ABSTRACT
Extracellular vesicles (EVs) are important mediators of communication between cells. Here, we reveal a new mode of intercellular communication by melanosomes, large EVs secreted by melanocytes for melanin transport. Unlike small EVs, which are disintegrated within the receiver cell, melanosomes stay intact within them, gain a unique protein signature, and can then be further transferred to another cell as "second-hand" EVs. We show that melanoma-secreted melanosomes passaged through epidermal keratinocytes or dermal fibroblasts can be further engulfed by resident macrophages. This process leads to macrophage polarization into pro-tumor or pro-immune cell infiltration phenotypes. Melanosomes that are transferred through fibroblasts can carry AKT1, which induces VEGF secretion from macrophages in an mTOR-dependent manner, promoting angiogenesis and metastasis in vivo. In melanoma patients, macrophages that are co-localized with AKT1 are correlated with disease aggressiveness, and immunotherapy non-responders are enriched in macrophages containing melanosome markers. Our findings suggest that interactions mediated by second-hand extracellular vesicles contribute to the formation of the metastatic niche, and that blocking the melanosome cues of macrophage diversification could be helpful in halting melanoma progression.
ABSTRACT
Natural killer (NK) cells are innate lymphoid cells, and their presence within human tumors correlates with better prognosis. However, the mechanisms by which NK cells control tumors in vivo are unclear. Here, we used reflectance confocal microscopy (RCM) imaging in humans and in mice to visualize tumor architecture in vivo. We demonstrated that signaling via the NK cell receptor NKp46 (human) and Ncr1 (mouse) induced interferon-γ (IFN-γ) secretion from intratumoral NK cells. NKp46- and Ncr1-mediated IFN-γ production led to the increased expression of the extracellular matrix protein fibronectin 1 (FN1) in the tumors, which altered primary tumor architecture and resulted in decreased metastases formation. Injection of IFN-γ into tumor-bearing mice or transgenic overexpression of Ncr1 in NK cells in mice resulted in decreased metastasis formation. Thus, we have defined a mechanism of NK cell-mediated control of metastases in vivo that may help develop NK cell-dependent cancer therapies.
Subject(s)
Antigens, Ly/metabolism , Fibronectins/metabolism , Interferon-gamma/metabolism , Killer Cells, Natural/metabolism , Natural Cytotoxicity Triggering Receptor 1/metabolism , Neoplasms/metabolism , Animals , Blotting, Western , Female , Flow Cytometry , Fluorescent Antibody Technique , Gene Expression Regulation, Neoplastic/genetics , Humans , Male , Mice , Microscopy, Confocal , Neoplasm Metastasis/genetics , Neoplasms/pathology , Real-Time Polymerase Chain Reaction , Signal Transduction/geneticsABSTRACT
BACKGROUND: Lentigo maligna (LM) can mimic benign, flat, pigmented lesions and can be challenging to diagnose. OBJECTIVE: To describe a new dermatoscopic feature termed "perifollicular linear projections (PLP)" as a diagnostic criterion for LM on the face. METHODS: Retrospective study on reflectance confocal microscopy and dermatoscopy images of flat facial pigmented lesions originating from 2 databases. PLP were defined as short, linear, pigmented projections emanating from hair follicles. Dermatoscopy readers were blinded to the final histopathologic diagnosis. RESULTS: From 83 consecutive LMs, 21/83 (25.3%) displayed "bulging of hair follicles" on reflectance confocal microscopy and 18 of these 21 (85.7%), displayed PLP on dermatoscopy. From a database of 2873 consecutively imaged and biopsied lesions, 252 flat-pigmented facial lesions were included. PLP was seen in 47/76 melanomas (61.8%), compared with 7/176 lesions (3.9%) with other diagnosis (P < .001). The sensitivity was 61.8% (95% CI, 49.9%-72.7%), specificity 96.0% (95% CI, 92.9%-98.4%). PLP was independently associated with LM diagnosis on multivariate analysis (OR 26.1 [95% CI, 9.6%-71.0]). LIMITATIONS: Retrospective study. CONCLUSION: PLP is a newly described dermatoscopic criterion that may add specificity and sensitivity to the early diagnosis of LM located on the face. We postulate that PLP constitutes an intermediary step in the LM progression model.
