Computer algorithms show potential for improving dermatologists' accuracy to diagnose cutaneous melanoma: Results of the International Skin Imaging Collaboration 2017 - 11/02/20
, Konstantinos Liopyris, MD a, Stephen W. Dusza, DrPH a, Noel C.F. Codella, PhD b, David A. Gutman, MD, PhD c, d, e, Brian Helba, BS f, Aadi Kalloo, MHS a, Allan C. Halpern, MD afor the
International Skin Imaging Collaboration
Abstract |
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.
El texto completo de este artículo está disponible en PDF.Key words : automated melanoma diagnosis, computer algorithm, computer vision, deep learning, dermatologist, International Skin Imaging Collaboration, International Symposium on Biomedical Imaging, machine learning, melanoma, reader study, skin cancer
Abbreviations used : CI, ISBI, ISIC, ROC, SK
Esquema
| Funding sources: Supported in part through the National Institutes of Health, National Cancer Institute, Cancer Center Support Grant P30 CA008748. |
|
| Conflicts of interest: Dr Codella is an employee of IBM and an IBM stockholder. Dr Halpern is a consultant for Canfield Scientific Inc, Caliber ID, and SciBase. Dr Marchetti, Dr Liopyris, Dr Dusza, Dr Gutman, Mr Helba, and Mr Kalloo have no financial conflicts of interest to disclose. |
|
| Previously presented preliminary study data at the World Congress of Melanoma in Brisbane, Australia, in October 2017. |
Vol 82 - N° 3
P. 622-627 - mars 2020 Regresar al número
Artículo precedente
- Janus kinase inhibition induces disease remission in cutaneous sarcoidosis and granuloma annulare
- William Damsky, Durga Thakral, Meaghan K. McGeary, Jonathan Leventhal, Anjela Galan, Brett King
- Comorbid diseases of vitiligo: A 10-year cross-sectional retrospective study of an urban US population
- Ali Hadi, Jason F. Wang, Pushpinder Uppal, Lauren A. Penn, Nada Elbuluk
Bienvenido a EM-consulte, la referencia de los profesionales de la salud.
El acceso al texto completo de este artículo requiere una suscripción.
¿Ya suscrito a @@106933@@ revista ?