Melanoma in skin of color: Part II. Racial disparities, role of UV, and interventions for earlier detection

Weakly supervised deep learning image analysis can differentiate melanoma from naevi on haematoxylin and eosin-stained histopathology slides

BACKGROUND

The broad histomorphological spectrum of melanocytic pathologies requires large data sets to develop accurate and generalisable deep learning (DL)-based diagnostic pathology classifiers. Weakly supervised DL promotes utilisation of larger training data sets compared to fully supervised (patch annotation) approaches.

OBJECTIVES

To evaluate weakly supervised DL image classifiers for discriminating melanomas from naevi on haematoxylin and eosin (H&E)-stained pathology slides.

METHODS

A representative H&E slide for 260 naevi and 260 melanomas from mucocutaneous sites at one tertiary institution was digitized. Clinicopathological features were recorded for each case including thickness and histological subtype. Whole-slide or whole-tissue section labels were applied. The ground truth was established by consensus diagnosis from two pathologists. Multiple-instance learning models, Trans-MIL, CLAM and DTFD-MIL were evaluated at 10×, 20× and 40× magnifications using stratified fivefold Monte Carlo cross-validation, with 80/10/10 splits for training/validation/test groups, to predict melanoma from naevus. Heatmaps were generated to understand model performance.

RESULTS

Naevi cases were younger (median age: 51 years; melanoma median age: 71.5 years), with more balanced sex distribution (males: 48.8%, melanoma male subgroup: 64.2%). The most frequent histological subtypes of naevi and melanomas were dysplastic compound (n = 99, 38.1%) and superficial spreading (n = 124, 47.7%), respectively. Average AUC (±1 SD) for Trans-MIL, CLAM and DTFD-MIL across test groups were 0.9952 ± 0.006, 0.9925 ± 0.0052 and 0.9708 ± 0.0328, at 20× magnification, respectively. Performance of the models varied according to the magnification used. Heatmaps from the two best performing models, Trans-MIL and CLAM, generally indicated attention on appropriate tissue regions for interpretation.

CONCLUSIONS

Weakly supervised DL on pathological slides of common mucocutaneous melanocytic tumours provides highly accurate diagnostic value for discrimination of melanomas and naevi. External validation and further assessment on less frequently occurring histologic subtypes and borderline cases using this method is required.

Malignant Melanoma: An Overview, New Perspectives, and Vitamin D Signaling

Melanoma, originating through malignant transformation of melanin-producing melanocytes, is a formidable malignancy, characterized by local invasiveness, recurrence, early metastasis, resistance to therapy, and a high mortality rate. This review discusses etiologic and risk factors for melanoma, diagnostic and prognostic tools, including recent advances in molecular biology, omics, and bioinformatics, and provides an overview of its therapy. Since the incidence of melanoma is rising and mortality remains unacceptably high, we discuss its inherent properties, including melanogenesis, that make this disease resilient to treatment and propose to use AI to solve the above complex and multidimensional problems. We provide an overview on vitamin D and its anticancerogenic properties, and report recent advances in this field that can provide solutions for the prevention and/or therapy of melanoma. Experimental papers and clinicopathological studies on the role of vitamin D status and signaling pathways initiated by its active metabolites in melanoma prognosis and therapy are reviewed. We conclude that vitamin D signaling, defined by specific nuclear receptors and selective activation by specific vitamin D hydroxyderivatives, can provide a benefit for new or existing therapeutic approaches. We propose to target vitamin D signaling with the use of computational biology and AI tools to provide a solution to the melanoma problem.

Multi-scale feature fusion and class weight loss for skin lesion classification

Skin cancer is one of the common types of cancer. It spreads quickly and is not easy to detect in the early stages, posing a major threat to human health. In recent years, deep learning methods have attracted widespread attention for skin cancer detection in dermoscopic images. However, training a practical classifier becomes highly challenging due to inter-class similarity and intra-class variation in skin lesion images. To address these problems, we propose a multi-scale fusion structure that combines shallow and deep features for more accurate classification. Simultaneously, we implement three approaches to the problem of class imbalance: class weighting, label smoothing, and resampling. In addition, the HAM10000_RE dataset strips out hair features to demonstrate the role of hair features in the classification process. We demonstrate that the region of interest is the most critical classification feature for the HAM10000_SE dataset, which segments lesion regions. We evaluated the effectiveness of our model using the HAM10000 and ISIC2019 dataset. The results showed that this method performed well in dermoscopic classification tasks, with ACC and AUC of 94.0% and 99.3%, on the HAM10000 dataset and ACC of 89.8% for the ISIC2019 dataset. The overall performance of our model is excellent in comparison to state-of-the-art models.

