The dermoscopic inverse approach significantly improves the accuracy of human readers for lentigo maligna diagnosis.
Adult
Aged
Datasets as Topic
Dermatologists
/ statistics & numerical data
Dermoscopy
/ methods
Diagnosis, Differential
Early Detection of Cancer
/ methods
Female
Humans
Hutchinson's Melanotic Freckle
/ diagnosis
Image Processing, Computer-Assisted
/ statistics & numerical data
Keratosis, Actinic
/ diagnosis
Keratosis, Seborrheic
/ diagnosis
Male
Middle Aged
Neural Networks, Computer
Sensitivity and Specificity
Skin
/ diagnostic imaging
Skin Neoplasms
/ diagnosis
Young Adult
artificial intelligence
dermatoscopy
dermoscopy
diagnosis
inverse approach
maligna
melanoma
pigmented actinic keratosis
solar lentigo
Journal
Journal of the American Academy of Dermatology
ISSN: 1097-6787
Titre abrégé: J Am Acad Dermatol
Pays: United States
ID NLM: 7907132
Informations de publication
Date de publication:
Feb 2021
Feb 2021
Historique:
received:
21
04
2020
revised:
16
06
2020
accepted:
19
06
2020
pubmed:
28
6
2020
medline:
30
7
2021
entrez:
28
6
2020
Statut:
ppublish
Résumé
A recently introduced dermoscopic method for the diagnosis of early lentigo maligna (LM) is based on the absence of prevalent patterns of pigmented actinic keratosis and solar lentigo/flat seborrheic keratosis. We term this the inverse approach. To determine whether training on the inverse approach increases the diagnostic accuracy of readers compared to classic pattern analysis. We used clinical and dermoscopic images of histopathologically diagnosed LMs, pigmented actinic keratoses, and solar lentigo/flat seborrheic keratoses. Participants in a dermoscopy masterclass classified the lesions at baseline and after training on pattern analysis and the inverse approach. We compared their diagnostic performance among the 3 timepoints and to that of a trained convolutional neural network. The mean sensitivity for LM without training was 51.5%; after training on pattern analysis, it increased to 56.7%; and after learning the inverse approach, it increased to 83.6%. The mean proportions of correct answers at the 3 timepoints were 62.1%, 65.5, and 78.5%. The percentages of readers outperforming the convolutional neural network were 6.4%, 15.4%, and 53.9%, respectively. The experimental setting and the inclusion of histopathologically diagnosed lesions only. The inverse approach, added to the classic pattern analysis, significantly improves the sensitivity of human readers for early LM diagnosis.
Sections du résumé
BACKGROUND
BACKGROUND
A recently introduced dermoscopic method for the diagnosis of early lentigo maligna (LM) is based on the absence of prevalent patterns of pigmented actinic keratosis and solar lentigo/flat seborrheic keratosis. We term this the inverse approach.
OBJECTIVE
OBJECTIVE
To determine whether training on the inverse approach increases the diagnostic accuracy of readers compared to classic pattern analysis.
METHODS
METHODS
We used clinical and dermoscopic images of histopathologically diagnosed LMs, pigmented actinic keratoses, and solar lentigo/flat seborrheic keratoses. Participants in a dermoscopy masterclass classified the lesions at baseline and after training on pattern analysis and the inverse approach. We compared their diagnostic performance among the 3 timepoints and to that of a trained convolutional neural network.
RESULTS
RESULTS
The mean sensitivity for LM without training was 51.5%; after training on pattern analysis, it increased to 56.7%; and after learning the inverse approach, it increased to 83.6%. The mean proportions of correct answers at the 3 timepoints were 62.1%, 65.5, and 78.5%. The percentages of readers outperforming the convolutional neural network were 6.4%, 15.4%, and 53.9%, respectively.
LIMITATIONS
CONCLUSIONS
The experimental setting and the inclusion of histopathologically diagnosed lesions only.
CONCLUSIONS
CONCLUSIONS
The inverse approach, added to the classic pattern analysis, significantly improves the sensitivity of human readers for early LM diagnosis.
Identifiants
pubmed: 32592885
pii: S0190-9622(20)31184-1
doi: 10.1016/j.jaad.2020.06.085
pii:
doi:
Types de publication
Comparative Study
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
381-389Informations de copyright
Copyright © 2020 American Academy of Dermatology, Inc. Published by Elsevier Inc. All rights reserved.