Your browser doesn't support javascript.
Mostrar: 20 | 50 | 100
Resultados 1 - 20 de 271
Filtrar
1.
Cells ; 9(5)2020 May 05.
Artigo em Inglês | MEDLINE | ID: mdl-32380691

RESUMO

INTRODUCTION: Mucosal melanoma is rare and associated with poorer prognosis in comparison to conventional melanoma subtypes. Little is known about the prognostic significance as well as possible associations between PARP1 and immunologic response in mucosal melanoma. METHODS: PARP1, PD-L1 and IDO1 immunostains were performed on 192 mucosal melanomas including 86 vulvar, 89 sinonasal, and 17 anorectal melanomas. RESULTS: By Kaplan-Meier analyses, high PARP1 expression correlated with worse overall and melanoma-specific survival (log-rank p values = 0.026 and 0.047, respectively). Tumors with combined PARP1 and IDO1 high expression correlated with worse overall and melanoma-specific survival (p = 0.015, 0.0034 respectively). By multivariate analyses, high PARP1 expression remained a predictor of worse survival independent of stage. By Fisher's exact test, high PARP1 expression correlated with highly mitogenic tumors (p = 0.02). High tumoral PD-L1 and IDO1 expression were associated with ulcerated primary tumors (p = 0.019, 0.0019, respectively). By linear regression analyses, correlations between PARP1 expression versus IDO1 expression (p = 0.0001) and mitotic index (p = 0.0052) were observed. CONCLUSION: Increased expression of PARP1 is an independent negative prognostic marker in mucosal melanomas. The association between PARP1 and IDO1 and their combined adverse prognostic role raise the potential of combined therapy in mucosal melanoma.

2.
Artigo em Inglês | MEDLINE | ID: mdl-32323445

RESUMO

Congenital melanocytic nevi (CMN) are cutaneous malformations whose prevalence is inversely correlated with projected adult size. CMN are caused by somatic mutations, but epidemiological studies suggest that germline genetic factors may influence CMN development. In CMN patients from the U.K., genetic variants in MC1R, such as p.V92M and loss-of-function variants, have been previously associated with larger CMN. We analyzed the association of MC1R variants with CMN characteristics in two distinct cohorts of medium-to-giant CMN patients from Spain (N=113) and from France, Norway, Canada and the U.S. (N=53), similar at the clinical and phenotypical level except for the number of nevi per patient. We found that the p.V92M or loss-of-function MC1R variants either alone or in combination did not correlate with CMN size, in contrast to the U.K. CMN patients. An additional case-control analysis with 259 unaffected Spanish individuals, showed a higher frequency of MC1R compound heterozygous or homozygous variant genotypes in Spanish CMN patients compared to the control population (15.9% vs. 9.3%; P=0.075). Altogether, this study suggests that MC1R variants are not associated with CMN size in these non-UK cohorts. Additional studies are required to define the potential role of MC1R as a risk factor in CMN development.

3.
J Am Acad Dermatol ; 2020 Apr 10.
Artigo em Inglês | MEDLINE | ID: mdl-32283231

RESUMO

BACKGROUND: CDKN2A, CDK4, and POT1 are well-established melanoma-susceptibility genes. OBJECTIVE: We evaluated melanoma histopathology for individuals with germline mutations of CDKN2A, CDK4, and POT1. METHODS: We assessed histopathology for melanomas diagnosed in melanoma-prone families (≥2 individuals with melanoma) from the United States, Italy, and Spain. Comparisons between mutation carriers and non-carriers (no mutation) were adjusted for age, sex, Breslow depth, and correlations among individuals within the same family. RESULTS: Histologic slides were evaluated for 290 melanomas (139 from 132 non-carriers, 122 from 68 CDKN2A carriers, 10 from 6 CDK4 carriers, 19 from 16 POT1 carriers). Superficial spreading was the predominant subtype for all groups. Spitzoid morphology (>25% of tumor) was observed in the majority of invasive melanomas from POT1 carriers (10 of 15 [67%], P<.0001 vs. non-carriers). This finding was independently confirmed by three expert melanoma dermatopathologists in 9 of 15 (60%) invasive melanomas. In situ and invasive melanomas from CDKN2A and CDK4 carriers were histologically similar to melanomas from non-carriers. LIMITATIONS: Limited sample sizes for rare melanoma-susceptibility syndromes (CDK4, POT1) CONCLUSION: Spitzoid morphology was associated with POT1 mutations suggesting that telomere dysfunction (POT1 mutations) may contribute to Spitzoid differentiation in melanocytic tumors.