Sunscreen use and affordability attitudes based on ethnicity, socioeconomic status, and Fitzpatrick skin type

IMPORTANCE

One in five Americans will develop skin cancer during their lifetime. While use of sunscreen can help prevent the development cutaneous cancer, regular use remains low nationwide.

OBJECTIVE

To assess and better understand health care consumer preferences for sun protection products and perceived product accessibility and availability based on socioeconomic factors, race, and ethnicity.

DESIGN

This quantitative survey study was conducted March through June of 2023.

SETTING

Participants were recruited from two university family medicine clinical sites in the Buffalo, New York area, one located in a low and one located in a middle-to-upper socioeconomic neighborhood.

PARTICIPANTS

Eligible participants were 18 years or older, fluent in English, and residents of the Buffalo, New York area. Surveys and consent forms were distributed by scripted verbal invitation, inviting all clinic patients who met eligibility criteria to participate. Participants were asked to self-report their racial/ethnic group as well as other demographic information including age, gender identification, household income, and household size. Information regarding sun exposure behaviors, and affordability/access was obtained using a combination of multiple choice and yes/no questions. A total of 405 participants were recruited. After excluding 235 incomplete responses, 170 surveys were available for analysis.

INTERVENTIONS

None.

MAIN OUTCOMES AND MEASURES

Our study aim was to expose health care consumer preferences as well as barriers to access based on socioeconomic factors, race, and ethnicity.

RESULTS

Using a 25-question anonymous survey, 405 participants from two university family medicine clinical sites representing low- and middle-to-high-income neighborhoods, participated in the survey. 170 participants completed the survey questions and were included for analysis. Of those, 61.8% identified as female, 37.6% as male, and 0.6% as other. 51.2% of participants identified as lower income, 38.2% as middle-income, and 10.6% as upper income. The results of the survey revealed disparities in sunscreen use and affordability perceptions across demographic groups. Compared with Hispanics, Caucasians exhibited higher rates of sunscreen use (85 Caucasians, 7 Hispanics; p = 0.0073), prioritized SPF (95 Caucasians, 10 Hispanics; p = 0.0178), and were more likely to perceive sunscreen as unaffordable (6 Caucasians, 4 Hispanics; p = 0.0269). Analysis by Fitzpatrick Skin Type demonstrated differences in sunscreen utilization, with Types I-III using more compared to Types IV-VI (70 Types I-III, 51 Types IV-VI; p = 0.0173); additionally, Type I-III individuals were significantly more likely to cite cost as barrier to sunscreen purchase (40 Type I-III, 65 Types IV-VI; p < 0.0001). Moreover, lower-income individuals were significantly more likely to perceive sunscreen as unaffordable (12 lower-income, 1 middle & upper income; p = 0.0025) and cited cost as a barrier to purchase (46 lower-income, 59 middle & upper income; p = 0.0146) compared to middle-to-upper income counterparts. Though statistical significance was not established, respondents from middle & upper income groups reported higher sunscreen usage rates compared with their lower-income peers.

CONCLUSIONS AND RELEVANCE

These findings highlight the importance of socioeconomic factors and ethnicity on accessibility to sunscreen and the impact of disparities in utilization among different ethnic and socioeconomic groups.

