion of diverse ancient and contemporary populations. Here we present on how potential biases and limitations in skin pigmentation research can be overcome with the integration of interdisciplinary data that includes both cultural and biological elements, which have shaped the evolutionary history of skin pigmentation in humans.

Human Skin Pigmentation: From a Biological Feature to a Social Determinant

Skin pigmentation is the most variable human characteristic that can be observed and has been used throughout history to classify humans into distinct groups. Many factors influence skin colour, but the melanin pigment is considered the most important because its type and quantity can determine variations in pigmentation shades. The evolution of skin pigmentation started around 1.6-2 million years ago. As a result of migratory phenomena to places with less ultraviolet radiation (UVR) and other seasonal regimes, the selection of depigmented skin and different tanning capabilities occurred over time. Thus, genetic adaptation to new environmental conditions gradually led to changes in skin pigmentation. Despite the biological importance of pigmentation, variation in skin colour has led to social and health inequalities. Since Linnaeus, skin colour classifications have been used to describe different human groups, encouraging the misuse of a biological characteristic. This review examines the characterisation of pigmentation and its evolution through history and society. The unequal perception of pigmentation diversity has led to an incomplete state of dermatological training and issues in medical approach in dermatology. The consciousness of all these aspects increases the need to address and overcome dermatologic and social health disparities related to skin pigmentation.

Managing Atopic Dermatitis in Patients With Skin of Color

Managing atopic dermatitis (AD) in patients with skin of color presents unique challenges for the clinician. There is increasing evidence that AD has higher prevalence, persistence, and severity among skin of color populations. This is likely to be partly related to differences in living conditions and exposure to irritants and allergens, among other factors. Assessment of AD severity in patients with darker skin can be challenging, in particular the assessment of erythema, leading to the potential for underscoring AD severity. Variations in disease have also been described, with the potential for a greater risk of inflammation-induced nodularity and hyper- or hypopigmentation. Management challenges include variable adherence to treatment, potential disparities in access to health care, and differences in the metabolism of cyclosporine. Optimal management of AD in patients with skin of color requires a tailored approach. Here, we review approaches to diagnosing AD, evaluating extent and severity with subjective and objective measures, considering treatment options for patients with skin of color, and highlighting areas for improvement in AD care for skin of color populations.

Skin of colour: essentials for the non-dermatologist

Doctors-in-training often receive an inadequate dermatology education. Furthermore, studies have highlighted the under-representation of skin of colour (SOC) in dermatological teaching, learning resources and research. Our image-based questionnaire, distributed to all internal medicine trainees in southwest England, highlighted knowledge gaps regarding SOC among training physicians. It is intrinsically more challenging for clinicians to confidently formulate dermatological diagnoses in SOC. In this review, we provide guidance for physicians to help make the diagnostic process more straightforward. First, we outline how skin colour is determined and classified. We discuss how inflammation presents in SOC, with the typical 'erythema' that physicians often associate with inflammation being a less prominent feature in darker skin tones. We then summarise nine important conditions that we believe physicians working in all specialties should be able to identify in patients with SOC, covering both conditions encountered on the medical take and conditions disproportionately affecting individuals with SOC. The population of the UK is rapidly diversifying; thus, as physicians, we have a professional duty to educate ourselves on dermatological conditions in SOC to provide the best quality of care for all our patients, regardless of their skin type.

UK ethnic minority healthcare workers' perspectives on COVID-19 vaccine hesitancy in the UK ethnic minority community: A qualitative study

Background

The experiences of UK ethnic minority (UKEM) healthcare workers are crucial to ameliorating the disproportionate COVID-19 infection rate and outcomes in the UKEM community. We conducted a qualitative study on UKEM healthcare workers' perspectives on COVID-19 vaccine hesitancy (CVH) in the UKEM community.

Methods

Participants were 15 UKEM healthcare workers (11 females; age range: 26-58 [43.3 ± 9.4] years). Data were collected using individual and joint interviews, and a focus group, and analyzed using thematic analysis.

Results

We generated three themes: heterogeneity (two subthemes), mistrust (six subthemes), and mitigating (six subthemes). Therein, participants distinguished CVH in the UKEM community in educational attainment and ethnicity. They pointed to the role of mistrust in CVH in the UKEM community. They opined that the mistrust underlying CVH in the UKEM community is rooted in history and religion, conspiracy theories, the speedy development and novelty of the vaccines, post-vaccination complications/side effects, false positive test results, and social media and social support/influence. Participants recommended that interventions targeted at mitigating CVH in the UKEM community need to, in a non-judgmental way, tackle dis/misinformation and provide education, and incorporate UKEM healthcare worker endorsement. They also suggested such interventions be community-oriented, enhance the convenience of vaccination centers and the possibility of vaccine choice, and appreciate that overcoming CVH and accepting vaccination is a gradual process involving personal assessment of risks and benefits.

Conclusion

CVH in the UKEM community is a multifaceted phenomenon requiring multicomponent interventions.