4.
Acta Derm Venereol ; 100(10): adv00134, 2020 May 11.
Artigo em Inglês | MEDLINE | ID: mdl-32318743

RESUMO

Sutton naevi can sometimes present a challenging appearance with atypical presentation, also by dermoscopy. Reflectance confocal microscopy could help in making a diagnosis. This study prospectively collected two groups of Sutton nevi: the first one was composed by typical white halo naevi monitored for one year (13, 23%) and the second one was made up of atypical lesions excised in order to rule out melanoma, which were histologically diagnosed as Sutton naevi (21, 37%). These two groups of Sutton naevi were compared to a retrospectively collected cohort of thin melanomas with histologic regression features (23, 40%). On dermoscopy, atypical Sutton naevi and melanomas were indistinguishable. Reflectance confocal microscopy demonstrated significant differences at the dermo-epidermal junction: marked dermo-epidermal junction thickening and non-edged papilla were associated with melanoma, while the presence of nests was associated with Sutton naevi. However, reflectance confocal microscopy also detected marked intraepidermal pagetoid cells in Sutton naevi that were a combination of MelanA+ and CD1a+ cells. Sutton naevi can simulate melanoma, under both dermoscopy and reflectance confocal microscopy. Nevertheless, relevant confocal dermo-epidermal junction features and the clinical scenario can be helpful to make a final diagnosis, especially in those situations where melanoma must be ruled out.

5.
J Am Acad Dermatol ; 2020 Mar 05.
Artigo em Inglês | MEDLINE | ID: mdl-32147389

RESUMO

BACKGROUND: Pigmented labial macules (PLMs) are clinical, dermoscopic, and histopathologic challenges. OBJECTIVE: To describe and evaluate the utility of reflectance confocal microscopy (RCM) in PLMs and to establish a correlation between dermoscopy, RCM, histopathology, and immunohistochemistry. METHODS: Prospective study of PLMs from 4 tertiary referral dermatology centers. The study included 51 biopsy specimen-proven PLMs. Dermoscopic, RCM images, and histopathologic preparations were evaluated for malignant criteria. Diagnostic accuracy of RCM for melanoma diagnosis, RCM Lip Score previously reported, and κ values between techniques were calculated. RESULTS: Included were 5 melanomas and 46 benign PLMs. Dermoscopically, melanomas exhibited more frequently ≥3 colors and ≥3 structures. With RCM, pagetoid spreading, epithelial disarray, continuous proliferation of atypical cells around papillae, nonhomogeneously distributed papillae, marked cellular atypia, and a higher number of dendritic cells per papillae were more frequent in melanomas. The RCM Lip Score was significantly higher in malignant lesions. Good κ values were observed in most of the evaluated features. A perfect sensitivity and specificity was obtained combining dermoscopy and RCM. LIMITATIONS: A low number of melanomas were obtained. CONCLUSIONS: RCM improves lip melanoma diagnosis, and the RCM Lip Score represents a useful tool for the evaluation of a PLM.

7.
Acta Derm Venereol ; 100(8): adv00106, 2020 Apr 06.
Artigo em Inglês | MEDLINE | ID: mdl-32110813

RESUMO

Numerous dermoscopic structures for the early detection of melanoma have been described. The aim of this study was to illustrate the characteristics of dermoscopic structures that are similar to blotches, but smaller (termed microblotches), and to evaluate their association with other well-known dermoscopic structures. A cross-sectional study design, including 165 dermoscopic images of melanoma was used to define microblotches, and 241 consecutive images of naevi from the HAM10000 database, were studied to evaluate the prevalence of this criterion in both groups. Microblotches were defined as sharply demarcated structures ≤1 mm, with geographical borders visible only with dermoscopy. Microblotches were present in 38.7% of the melanomas and 6.7% of the naevi. Moreover, microblotches were associated with an odds ratio (OR) of malignancy of 5.79, and were more frequent in invasive melanoma than in the in-situ subtype (OR 2.92). Histologically, they correspond to hyperpigmented parakeratosis or epidermal consumption. In conclusion, microblotches are related to melanomas. This finding could help dermatologists to differentiate between naevi and melanomas.