Subject(s)
Hutchinson's Melanotic Freckle , Melanoma , Skin Neoplasms , Humans , Hutchinson's Melanotic Freckle/diagnostic imaging , Hutchinson's Melanotic Freckle/pathology , Skin Neoplasms/diagnostic imaging , Skin Neoplasms/pathology , Retrospective Studies , Diagnosis, Differential , Melanoma/pathology , Microscopy, Confocal/methods , Dermoscopy/methodsABSTRACT
BACKGROUND: A common terminology for diagnosis is critically important for clinical communication, education, research and artificial intelligence. Prevailing lexicons are limited in fully representing skin neoplasms. OBJECTIVES: To achieve expert consensus on diagnostic terms for skin neoplasms and their hierarchical mapping. METHODS: Diagnostic terms were extracted from textbooks, publications and extant diagnostic codes. Terms were hierarchically mapped to super-categories (e.g. 'benign') and cellular/tissue-differentiation categories (e.g. 'melanocytic'), and appended with pertinent-modifiers and synonyms. These terms were evaluated using a modified-Delphi consensus approach. Experts from the International-Skin-Imaging-Collaboration (ISIC) were surveyed on agreement with terms and their hierarchical mapping; they could suggest modifying, deleting or adding terms. Consensus threshold was >75% for the initial rounds and >50% for the final round. RESULTS: Eighteen experts completed all Delphi rounds. Of 379 terms, 356 (94%) reached consensus in round one. Eleven of 226 (5%) benign-category terms, 6/140 (4%) malignant-category terms and 6/13 (46%) indeterminate-category terms did not reach initial agreement. Following three rounds, final consensus consisted of 362 terms mapped to 3 super-categories and 41 cellular/tissue-differentiation categories. CONCLUSIONS: We have created, agreed upon, and made public a taxonomy for skin neoplasms and their hierarchical mapping. Further study will be needed to evaluate the utility and completeness of the lexicon.
ABSTRACT
BACKGROUND: Treatment management for congenital melanocytic nevi (CMN) on the face (FCMN) is highly variable and requires a thorough assessment of multiple factors. To date, a systematic review of FCMN treatment is lacking. The purpose of the present study was to elucidate the frequency, variety, and outcomes of treatment modalities for FCMN with different levels of complexity. METHODS: A comprehensive review of Pubmed, Embase, and Google Scholar databases from 1950 to 2022 was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Articles reporting on FCMN treatment approaches, outcomes, and associated complications were screened and data were extracted according to inclusion criteria. Data were tabulated for thematic analysis of FCMN treatment types, anatomic locations, outcomes, and complications. RESULTS: Of the 561 studies retrieved, 34 met inclusion criteria including 19 surgical treatments, 14 nonsurgical treatments, and one combined surgical and nonsurgical treatment study, totaling 356 patients. The majority of treated FCMN were small-to-medium-sized (56%). Facial CMN treated conservatively were mostly located on the cheek (27%) and/or perinasal region (21%), whereas FCMN treated with surgery were primarily located in the periorbital region (44%) and/or the cheek (17%). Across all treatment cohorts, 22% of patients experienced at least one complication, with 12% of complications experienced by patients treated by surgery. CONCLUSIONS: There is a greater need for standardized FCMN nomenclature that encompasses nevi pattern, dimensions, anatomical coverage, and quantitative measurements of treatment outcome. Future studies should focus on identifying anatomic locations of FCMN that are more prone to complications and determine which treatment approach optimizes outcomes.
