Non-melanoma skin cancer diagnosis: a comparison between dermoscopic and smartphone images by unified visual and sonification deep learning algorithms.
Deep learning
Dermoscopy
Non-melanoma skin cancer
Preventive medicine
Sonification
Telemedicine
Journal
Journal of cancer research and clinical oncology
ISSN: 1432-1335
Titre abrégé: J Cancer Res Clin Oncol
Pays: Germany
ID NLM: 7902060
Informations de publication
Date de publication:
Sep 2022
Sep 2022
Historique:
received:
08
09
2021
accepted:
15
09
2021
pubmed:
22
9
2021
medline:
6
8
2022
entrez:
21
9
2021
Statut:
ppublish
Résumé
Non-melanoma skin cancer (NMSC) is the most frequent keratinocyte-origin skin tumor. It is confirmed that dermoscopy of NMSC confers a diagnostic advantage as compared to visual face-to-face assessment. COVID-19 restrictions diagnostics by telemedicine photos, which are analogous to visual inspection, displaced part of in-person visits. This study evaluated by a dual convolutional neural network (CNN) performance metrics in dermoscopic (DI) versus smartphone-captured images (SI) and tested if artificial intelligence narrows the proclaimed gap in diagnostic accuracy. A CNN that receives a raw image and predicts malignancy, overlaid by a second independent CNN which processes a sonification (image-to-sound mapping) of the original image, were combined into a unified malignancy classifier. All images were histopathology-verified in a comparison between NMSC and benign skin lesions excised as suspected NMSCs. Study criteria outcomes were sensitivity and specificity for the unified output. Images acquired by DI (n = 132 NMSC, n = 33 benign) were compared to SI (n = 170 NMSC, n = 28 benign). DI and SI analysis metrics resulted in an area under the curve (AUC) of the receiver operator characteristic curve of 0.911 and 0.821, respectively. Accuracy was increased by DI (0.88; CI 81.9-92.4) as compared to SI (0.75; CI 68.1-80.6, p < 0.005). Sensitivity of DI was higher than SI (95.3%, CI 90.4-98.3 vs 75.3%, CI 68.1-81.6, p < 0.001), but not specificity (p = NS). Telemedicine use of smartphone images might result in a substantial decrease in diagnostic performance as compared to dermoscopy, which needs to be considered by both healthcare providers and patients.
Identifiants
pubmed: 34546412
doi: 10.1007/s00432-021-03809-x
pii: 10.1007/s00432-021-03809-x
pmc: PMC8453469
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
2497-2505Informations de copyright
© 2021. The Author(s).
Références
Baumann BC, MacArthur KM, Brewer JD, Mendenhall WM, Barker BA, Etzkorn JR, Jellinek NJ et al (2020) Management of primary skin cancer during a pandemic: multidisciplinary recommendations. Cancer 126(17):3900–3906. https://doi.org/10.1002/cncr.32969
doi: 10.1002/cncr.32969
pubmed: 32478867
Bridges C, Calli M, McElroy JA, Quinn K, Dyer J, Becevic M (2019) Utility of dermatology extension for community healthcare outcomes (ECHO) sessions in the adult and paediatric population. J Telemed Telecare 27(6):376–381. https://doi.org/10.1177/1357633X19874200
doi: 10.1177/1357633X19874200
pubmed: 31526083
Celebi ME, Hassan AK, Vela PA (2013) A comparative study of efficient initialization methods for the k-means clustering algorithm. Expert Syst Appl 40(1):200–210. https://doi.org/10.1016/J.ESWA.2012.07.021
doi: 10.1016/J.ESWA.2012.07.021
Codella NCF, Gutman D, Celebi ME, Helba B, Marchetti MA, Dusza SW, Kalloo A et al (2018) Skin lesion analysis toward melanoma detection: a challenge at the 2017 international symposium on biomedical imaging (ISBI), hosted by the International Skin Imaging Collaboration (ISIC). http://arxiv.org/abs/1710.05006
