Highlights
- •Correct clinical diagnoses in skin lesions of face and scalp (FSL) are challenging.
- •Convolutional neural networks (CNN) are increasingly applied for skin diagnostics.
- •We investigated CNN-support for FSL diagnostically ‘unclear’ to dermatologists.
- •Following CNN classifications significantly reduced incorrect management decisions.
Abstract
Background
The clinical diagnosis of face and scalp lesions (FSL) is challenging due to overlapping
features. Dermatologists encountering diagnostically ‘unclear’ lesions may benefit
from artificial intelligence support via convolutional neural networks (CNN).
Methods
In a web-based classification task, dermatologists (n = 64) diagnosed a convenience
sample of 100 FSL as ‘benign’, ‘malignant’, or ‘unclear’ and indicated their management
decisions (‘no action’, ‘follow-up’, ‘treatment/excision’). A market-approved CNN
(Moleanalyzer-Pro®, FotoFinder Systems, Germany) was applied for binary classifications
(benign/malignant) of dermoscopic images.
Results
After reviewing one dermoscopic image per case, dermatologists labelled 562 of 6400
diagnoses (8.8%) as ‘unclear’ and mostly managed these by follow-up examinations (57.3%,
n = 322) or excisions (42.5%, n = 239). Management was incorrect in 58.8% of 291 truly
malignant cases (171 ‘follow-up’ or ‘no action’) and 43.9% of 271 truly benign cases
(119 ‘excision’). Accepting CNN classifications in unclear cases would have reduced
false management decisions to 4.1% in truly malignant and 31.7% in truly benign lesions
(both p < 0.01). After receiving full case information 239 diagnoses (3.7%) remained ‘unclear’
to dermatologists, now triggering more excisions (72.0%) than follow-up examinations
(28.0%). These management decisions were incorrect in 32.8% of 116 truly malignant
cases and 76.4% of 123 truly benign cases. Accepting CNN classifications would have
reduced false management decisions to 6.9% in truly malignant lesions and to 38.2%
in truly benign cases (both p < 0.01).
Conclusions
Dermatologists mostly managed diagnostically ‘unclear’ FSL by treatment/excision or
follow-up examination. Following CNN classifications as guidance in unclear cases
seems suitable to significantly reduce incorrect decisions.
Abbreviations:
95% CI (95% confidence interval), AI (artificial intelligence), AK (actinic keratosis), BCC (basal cell carcinoma), CNN (convolutional neural network), FSL (face and scalp lesion), LM (lentigo maligna), LMM (lentigo maligna melanoma), SK (seborrhoeic keratosis)Keywords
To read this article in full you will need to make a payment
Purchase one-time access:
Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online accessOne-time access price info
- For academic or personal research use, select 'Academic and Personal'
- For corporate R&D use, select 'Corporate R&D Professionals'
Subscribe:
Subscribe to European Journal of CancerAlready a print subscriber? Claim online access
Already an online subscriber? Sign in
Register: Create an account
Institutional Access: Sign in to ScienceDirect
References
- Basic histological structure and functions of facial skin.Clin Dermatol. 2014; 32: 3-13
- Dermoscopic clues to differentiate facial lentigo maligna from pigmented actinic keratosis.Br J Dermatol. 2016; 174: 1079-1085
- Dermatoscopy of flat pigmented facial lesions.J Eur Acad Dermatol Venereol. 2015; 29: 120-127
- Collision skin lesions-results of a multicenter study of the International Dermoscopy Society (IDS).Dermatol Pract Concept. 2017; 7: 51-62
- The dermatoscopic inverse approach significantly improves the accuracy of human readers for lentigo maligna diagnosis.J Am Acad Dermatol. 2020; 84: 381-389
- Improvement of early recognition of lentigo maligna using dermatoscopy.J Am Acad Dermatol. 2000; 42: 25-32
- Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists.Ann Oncol. 2018; 29: 1836-1842
- Dermatologist-level classification of skin cancer with deep neural networks.Nature. 2017; 542: 115-118
- Computer algorithms show potential for improving dermatologists' accuracy to diagnose cutaneous melanoma: Results of the International Skin Imaging Collaboration 2017.J Am Acad Dermatol. 2020; 82: 622-627
- World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects.JAMA. 2013; 310: 2191-2194
- Skin lesions of face and scalp - Classification by a market-approved convolutional neural network in comparison with 64 dermatologists.Eur J Cancer. 2021; 144: 192-199
- Man against machine reloaded: performance of a market-approved convolutional neural network in classifying a broad spectrum of skin lesions in comparison with 96 dermatologists working under less artificial conditions.Ann Oncol. 2020; 31: 137-143
- Association Between Surgical Skin Markings in Dermoscopic Images and Diagnostic Performance of a Deep Learning Convolutional Neural Network for Melanoma Recognition.JAMA Dermatol. 2019; 155: 1135-1141
- Human-computer collaboration for skin cancer recognition.Nat Med. 2020; 26: 1229-1234
- Addition of dermoscopy to conventional naked-eye examination in melanoma screening: a randomized study.J Am Acad Dermatol. 2004; 50: 683-689
- Dermoscopic predictors to discriminate between in situ and early invasive lentigo maligna melanoma: A retrospective observational study.J Am Acad Dermatol. 2020; 83: 269-271
- The peculiar dermatoscopic pattern of scalp melanoma.J Eur Acad Dermatol Venereol. 2022; 36: 1564-1567
- Clinical findings are not helpful in detecting lentigo maligna melanoma in patients with biopsy-proven lentigo maligna.J Eur Acad Dermatol Venereol. 2022; 36: 2325-2330
- Comparison of the accuracy of human readers versus machine-learning algorithms for pigmented skin lesion classification: an open, web-based, international, diagnostic study.Lancet Oncol. 2019; 20: 938-947
- Melanoma recognition by a deep learning convolutional neural network-Performance in different melanoma subtypes and localisations.Eur J Cancer. 2020; 127: 21-29
Article info
Publication history
Published online: March 03, 2023
Accepted:
February 26,
2023
Received in revised form:
February 24,
2023
Received:
January 19,
2023
Identification
Copyright
© 2023 Elsevier Ltd. All rights reserved.