8.
JAMA Oncol ; 2020 Jan 02.
Artigo em Inglês | MEDLINE | ID: mdl-31895407

RESUMO

Importance: Whether immune-related adverse events (irAEs) indicate drug activity in patients treated with immune checkpoint inhibitors remains unknown. Objective: To investigate the association between irAEs and recurrence-free survival (RFS) in the double-blind EORTC 1325/KEYNOTE-054 clinical trial comparing pembrolizumab therapy and placebo for the treatment of patients with high-risk stage III melanoma. Design, Setting, and Participants: A total of 1019 adults with stage III melanoma were randomly assigned on a 1:1 ratio to receive treatment with pembrolizumab therapy or placebo. Eligible patients were adults 18 years and older with complete resection of cutaneous melanoma metastatic to lymph nodes, classified with stage IIIA (at least 1 micrometastasis measuring >1 mm in greatest diameter), IIIB, or IIIC (without in-transit metastasis) cancer. Patients were randomized from August 26, 2015, to November 14, 2016. The clinical cutoff for the data set was October 2, 2017. Analyses were then performed on the database, which was locked on November 28, 2017. Interventions: Participants were scheduled to receive 200 mg of pembrolizumab or placebo every 3 weeks for a total of 18 doses for approximately 1 year or until disease recurrence, unacceptable toxic effects, major protocol violation, or withdrawal of consent. Main Outcomes and Measures: The association between irAEs and RFS was estimated using a Cox model adjusted for sex, age, and AJCC-7 stage, with a time-varying covariate that had a value of 0 before irAE onset and 1 after irAE onset. Results: Of 1011 patients who began treatment with pembrolizumab therapy or placebo, 622 (61.5%) were men and 389 (38.5%) were women; 386 patients (38.2%) were aged 50 to 64 years, 377 (37.3%) were younger than 50 years, and 248 (24.5%) were 65 years and older. Consistent with the reported main analysis in the intent-to-treat population, RFS was longer in the pembrolizumab arm compared with the placebo arm (hazard ratio [HR], 0.56; 98.4% CI, 0.43-0.74) among patients who started treatment. The incidence of irAEs was 190 (37.4%) in the pembrolizumab arm (n = 509) and 45 (9.0%) in the placebo arm (n = 502); in each treatment group, the incidence was similar for men and women. The occurrence of an irAE was associated with a longer RFS in the pembrolizumab arm (HR, 0.61; 95% CI, 0.39-0.95; P = .03) in both men and women. However, in the placebo arm, this association was not significant. Compared with the placebo arm, the reduction in the hazard of recurrence or death in the pembrolizumab arm was greater after the onset of an irAE than without or before an irAE (HR, 0.37; 95% CI, 0.24-0.57 vs HR, 0.61; 95% CI, 0.49-0.77, respectively; P = .03). Conclusions and Relevance: In this study, the occurrence of an irAE was associated with a longer RFS in the pembrolizumab arm. Trial Registrations: ClinicalTrials.gov identifier: NCT02362594; EudraCT identifier: 2014-004944-37.

9.
Eur J Surg Oncol ; 46(2): 263-271, 2020 Feb.
Artigo em Inglês | MEDLINE | ID: mdl-31594672

RESUMO

INTRODUCTION: Sentinel lymph node (SLN) biopsy is useful for the prognostic stratification of patients with thick melanoma. Identifying which variables are associated with SLN involvement and establishing risk in different subgroups of patients could be useful for guiding the indication of SLN biopsy. The value of complete lymph node dissection (CLND) in patients with a positive SLN biopsy is currently under debate. MATERIALS AND METHODS: To identify factors associated with SLN involvement in thick melanoma we performed a multicentric retrospective cohort study involving 660 patients with thick melanoma who had undergone SLN biopsy. To analyze the role of CLND in thick melanoma patients with a positive SLN biopsy, we built a multivariate Cox proportional hazards model for melanoma-specific survival (MSS) and disease-free survival (DFS) and compared 217 patients who had undergone CLND with 44 who had not. RESULTS: The logistic regression analysis showed that age, histologic subtype, ulceration, microscopic satellitosis, and lymphovascular invasion were associated with nodal disease. The CHAID (Chi-squared Automatic Interaction Detection) decision tree showed ulceration to be the most important predictor of lymphatic involvement. For nonulcerated melanomas, the histologic subtype lentigo maligna melanoma was associated with a low rate of SLN involvement (4.3%). No significant differences were observed for DFS and MSS between the CLND performed and not-performed groups. Nodal status on CLND was associated with differences in DFS and MSS rates. CONCLUSION: We identified subgroups of thick melanoma patients with a low likelihood of SLN involvement. CLND does not offer survival benefit, but provides prognostic information.