Subject(s)
Nevus, Pigmented , Skin Neoplasms , Humans , Nevus, Pigmented/surgery , Treatment Outcome , Cheek , Databases, Factual , Skin Neoplasms/surgery , Skin Neoplasms/complicationsABSTRACT
INTRODUCTION: Early detection may lead to reduced morbidity and mortality from melanoma. This study aims to establish guidelines for selecting patients suitable for digital monitoring of skin lesions. METHODS: A literature review was conducted, followed by consensus among experts appointed by the Israeli Dermatology Association. RESULTS: Two effective methods for early melanoma diagnosis were identified: Total-body photography (TBP) and digital dermoscopy. TBP involves capturing clinical images of the entire skin area for long-term monitoring (6-12 months). Digital dermoscopy focuses on close-up images of distinct lesions for short-term monitoring (3-4 months). Various risk factors for melanoma were identified, including genetic and familial factors, as well as demographic and phenotypic characteristics. Based on these risk factors and feasibility of clinical follow-up, a comprehensive list of indications for TBP was developed, categorized into three groups based on the expected level of benefit. Digital dermoscopy surveillance is recommended for patients with flat or slightly raised skin lesions showing dermoscopic features that do not definitively indicate melanoma. DISCUSSION: TBP significantly improves early melanoma detection, enhancing sensitivity and specificity while reducing unnecessary biopsies. However, due to its high cost and limited coverage by the Israeli public health care system, prioritizing patients who would benefit most from TBP is crucial. The compiled list of indications aligns with international recommendations and provides further details within the article.
Subject(s)
Dermatology , Melanoma , Humans , Israel , Melanoma/diagnosis , Biopsy , ConsensusABSTRACT
BACKGROUND: Facial skin is characterized by high density of follicles. Facial neoplasms may present overlapping clinical and dermoscopic findings. Our goal was to evaluate and compare, via reflectance confocal microscopy (RCM), follicular involvement in facial neoplasms. METHODS: We retrospectively searched our image database, between January 2008 and December 2020, for all facial lesions with (1) a standardized set of clinical, dermoscopic, and RCM images, and (2) a biopsy-proven diagnosis of lentigo maligna/lentigo maligna melanoma (LM/LMM, n = 39), basal cell carcinoma (BCC, n = 51), squamous cell carcinoma in situ (SCCIS, n = 5), actinic keratosis (AK, n = 11), and lichen-planus-like keratosis (LPLK, n = 18). Two readers jointly evaluated the RCM images for a set of predefined features of follicular involvement. RESULTS: Diffuse obliteration of follicles was frequent in BCC (88%), while follicular infiltration by refractile dendritic cells and/or by bright round nucleated cells was common in melanoma (90% and 44%, respectively). Extension of atypical keratinocytes down follicles was more prominent among SCCIS than AK (80% vs. 45%, p = 0.01). In most LPLK (89%), there was follicular sparing. CONCLUSIONS: Evaluation of RCM criteria centering on the follicles can be useful in the differential diagnosis between common facial neoplasms.
Subject(s)
Facial Neoplasms , Hutchinson's Melanotic Freckle , Keratosis, Actinic , Melanoma , Skin Neoplasms , Humans , Hutchinson's Melanotic Freckle/diagnosis , Hutchinson's Melanotic Freckle/pathology , Skin Neoplasms/pathology , Retrospective Studies , Melanoma/diagnosis , Melanoma/pathology , Keratosis, Actinic/diagnosis , Facial Neoplasms/pathology , Diagnosis, Differential , Dermoscopy/methods , Microscopy, Confocal/methodsABSTRACT
BACKGROUND: Folliculitis decalvans (FD) is a type of primary neutrophilic cicatricial alopecia often leading to irreversible hair loss. Data on its epidemiology, clinical features, outcomes, and prognostic factors are limited. OBJECTIVE: To evaluate a cohort of patients with FD and identify characteristics of severe disease and prognostic factors which impede remission. PATIENTS AND METHODS: This retrospective cohort study included 192 patients diagnosed with FD and followed for at least six months at a tertiary center between 2010 and 2020. RESULTS: There was a diagnostic delay averaging 22.2 (± 29.7) months. Comorbid follicular occlusion disorders were common. Bacterial cultures were positive in 45.6% of the cases, with Staphylococcus (S.) aureus being the most common pathogen. Severe disease was associated with comorbid hidradenitis suppurativa and a positive bacterial culture, particularly S. aureus. 50.7% of patients experienced complete remission: 32% within the first six months of treatment and 18.7% later during follow-up. Relapses were frequent. Negative prognostic factors for achieving remission included younger age and a positive bacterial culture. CONCLUSIONS: There is a need for the education of dermatologists to reduce the diagnostic delay. Screening FD patients for comorbid hidradenitis suppurativa and obtaining bacterial cultures is important for treatment planning.
