Skin cancer detection by deep learning and sound analysis algorithms: A prospective clinical study of an elementary dermoscope.
Artificial intelligence
Deep learning
Dermoscopy
Melanoma
Skin cancer
Sonification
Telemedicine
Journal
EBioMedicine
ISSN: 2352-3964
Titre abrégé: EBioMedicine
Pays: Netherlands
ID NLM: 101647039
Informations de publication
Date de publication:
May 2019
May 2019
Historique:
received:
18
03
2019
revised:
16
04
2019
accepted:
29
04
2019
pubmed:
19
5
2019
medline:
26
11
2019
entrez:
19
5
2019
Statut:
ppublish
Résumé
Skin cancer (SC), especially melanoma, is a growing public health burden. Experimental studies have indicated a potential diagnostic role for deep learning (DL) algorithms in identifying SC at varying sensitivities. Previously, it was demonstrated that diagnostics by dermoscopy are improved by applying an additional sonification (data to sound waves conversion) layer on DL algorithms. The aim of the study was to determine the impact of image quality on accuracy of diagnosis by sonification employing a rudimentary skin magnifier with polarized light (SMP). Dermoscopy images acquired by SMP were processed by a first deep learning algorithm and sonified. Audio output was further analyzed by a different secondary DL. Study criteria outcomes of SMP were specificity and sensitivity, which were further processed by a F2-score, i.e. applying a twice extra weight to sensitivity over positive predictive values. Patients (n = 73) fulfilling inclusion criteria were referred to biopsy. SMP analysis metrics resulted in a receiver operator characteristic curve AUC's of 0.814 (95% CI, 0.798-0.831). SMP achieved a F2-score sensitivity of 91.7%, specificity of 41.8% and positive predictive value of 57.3%. Diagnosing the same set of patients' lesions by an advanced dermoscope resulted in a F2-score sensitivity of 89.5%, specificity of 57.8% and a positive predictive value of 59.9% (P=NS). DL processing of dermoscopic images followed by sonification results in an accurate diagnostic output for SMP, implying that the quality of the dermoscope is not the major factor influencing DL diagnosis of skin cancer. Present system might assist all healthcare providers as a feasible computer-assisted detection system. FUND: Bostel Technologies. Trial Registration clinicaltrials.gov Identifier: NCT03362138.
Sections du résumé
BACKGROUND
BACKGROUND
Skin cancer (SC), especially melanoma, is a growing public health burden. Experimental studies have indicated a potential diagnostic role for deep learning (DL) algorithms in identifying SC at varying sensitivities. Previously, it was demonstrated that diagnostics by dermoscopy are improved by applying an additional sonification (data to sound waves conversion) layer on DL algorithms. The aim of the study was to determine the impact of image quality on accuracy of diagnosis by sonification employing a rudimentary skin magnifier with polarized light (SMP).
METHODS
METHODS
Dermoscopy images acquired by SMP were processed by a first deep learning algorithm and sonified. Audio output was further analyzed by a different secondary DL. Study criteria outcomes of SMP were specificity and sensitivity, which were further processed by a F2-score, i.e. applying a twice extra weight to sensitivity over positive predictive values.
FINDINGS
RESULTS
Patients (n = 73) fulfilling inclusion criteria were referred to biopsy. SMP analysis metrics resulted in a receiver operator characteristic curve AUC's of 0.814 (95% CI, 0.798-0.831). SMP achieved a F2-score sensitivity of 91.7%, specificity of 41.8% and positive predictive value of 57.3%. Diagnosing the same set of patients' lesions by an advanced dermoscope resulted in a F2-score sensitivity of 89.5%, specificity of 57.8% and a positive predictive value of 59.9% (P=NS).
INTERPRETATION
CONCLUSIONS
DL processing of dermoscopic images followed by sonification results in an accurate diagnostic output for SMP, implying that the quality of the dermoscope is not the major factor influencing DL diagnosis of skin cancer. Present system might assist all healthcare providers as a feasible computer-assisted detection system. FUND: Bostel Technologies. Trial Registration clinicaltrials.gov Identifier: NCT03362138.
Identifiants
pubmed: 31101596
pii: S2352-3964(19)30294-4
doi: 10.1016/j.ebiom.2019.04.055
pmc: PMC6562065
pii:
doi:
Banques de données
ClinicalTrials.gov
['NCT03362138']
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
107-113Informations de copyright
Copyright © 2019 The Authors. Published by Elsevier B.V. All rights reserved.
Références
J Am Acad Dermatol. 2007 May;56(5):759-67
pubmed: 17316894
Arch Dermatol. 2007 Mar;143(3):329-38
pubmed: 17372097
J Am Acad Dermatol. 2012 Jul;67(1):54-9
pubmed: 21982636
Br J Dermatol. 2014 Nov;171(5):1099-107
pubmed: 24841846
J Eur Acad Dermatol Venereol. 2014 Nov;28(11):1442-9
pubmed: 25493316
JAMA Dermatol. 2016 Dec 1;152(12):1327-1334
pubmed: 27542070
J Eur Acad Dermatol Venereol. 2017 Jun;31(6):972-977
pubmed: 27896853
BMJ. 2017 Jun 28;357:j2813
pubmed: 28659278
J Am Acad Dermatol. 2018 Apr;78(4):701-709.e1
pubmed: 29180093
PLoS One. 2017 Dec 22;12(12):e0189828
pubmed: 29272283
Ann Oncol. 2018 Aug 1;29(8):1836-1842
pubmed: 29846502
J Invest Dermatol. 2018 Oct;138(10):2277-2279
pubmed: 29864435
Int J Cancer. 2019 Apr 15;144(8):1941-1953
pubmed: 30350310
Eur J Cancer. 2018 Dec;105:33-40
pubmed: 30384014
Cochrane Database Syst Rev. 2018 Dec 04;12:CD011902
pubmed: 30521682
EBioMedicine. 2019 Feb;40:176-183
pubmed: 30674442
IEEE J Biomed Health Inform. 2019 Mar;23(2):474-478
pubmed: 30703051