12.
PLoS Genet ; 15(11): e1008490, 2019 11.
Artigo em Inglês | MEDLINE | ID: mdl-31730655

RESUMO

Despite genetics being accepted as the primary cause of familial aggregation for most diseases, it is still unclear whether afflicted families are likely to share a single highly penetrant rare variant, many minimally penetrant common variants, or a combination of the two types of variants. We therefore use recent estimates of SNP heritability and the liability threshold model to estimate the proportion of afflicted families likely to carry a rare, causal variant. We then show that Polygenic Risk Scores (PRS) may be useful for identifying families likely to carry such a rare variant and therefore for prioritizing families to include in sequencing studies with that aim. Specifically, we introduce a new statistic that estimates the proportion of individuals carrying causal rare variants based on the family structure, disease pattern, and PRS of genotyped individuals. Finally, we consider data from the MelaNostrum consortium and show that, despite an estimated PRS heritability of only 0.05 for melanoma, families carrying putative causal variants had a statistically significantly lower PRS, supporting the idea that PRS prioritization may be a useful future tool. However, it will be important to evaluate whether the presence of rare mendelian variants are generally associated with the proposed test statistic or lower PRS in future and larger studies.


Assuntos
Doenças Genéticas Inatas/genética , Predisposição Genética para Doença , Melanoma/genética , Análise de Sequência de DNA , Alelos , Variação Genética/genética , Estudo de Associação Genômica Ampla , Genótipo , Humanos , Melanoma/epidemiologia , Melanoma/patologia , Herança Multifatorial/genética , Linhagem , Fenótipo , Polimorfismo de Nucleotídeo Único/genética , Fatores de Risco
13.
Pharm Pract (Granada) ; 17(3): 1455, 2019.
Artigo em Inglês | MEDLINE | ID: mdl-31592287

RESUMO

Background: Skin cancer incidence is increasing alarmingly, despite current efforts trying to improve its early detection. Community pharmacists have proven success in implementing screening protocols for a number of diseases because of their skills and easy access. Objective: To evaluate the prevalence of skin cancer risk factors and the photoprotection habits with a questionnaire in community pharmacy users. Methods: A research group consisting of pharmacists and dermatologists conducted a descriptive cross-sectional study to assess photoprotection habits and skin cancer risk factors by using a validated questionnaire in 218 community pharmacies in Barcelona from May 23rd to June 13th 2016. All participants received health education on photoprotection and skin cancer prevention. Patients with ≥1 skin cancer risk factor were referred to their physician, as they needed further screening of skin cancer. Results: A total of 5,530 participants were evaluated. Of those, only 20.2% participants had received a total body skin examination for skin cancer screening in the past by a physician and 57.1% reported using a SPF 50+ sunscreen. 53.9% participants presented ≥1 skin cancer risk factor: 11.8% participants reported having skin cancer familial history and 6.2% reported skin cancer personal history; pharmacists found ≥10 melanocytic nevi in 43.8% participants and chronically sun-damaged skin in 21.4%. Lesions suspicious for melanoma were reported in 10.9% of the participants and urgent dermatological evaluation was recommended. Conclusions: Pharmacists can detect people with skin cancer risk factors amongst their users. This intervention can be considered in multidisciplinary strategies of skin cancer screening.