Dermatology in black skin

The vast majority of publications in dermatology refer to lightly pigmented skin, with few addressing the peculiarities of black skin. In addition there is no consensus on what it means to be black in different regions of the world. The lack of knowledge on the subject makes it difficult to recognize and manage dermatoses in this type of skin. This article aims to review the literature on intrinsic characteristics, as well as epidemiological and clinical aspects of the cutaneous manifestations of different dermatoses in black skin. It was found that there are sometimes striking differences, in the structural, biological, and functional aspects when comparing lightly pigmented and black skin. There are also physiological changes that need to be recognized to avoid unnecessary interventions. Some dermatoses have a higher incidence in black skin, such as acne, eczema, dyschromia and dermatophytosis. On the other hand, several dermatoses are more specific to black skin, such as pseudofolliculitis barbae, keloid, dermatosis papulosa nigra, ulcers caused by sickle-cell anemia, dactylolysis spontanea, confluent and reticulated papillomatosis of Gougerot and Carteaud, and some diseases of the hair and scalp (including fragile and brittle hair, traction alopecia, folliculitis keloidalis nuchae, folliculitis dissecans and central centrifugal cicatricial alopecia). A spectrum of peculiar aspects of specific dermatoses, including sarcoidosis, lichen planus (with emphasis on the pigmentosus variant), psoriasis, lupus erythematosus, vitiligo, syphilis, pityriasis versicolor, and neoplasms are highlighted. In the latter, characteristics of basal cell carcinoma, squamous cell carcinoma, and melanoma are compared, in addition to highlighting unusual aspects of primary cutaneous T-cell lymphoma, endemic Kaposi sarcoma, and dermatofibrosarcoma protuberans.

Cutaneous Oncology: Strategies for Melanoma Prevention, Diagnosis, and Therapy

Skin cancer comprises one-third of all diagnosed cancer cases and remains a major health concern. Genetic and environmental parameters serve as the two main risk factors associated with the development of skin cancer, with ultraviolet radiation being the most common environmental risk factor. Studies have also found fair complexion, arsenic toxicity, indoor tanning, and family history among the prevailing causes of skin cancer. Prevention and early diagnosis play a crucial role in reducing the frequency and ensuring effective management of skin cancer. Recent studies have focused on exploring minimally invasive or non-invasive diagnostic technologies along with artificial intelligence to facilitate rapid and accurate diagnosis. The treatment of skin cancer ranges from traditional surgical excision to various advanced methods such as phototherapy, radiotherapy, immunotherapy, targeted therapy, and combination therapy. Recent studies have focused on immunotherapy, with the introduction of new checkpoint inhibitors and personalized immunotherapy enhancing treatment efficacy. Advancements in multi-omics, nanotechnology, and artificial intelligence have further deepened the understanding of the mechanisms underlying tumoral growth and their interaction with therapeutic effects, which has paved the way for precision oncology. This review aims to highlight the recent advancements in the understanding and management of skin cancer, and provide an overview of existing and emerging diagnostic, prognostic, and therapeutic modalities, while highlighting areas that require further research to bridge the existing knowledge gaps.

Melanoma Detection in Alaska Native, American Indian, Asian, Black, Hispanic, and Pacific Islander Patients in a Large Skin Cancer Screening Initiative

This cohort study investigates patterns in melanoma detection in racial and ethnic minority populations as part of a large, primary care–based screening initiative.

Skin Cancer in Skin of Color

Skin cancer is often associated with greater morbidity and mortality in skin of color patients because most medical literature and research on skin cancer to date has been predominantly focused on lighter skin types. It is crucial that dermatologic providers be able to recognize different presentations of skin cancer in skin of color patients to optimize the early detection of these tumors and ensure equitable outcomes. This article details the epidemiology, risk factors, clinical features, and disparities in the treatment of melanoma, squamous cell carcinoma, basal cell carcinoma, and mycosis fungoides subtype of cutaneous T-cell lymphoma in skin of color patients.

Advances in Early Detection of Melanoma and the Future of At-Home Testing

The past decade has seen numerous advancements in approaches to melanoma detection, each with the common goal to stem the growing incidence of melanoma and its mortality rate. These advancements, while well documented to increase early melanoma detection, have also garnered considerable criticism of their efficacy for improving survival rates. In this review, we discuss the current state of such early detection approaches that do not require direct dermatologist intervention. Our findings suggest that a number of at-home and non-specialist methods exist with high accuracy for detecting melanoma, albeit with a few notable concerns worth further investigation. Additionally, research continues to find new approaches using artificial intelligence which have promise for the future.