Subject(s)
Folliculitis , Hidradenitis Suppurativa , Humans , Cohort Studies , Delayed Diagnosis , Hidradenitis Suppurativa/diagnosis , Hidradenitis Suppurativa/epidemiology , Hidradenitis Suppurativa/complications , Prognosis , Retrospective Studies , Folliculitis/diagnosis , Folliculitis/epidemiology , Folliculitis/drug therapy , Staphylococcus aureus , Alopecia/diagnosis , Alopecia/epidemiology , Alopecia/drug therapyABSTRACT
INTRODUCTION: Melanocytic nevi present microscopic patterns, which differ in their associated melanoma risk, and can be non-invasively recognized under Reflectance Confocal Microscopy (RCM). AIMS: To train a Generative Adversarial Network (GAN) deep-learning model to produce synthetic images that recapitulate RCM patterns of nevi, enabling reliable classification by human readers and by a Convolutional Neural Network (CNN) computer model. METHODS: A dataset of RCM images of nevi, presenting a uniform pattern, were chosen and classified into one of three patterns - Meshwork, Ring or Clod. Images were used for training a GAN model, which in turn, produced synthetic images recapitulating RCM patterns of nevi. A random sample of synthetic images was classified by two independent human readers and by a CNN model. Human and computer-model classifications were compared. RESULTS: The training set for the GAN model included 1496 RCM images, including 977 images (65.3%) with Meshwork pattern, 261 (17.4%) with Ring and 258 (17.2%) with Clod pattern. The GAN model produced 6000 synthetic RCM-like images. Of these, 302 images were randomly chosen and classified by human readers, including 83 (27.5%) classified as Meshwork, 131 (43.4%) as Ring, and 88 (29.1%) as Clod pattern. Human inter-observer concordance in pattern classification was 91.7%, and human-to-CNN concordance was 87.7%. CONCLUSIONS: We demonstrate feasibility of producing synthetic images, which recapitulate RCM patterns of nevi and can be reproducibly recognized by human readers and by deep-learning models. Synthetic image datasets may allow teaching RCM patterns to novices, training of computer models, and data sharing between research centers.
Subject(s)
Deep Learning , Image Interpretation, Computer-Assisted , Melanoma , Nevus, Pigmented , Skin Neoplasms , Humans , Microscopy, Confocal/methods , Nevus, Pigmented/diagnostic imaging , Nevus, Pigmented/ultrastructure , Skin Neoplasms/diagnostic imagingABSTRACT
BACKGROUND: Congenital nail matrix nevi (NMN) are difficult to diagnose because they feature clinical characteristics suggestive of adult subungual melanoma. Nail matrix biopsy is difficult to perform, especially in children. OBJECTIVE: To describe the initial clinical and dermatoscopic features of NMN appearing at birth (congenital) or after birth but before the age of 5 years (congenital-type). METHODS: We conducted a prospective, international, and consecutive data collection in 102 hospitals or private medical offices across 30 countries from 2009 to 2019. RESULTS: There were 69 congenital and 161 congenital-type NMNs. Congenital and congenital-type NMN predominantly displayed an irregular pattern of longitudinal microlines (n = 146, 64%), reminiscent of subungual melanoma in adults. The distal fibrillar ("brush-like") pattern, present in 63 patients (27.8%), was more frequently encountered in congenital NMN than in congenital-type NMN (P = .012). Moreover, congenital NMN more frequently displayed a periungual pigmentation (P = .029) and Hutchinson's sign (P = .027) than did congenital-type NMN. LIMITATIONS: Lack of systematic biopsy-proven diagnosis and heterogeneity of clinical and dermatoscopic photographs. CONCLUSION: Congenital and congenital-type NMN showed worrisome clinical and dermatoscopic features similar to those observed in adulthood subungual melanoma. The distal fibrillar ("brush-like") pattern is a suggestive feature of congenital and congenital-type NMN.