14.
Pharm. pract. (Granada, Internet) ; 17(3): 0-0, jul.-sept. 2019. tab
Artigo em Inglês | IBECS | ID: ibc-ET1-4314

RESUMO

Background: Skin cancer incidence is increasing alarmingly, despite current efforts trying to improve its early detection. Community pharmacists have proven success in implementing screening protocols for a number of diseases because of their skills and easy access. Objective: To evaluate the prevalence of skin cancer risk factors and the photoprotection habits with a questionnaire in community pharmacy users. Methods: A research group consisting of pharmacists and dermatologists conducted a descriptive cross-sectional study to assess photoprotection habits and skin cancer risk factors by using a validated questionnaire in 218 community pharmacies in Barcelona from May 23rd to June 13th 2016. All participants received health education on photoprotection and skin cancer prevention. Patients with ≥1 skin cancer risk factor were referred to their physician, as they needed further screening of skin cancer. Results: A total of 5,530 participants were evaluated. Of those, only 20.2% participants had received a total body skin examination for skin cancer screening in the past by a physician and 57.1% reported using a SPF 50+ sunscreen. 53.9% participants presented ≥1 skin cancer risk factor: 11.8% participants reported having skin cancer familial history and 6.2% reported skin cancer personal history; pharmacists found ≥10 melanocytic nevi in 43.8% participants and chronically sun-damaged skin in 21.4%. Lesions suspicious for melanoma were reported in 10.9% of the participants and urgent dermatological evaluation was recommended. Conclusions: Pharmacists can detect people with skin cancer risk factors amongst their users. This intervention can be considered in multidisciplinary strategies of skin cancer screening


No disponible


Assuntos
Humanos , Serviços Comunitários de Farmácia/organização & administração , Neoplasias Cutâneas/epidemiologia , Proteção Radiológica/estatística & dados numéricos , Protetores Solares/farmacocinética , Radiação Solar/efeitos adversos , Detecção Precoce de Câncer/métodos , Estudos Transversais , Fatores de Risco , Variação Biológica da População
15.
Biomed Opt Express ; 10(7): 3404-3409, 2019 Jul 01.
Artigo em Inglês | MEDLINE | ID: mdl-31467785

RESUMO

The effective and non-invasive diagnosis of skin cancer is a hot topic, since biopsy is a costly and time-consuming surgical procedure. As skin relief is an important biophysical feature that can be difficult to perceive with the naked eye and by touch, we developed a novel 3D imaging scanner based on fringe projection to obtain morphological parameters of skin lesions related to perimeter, area and volume with micrometric precision. We measured 608 samples and significant morphological differences were found between melanomas and nevi (p<0.001). The capacity of the 3D scanner to distinguish these lesions was supported by a supervised machine learning algorithm resulting in 80.0% sensitivity and 76.7% specificity.

17.
Cancer Med ; 8(9): 4235-4244, 2019 Aug.
Artigo em Inglês | MEDLINE | ID: mdl-31215168

RESUMO

Mitotic rate is no longer considered a staging criterion for thin melanoma in the 8th edition of the American Joint Committee on Cancer Staging Manual. The aim of this observational study was to identify prognostic factors for thin melanoma and predictors and prognostic significance of sentinel lymph node (SLN) involvement in a large multicenter cohort of patients with melanoma from nine tertiary care hospitals. A total of 4249 consecutive patients with thin melanoma diagnosed from January 1, 1998 to December 31, 2016 were included. The main outcomes were disease-free interval and melanoma-specific survival for the overall population and predictors of SLN metastasis (n = 1083). Associations between survival and SLN status and different clinical and pathologic variables (sex, age, tumor location, mitosis, ulceration, regression, lymphovascular invasion, histologic subtype, Clark level, and Breslow thickness) were analyzed by Cox proportional hazards regression and logistic regression. SLN status was the most important prognostic factor for melanoma-specific survival (hazard ratio, 13.8; 95% CI, 6.1-31.2; P < 0.001), followed by sex, ulceration, and Clark level for patients who underwent SLNB. A mitotic rate of >2 mitoses/mm2 was the only factor associated with a positive SLN biopsy (odds ratio, 2.9; 95% CI, 1.22-7; P = 0.01. SLN status is the most important prognostic factor in thin melanoma. A high mitotic rate is associated with metastatic SLN involvement. SLN biopsy should be discussed and recommended in patients with thin melanoma and a high mitotic rate.

