Computer algorithms show potential for improving dermatologists' accuracy to diagnose cutaneous melanoma: Results of the International Skin Imaging Collaboration 2017.

Marchetti, Michael A; Liopyris, Konstantinos; Dusza, Stephen W; Codella, Noel C F; Gutman, David A; Helba, Brian; Kalloo, Aadi; Halpern, Allan C

Marchetti, Michael A; Liopyris, Konstantinos; Dusza, Stephen W; Codella, Noel C F; Gutman, David A; Helba, Brian; Kalloo, Aadi; Halpern, Allan C.

Afiliação

Marchetti MA; Dermatology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, New York. Electronic address: marchetm@mskcc.org.
Liopyris K; Dermatology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, New York.
Dusza SW; Dermatology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, New York.
Codella NCF; IBM Research Division, Thomas J. Watson Research Center, Yorktown Heights, New York.
Gutman DA; Department of Neurology, Emory University School of Medicine, Atlanta, Georgia; Department of Psychiatry, Emory University School of Medicine, Atlanta, Georgia; Department of Biomedical Informatics, Emory University School of Medicine, Atlanta, Georgia.
Helba B; Kitware Inc, Clifton Park, New York.
Kalloo A; Dermatology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, New York.
Halpern AC; Dermatology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, New York.

J Am Acad Dermatol ; 82(3): 622-627, 2020 Mar.

Article em En | MEDLINE | ID: mdl-31306724

RESUMO

BACKGROUND: Computer vision has promise in image-based cutaneous melanoma diagnosis but clinical utility is uncertain. OBJECTIVE: To determine if computer algorithms from an international melanoma detection challenge can improve dermatologists' accuracy in diagnosing melanoma. METHODS: In this cross-sectional study, we used 150 dermoscopy images (50 melanomas, 50 nevi, 50 seborrheic keratoses) from the test dataset of a melanoma detection challenge, along with algorithm results from 23 teams. Eight dermatologists and 9 dermatology residents classified dermoscopic lesion images in an online reader study and provided their confidence level. RESULTS: The top-ranked computer algorithm had an area under the receiver operating characteristic curve of 0.87, which was higher than that of the dermatologists (0.74) and residents (0.66) (P < .001 for all comparisons). At the dermatologists' overall sensitivity in classification of 76.0%, the algorithm had a superior specificity (85.0% vs. 72.6%, P = .001). Imputation of computer algorithm classifications into dermatologist evaluations with low confidence ratings (26.6% of evaluations) increased dermatologist sensitivity from 76.0% to 80.8% and specificity from 72.6% to 72.8%. LIMITATIONS: Artificial study setting lacking the full spectrum of skin lesions as well as clinical metadata. CONCLUSION: Accumulating evidence suggests that deep neural networks can classify skin images of melanoma and its benign mimickers with high accuracy and potentially improve human performance.

Assuntos

Aprendizado Profundo; Dermoscopia/métodos; Interpretação de Imagem Assistida por Computador/métodos; Melanoma/diagnóstico; Neoplasias Cutâneas/diagnóstico; Colômbia; Estudos Transversais; Dermatologistas/estatística & dados numéricos; Dermoscopia/estatística & dados numéricos; Diagnóstico Diferencial; Humanos; Cooperação Internacional; Internato e Residência/estatística & dados numéricos; Israel; Ceratose Seborreica/diagnóstico; Melanoma/patologia; Nevo/diagnóstico; Curva ROC; Pele/diagnóstico por imagem; Pele/patologia; Neoplasias Cutâneas/patologia; Espanha; Estados Unidos

Palavras-chave

International Skin Imaging Collaboration; International Symposium on Biomedical Imaging; automated melanoma diagnosis; computer algorithm; computer vision; deep learning; dermatologist; machine learning; melanoma; reader study; skin cancer

Texto completo

Adicionar na Minha BVS

Imprimir

XML

PubMed Links

Buscar no Google

Texto completo: 1 Coleções: 01-internacional Base de dados: MEDLINE Assunto principal: Neoplasias Cutâneas / Interpretação de Imagem Assistida por Computador / Dermoscopia / Aprendizado Profundo / Melanoma Tipo de estudo: Diagnostic_studies / Observational_studies / Prevalence_studies / Prognostic_studies / Risk_factors_studies Limite: Humans País/Região como assunto: America do norte / America do sul / Asia / Colombia / Europa Idioma: En Revista: J Am Acad Dermatol Ano de publicação: 2020 Tipo de documento: Article País de publicação: Estados Unidos

Texto completo

Adicionar na Minha BVS

Imprimir

XML

PubMed Links

Buscar no Google