Subject(s)
Melanoma , Nail Diseases , Nevus , Skin Neoplasms , Adult , Child , Child, Preschool , Dermoscopy , Diagnosis, Differential , Humans , Infant, Newborn , Melanoma/diagnostic imaging , Melanoma/pathology , Nail Diseases/diagnostic imaging , Nail Diseases/pathology , Nevus/diagnosis , Prospective Studies , Skin Neoplasms/diagnostic imaging , Skin Neoplasms/pathologyABSTRACT
BACKGROUND: The adherence to a narrowband ultraviolet B (NB-UVB) treatment plan is derived, in large part, from the patient's skin tolerance to the phototherapy dose. At present, the initial and first-month incremental phototherapy doses are determined prior to treatment initiation based on the patient's Fitzpatrick skin phototyping. OBJECTIVES: To identify variables that predict adherence to NB-UVB first-month treatment dosage plan. METHODS: Charts of 1000 consecutive patients receiving NB-UVB at a hospital-based phototherapy unit were retrospectively analyzed. We included patients receiving NB-UVB for atopic dermatitis, psoriasis, vitiligo, and mycosis fungoides. The first-month NB-UVB treatment plan was determined based on the patient's Fitzpatrick phototype. Adherence to treatment was defined as receiving at least 80% of the planned first-month cumulative dose. We compared adherent vs. non-adherent patient groups for age, sex, Fitzpatrick phototype, presence of freckles, nevus count category, and type of dermatological disease. RESULTS: The study included 817 eligible patients, mean age 40 (2-95) years; 54% men; 32% had Fitzpatrick phototype I-II. Distribution by diagnosis was atopic dermatitis (29%), psoriasis (27%), vitiligo (23%), and mycosis fungoides (21%). Adherence to NB-UVB treatment plan was observed in 71% of patients. Adherence decreased with age, with 7% decrease per year (P = 0.03) and was higher among mycosis fungoides patients (77.3%) compared to all other diagnoses (69.8%; P = 0.02). CONCLUSIONS: Adherence to NB-UVB treatment may be related to age and diagnosis. Fitzpatrick phototype-based first-month treatment plans should be modified accordingly.