18.
J Am Acad Dermatol ; 2019 Jun 13.
Artigo em Inglês | MEDLINE | ID: mdl-31202873

RESUMO

BACKGROUND: Reflectance confocal microscopy (RCM) allows accurate, noninvasive, in vivo diagnosis for skin cancer. However, its impact on physicians' diagnostic confidence and management is unknown. OBJECTIVES: We sought to assess the physicians' diagnostic confidence and management before and after RCM of equivocal skin lesions. METHODS: Prospective, 2-center, observational study. During clinical practice, 7 dermatologists recorded their diagnostic confidence level (measured in a scale from 0 to 10), diagnosis, and management before and after RCM of clinically/dermoscopically equivocal lesions that raised concern for skin cancer. We also evaluated the diagnostic accuracy before and after RCM. RESULTS: We included 272 consecutive lesions from 226 individuals (mean age, 53.5 years). Diagnostic confidence increased from 6.2 to 8.1 after RCM (P < .001) when RCM confirmed or changed the diagnosis. Lesion management changed in 33.5% cases after RCM (to observation in 51 cases and to biopsy/excision in 31 cases). After RCM, the number needed to excise was 1.2. Sensitivity for malignancy before and after RCM was 78.2% and 85.1%, respectively. Specificity before and after RCM was 78.8% and 80%, respectively. LIMITATIONS: Small sample size, real-life environment, and different levels of expertise among RCM users. CONCLUSION: Physicians' diagnostic confidence and accuracy increased after RCM when evaluating equivocal tumors, frequently resulting in management changes while maintaining high diagnostic accuracy.

19.
Eur J Cancer ; 116: 148-157, 2019 Jul.
Artigo em Inglês | MEDLINE | ID: mdl-31200321

RESUMO

BACKGROUND: The American Joint Committee on Cancer-8 (AJCC) classification of melanoma was implemented in January 2018. It was based on data gathered when checkpoint inhibitors were not used as adjuvant therapy in stage III melanoma. The European Organization for Research and Treatment of Cancer (EORTC) 1325/KEYNOTE-054 double-blind phase III trial evaluated pembrolizumab vs placebo in AJCC-7 stage IIIA (excluding lymph node metastasis ≤1 mm), IIIB or IIIC (without in-transit metastasis) patients after complete lymphadenectomy. PATIENTS, METHODS AND RESULTS: Patients (n = 1019) were randomised 1:1 to pembrolizumab 200 mg or placebo every 3 weeks (total of 18 doses, ∼1 year). At 1.25-year median follow-up, pembrolizumab prolonged relapse-free survival (RFS) in the total population (1-year RFS rate: 75.4% vs 61.0%; hazard ratio [HR] 0.57; logrank P < 0.0001) and consistently in the AJCC-7 subgroups. Prognostic and predictive values of AJCC-8 for RFS were evaluated in this study. Patient distribution according to the AJCC-8 stage subgroups was 8% (IIIA), 34.7% (IIIB), 49.7% (IIIC), 3.7% (IIID) and 3.8% (unknown). AJCC-8 classification was strongly associated with RFS (HRs for stage IIIB, IIIC and IIID vs IIIA were 4.0, 5.7 and 12.2, respectively) but showed no predictive importance for the treatment comparison regarding RFS (test for interaction: P = 0.68). The 1-year RFS rate for pembrolizumab vs placebo and the HRs (99% confidence interval) within each AJCC-8 subgroup were as follows: stage IIIA (92.7% vs 92.5%; 0.76 [0.11-5.43]), IIIB (79.0% vs 65.5%; 0.59 [0.35-0.99]), IIIC (73.6% vs 53.9%; 0.48 [0.33-0.70]) and IIID (50.0% vs 33.3%; 0.69 [0.24-2.00]). CONCLUSIONS: AJCC-8 staging had a strong prognostic importance for RFS but no predictive importance: the RFS benefit of pembrolizumab was observed across AJCC-8 subgroups in resected high-risk stage III melanoma patients.

20.
Lancet Oncol ; 20(7): 938-947, 2019 Jul.
Artigo em Inglês | MEDLINE | ID: mdl-31201137

RESUMO

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 10 015 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.

SELEÇÃO DE REFERÊNCIAS
DETALHE DA PESQUISA