Subject(s)
Dermatitis, Atopic , Mycosis Fungoides , Psoriasis , Skin Neoplasms , Ultraviolet Therapy , Vitiligo , Male , Humans , Adult , Female , Ultraviolet Therapy/adverse effects , Vitiligo/diagnosis , Vitiligo/radiotherapy , Dermatitis, Atopic/etiology , Dermatitis, Atopic/therapy , Retrospective Studies , Psoriasis/radiotherapy , Mycosis Fungoides/radiotherapy , Mycosis Fungoides/etiology , Skin Neoplasms/radiotherapy , Skin Neoplasms/etiology , Treatment OutcomeABSTRACT
BACKGROUND: Nevus phenotype is a marker of melanoma risk. In vivo prediction of microscopic pattern is needed to more precisely classify nevi. OBJECTIVE: To identify dermoscopic and clinical predictors of microscopic patterns of typical nevi. METHODS: We used reflectance confocal microscopy (RCM) to classify microscopic patterns of nevi. We prospectively accrued adults presenting for periodic skin screening and imaged, with dermoscopy and RCM, 3 randomly selected nevi from the upper and lower back and lower extremity. RCM patterns were classified into ring, clod, meshwork, and composite types. Logistic regression was used to identify best predictors of RCM pattern. RESULTS: The study included 310 nevi from 112 participants (mean age 44 years; 51 women). Dermoscopic reticular pattern correlated most frequently (59.9%) with RCM ring pattern, dermoscopic globular with RCM composite (56.6%) and RCM clod (35.9%), dermoscopic complex with RCM composite (76.3%), and dermoscopic homogenous with RCM clod (50.8%). Integrating dermoscopic pattern with contour, diameter, color, and anatomic location of nevi improved prediction of microscopic patterns beyond dermoscopy alone. The dermoscopic clinical regression model correctly classified lesions to RCM ring versus RCM clod in 90% and to RCM ring versus RCM composite patterns in 81%. LIMITATIONS: The study was restricted to adults, back and lower extremities, and typical nevi. CONCLUSIONS: Integrating dermoscopic patterns with clinical attributes may improve prediction of microscopic patterns of nevi.
Subject(s)
Nevus , Adult , Cross-Sectional Studies , Dermoscopy , Female , Humans , Leg , Male , Microscopy, Confocal , Skin Neoplasms/diagnostic imagingABSTRACT
Reflectance confocal microscopy (RCM) is a noninvasive imaging tool used for in vivo visualization of the skin. It has been extensively studied for use in the evaluation of equivocal cutaneous neoplasms to decrease the number of biopsy procedures in patients with benign lesions. Furthermore, its applications are broadening to include presurgical cancer margin mapping, tumor recurrence surveillance, monitoring of ablative and noninvasive therapies, and stratification of inflammatory disorders. With the approval of category I Current Procedural Terminology reimbursement codes for RCM image acquisition and interpretation, use of this technology has been increasingly adopted by dermatologists. The first article in this 2-part continuing medical education series highlights basic terminology, principles, clinical applications, limitations, and practical considerations in the clinical use of RCM technology.
Subject(s)
Dermoscopy/methods , Image Interpretation, Computer-Assisted , Skin/diagnostic imaging , Dermatologic Surgical Procedures , Dermoscopy/instrumentation , Humans , Margins of Excision , Microscopy, Confocal/instrumentation , Microscopy, Confocal/methods , Preoperative Care/methods , Skin/pathologyABSTRACT
Reflectance confocal microscopy (RCM) is a high-resolution, noninvasive tool that is currently approved by the US Food and Drug Administration for obtaining and interpreting images of the skin and cutaneous neoplasms with the goal of decreasing unnecessary biopsy procedures in patients with benign lesions. The second article in this continuing medical education series focuses on identifying key criteria for the diagnosis of common skin cancers-melanoma, basal cell carcinoma, and squamous cell carcinoma. We contrast these findings with RCM features of common benign lesions-melanocytic nevi, solar lentigo, seborrheic keratosis, lichen planus-like keratosis, and sebaceous hyperplasia. We also correlate the dermoscopic and histopathologic findings with the RCM features.
Subject(s)
Dermoscopy/methods , Keratosis, Actinic/diagnosis , Keratosis, Seborrheic/diagnosis , Lentigo/diagnosis , Lichen Planus/diagnosis , Skin Neoplasms/diagnosis , Algorithms , Dermoscopy/instrumentation , Diagnosis, Differential , Humans , Keratosis, Actinic/pathology , Keratosis, Seborrheic/pathology , Lentigo/pathology , Lichen Planus/pathology , Microscopy, Confocal/instrumentation , Microscopy, Confocal/methods , Skin/diagnostic imaging , Skin/pathology , Skin Neoplasms/pathologyABSTRACT
BACKGROUND: There is lack of uniformity in the reflectance confocal microscopy (RCM) terminology for melanocytic lesions. OBJECTIVE: To review published RCM terms for melanocytic lesions and identify redundant, synonymous terms. METHODS: A systematic review of original research articles adhering to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines was conducted until August 15, 2018. Two investigators gathered all published RCM terms used to describe melanoma and melanocytic nevi. Synonymous terms were grouped based on similarity in definition and in histopathologic correlation. RESULTS: Out of 156 full-text screened articles, 59 studies met the inclusion criteria. We identified 209 terms; 191 (91.4%) corresponding to high-magnification/cellular-level terms and 18 (8.6%) corresponding to low-magnification/architectural patterns terms. The overall average use frequency of RCM terms was 3.1 times (range, 1-31). By grouping of individual RCM terms based on likely synonymous definitions and by eliminating terms lacking clear definition, the total number of RCM terms could be potentially reduced from 209 to 40 terms (80.8% reduction). LIMITATIONS: Non-English and non-peer-reviewed articles were excluded. CONCLUSIONS: This systematic review of published RCM terms identified significant terminology redundancy. It provides the basis for subsequent terminology consensus on melanocytic neoplasms.
Subject(s)
Melanoma/classification , Melanoma/pathology , Microscopy, Confocal , Skin Neoplasms/classification , Skin Neoplasms/pathology , Terminology as Topic , Humans , Melanoma/diagnostic imaging , Skin Neoplasms/diagnostic imagingABSTRACT
BACKGROUND: Melanoma in situ and dysplastic nevi with severe atypia present overlapping histopathologic features. Reflectance confocal microscopy findings can be integrated with the dermatopathology report to improve differentiation between melanoma and dysplastic nevi with severe atypia. OBJECTIVE: To compare prevalence of reflectance confocal microscopy findings between melanoma in situ and dysplastic nevi with severe atypia. METHODS: This retrospective observational study compared reflectance confocal microscopy findings in dermatopathologically diagnosed dysplastic nevi with severe atypia and melanoma in situ, collected between 2007 and 2017 at a private pigmented-lesion clinic. Concordant pathologic diagnosis was defined as unanimous agreement between 3 dermatopathologists who independently reviewed all cases; all other cases were classified as discordant. RESULTS: The study included 112 lesions, 62 concordant melanomas in situ, 28 concordant dysplastic nevi with severe atypia, and 22 discordant lesions. In comparing reflectance confocal microscopy findings in concordant cases, melanoma in situ showed more frequently than dysplastic nevi with severe atypia the presence of epidermal atypical melanocytes as round cells (19/62 vs 0/28; P < .001) and dendritic cells (50/62 vs 6/28; P < .001), as well as a diffuse distribution of epidermal atypical melanocytes (50/54 vs 3/6; P = .002). In contrast, dysplastic nevi with severe atypia showed the presence of dense melanocytic nests more frequently than melanoma in situ did (15/28 vs 14/62; P = .003). LIMITATIONS: The study was based on a limited number of lesions originating from a single clinic. CONCLUSIONS: Reflectance confocal microscopy findings may help differentiate a subset of dysplastic nevi with severe atypia from melanoma in situ.
Subject(s)
Dysplastic Nevus Syndrome/diagnostic imaging , Dysplastic Nevus Syndrome/pathology , Melanoma/diagnostic imaging , Melanoma/pathology , Skin Neoplasms/diagnostic imaging , Skin Neoplasms/pathology , Adult , Aged , Aged, 80 and over , Cross-Sectional Studies , Diagnosis, Differential , Female , Humans , Langerhans Cells/pathology , Male , Melanocytes/pathology , Microscopy, Confocal/methods , Middle Aged , Retrospective Studies , Young AdultABSTRACT
BACKGROUND: Whether machine-learning algorithms can diagnose all pigmented skin lesions as accurately as human experts is unclear. The aim of this study was to compare the diagnostic accuracy of state-of-the-art machine-learning algorithms with human readers for all clinically relevant types of benign and malignant pigmented skin lesions. METHODS: For this open, web-based, international, diagnostic study, human readers were asked to diagnose dermatoscopic images selected randomly in 30-image batches from a test set of 1511 images. The diagnoses from human readers were compared with those of 139 algorithms created by 77 machine-learning labs, who participated in the International Skin Imaging Collaboration 2018 challenge and received a training set of 10â015 images in advance. The ground truth of each lesion fell into one of seven predefined disease categories: intraepithelial carcinoma including actinic keratoses and Bowen's disease; basal cell carcinoma; benign keratinocytic lesions including solar lentigo, seborrheic keratosis and lichen planus-like keratosis; dermatofibroma; melanoma; melanocytic nevus; and vascular lesions. The two main outcomes were the differences in the number of correct specific diagnoses per batch between all human readers and the top three algorithms, and between human experts and the top three algorithms. FINDINGS: Between Aug 4, 2018, and Sept 30, 2018, 511 human readers from 63 countries had at least one attempt in the reader study. 283 (55·4%) of 511 human readers were board-certified dermatologists, 118 (23·1%) were dermatology residents, and 83 (16·2%) were general practitioners. When comparing all human readers with all machine-learning algorithms, the algorithms achieved a mean of 2·01 (95% CI 1·97 to 2·04; p<0·0001) more correct diagnoses (17·91 [SD 3·42] vs 19·92 [4·27]). 27 human experts with more than 10 years of experience achieved a mean of 18·78 (SD 3·15) correct answers, compared with 25·43 (1·95) correct answers for the top three machine algorithms (mean difference 6·65, 95% CI 6·06-7·25; p<0·0001). The difference between human experts and the top three algorithms was significantly lower for images in the test set that were collected from sources not included in the training set (human underperformance of 11·4%, 95% CI 9·9-12·9 vs 3·6%, 0·8-6·3; p<0·0001). INTERPRETATION: State-of-the-art machine-learning classifiers outperformed human experts in the diagnosis of pigmented skin lesions and should have a more important role in clinical practice. However, a possible limitation of these algorithms is their decreased performance for out-of-distribution images, which should be addressed in future research. FUNDING: None.
Subject(s)
Algorithms , Dermoscopy , Internet , Machine Learning , Pigmentation Disorders/pathology , Skin Neoplasms/pathology , Adult , Female , Humans , Male , Reproducibility of Results , Retrospective StudiesABSTRACT
BACKGROUND: Reflectance confocal microscopy (RCM)-based skin cancer diagnosis requires proficiency. OBJECTIVE: To identify a short list of key RCM features of skin cancers and test their diagnostic utility. METHODS: We identified key RCM features through consensus among 6 experts using a modified Delphi method. To test the diagnostic utility of these RCM key features, 10 novice RCM readers evaluated a subset of 100 RCM cases from a retrospective data set of benign and malignant skin neoplasms. RESULTS: From 56 features reported in the literature, the experts identified 18 RCM features as highly valuable for skin cancer diagnosis. On the basis of consensus definitions, these RCM features were further clustered into 2 melanoma-specific key features (atypical cells and dermoepidermal junction disarray), 1 basal cell carcinoma-specific key feature (basaloid cords/islands), and 1 squamous cell carcinoma-specific key feature (keratinocyte disarray). The novice reading study showed that the presence of at least 1 of the 4 key features was associated with an overall sensitivity for skin cancer diagnosis of 91%, with a sensitivity for melanoma of 93%, a sensitivity for basal cell carcinoma of 92%, and a sensitivity for squamous cell carcinoma of 67%, and an overall specificity of 57%. LIMITATIONS: The consensus was based on only six RCM experts and the validation study was retrospective. CONCLUSIONS: A consensus terminology short list identifying the 4 RCM key features for skin cancer diagnosis may facilitate dissemination of RCM to novice users.