Dascalu A, David EO (2019) Skin cancer detection by deep learning and sound analysis algorithms: a prospective clinical study of an elementary dermoscope. EBioMedicine 43(5):107–113. https://doi.org/10.1016/j.ebiom.2019.04.055
doi: 10.1016/j.ebiom.2019.04.055
pubmed: 31101596
pmcid: 6562065
Dinnes J, Deeks JJ, Chuchu N, Matin RN, Wong KY, Aldridge RB, Durack A et al (2018) Visual inspection and dermoscopy, alone or in combination, for diagnosing keratinocyte skin cancers in adults. Cochrane Database Syst Rev 12(12):CD011901. https://doi.org/10.1002/14651858.CD011901.pub2
doi: 10.1002/14651858.CD011901.pub2
pubmed: 30521688
Dréno B, Amici JM, Basset-Seguin N, Cribier B, Claudel JP, Richard MA (2014) Management of actinic keratosis: a practical report and treatment algorithm from expert clinicians. J Eur Acad Dermatol Venereol 28(9):1141–1149. https://doi.org/10.1111/jdv.12434
doi: 10.1111/jdv.12434
pubmed: 24612407
Fargnoli MC, Kostaki D, Piccioni A, Micantonio T, Peris K (2012) Dermoscopy in the diagnosis and management of non-melanoma skin cancers. Eur J Dermatol 22(4):456–463. https://doi.org/10.1684/ejd.2012.1727
doi: 10.1684/ejd.2012.1727
pubmed: 22534287
Ferrándiz L, Ojeda-Vila T, Corrales A, Martín-Gutiérrez FJ, Ruíz-de-Casas A, Galdeano R, o Álvarez-Torralba I, et al (2017) Internet-based skin cancer screening using clinical images alone or in conjunction with dermoscopic images: a randomized teledermoscopy trial. J Am Acad Dermatol 76(4):676–682. https://doi.org/10.1016/J.JAAD.2016.10.041
doi: 10.1016/J.JAAD.2016.10.041
pubmed: 28089728
Freeman K, Dinnes J, Chuchu N, Takwoingi Y, Bayliss SE, Matin RN, Jain A, Walter FM, Williams HC, Deeks JJ (2020) Algorithm based smartphone apps to assess risk of skin cancer in adults: systematic review of diagnostic accuracy studies. BMJ 368:m127. https://doi.org/10.1136/bmj.m127
doi: 10.1136/bmj.m127
pubmed: 32041693
pmcid: 7190019
Garg A, Noelle LW, Andrew IP, Dicker AP (2018) Clinical integration of digital solutions in health care: an overview of the current landscape of digital technologies in cancer care. JCO Clin Cancer Inform 2(12):1–9. https://doi.org/10.1200/CCI.17.00159
doi: 10.1200/CCI.17.00159
pubmed: 30652580
Gerd G, Gaissmaier W (2011) Heuristic decision making. Annu Rev Psychol 62(2):451–482. https://doi.org/10.1146/annurev-psych-120709-145346
doi: 10.1146/annurev-psych-120709-145346
Giavina-Bainchi M, Azevedo MFD, Sousa RM, Cordioli E (2020) Part II: accuracy of teledermatology in skin neoplasms. Front Med 7:598903. https://doi.org/10.3389/fmed.2020.598903
doi: 10.3389/fmed.2020.598903
Jeyaraj PR, Nadar ERS (2019) Computer-assisted medical image classification for early diagnosis of oral cancer employing deep learning algorithm. J Cancer Res Clin Oncol 145(4):829–837. https://doi.org/10.1007/S00432-018-02834-7
doi: 10.1007/S00432-018-02834-7
pubmed: 30603908
Kann BH, Hicks DF, Payabvash S, Mahajan A, Du J, Gupta V, Park HS et al (2020) Multi-institutional validation of deep learning for pretreatment identification of extranodal extension in head and neck squamous cell carcinoma. J Clinical Oncol 38(12):1304–1311. https://doi.org/10.1200/JCO.19.02031
doi: 10.1200/JCO.19.02031
Kuziemsky C, Maeder AJ, John O, Gogia SB, Basu A, Meher S, Ito M (2019) Role of artificial intelligence within the telehealth domain: official 2019 yearbook contribution by the members of IMIA telehealth working group. Yearb Med Inform 28(1):35–40. https://doi.org/10.1055/S-0039-1677897
doi: 10.1055/S-0039-1677897
pubmed: 31022750
pmcid: 6697552
Lio PA, Nghiem P (2004) Interactive atlas of dermoscopy. J Am Acad Dermatol 50(5):807–808. https://doi.org/10.1016/j.jaad.2003.07.029
doi: 10.1016/j.jaad.2003.07.029
Nault A, Zhang C, Kim KM, Saha S, Bennett DD, Xu YG (2015) Biopsy use in skin cancer diagnosis: comparing dermatology physicians and advanced practice professionals. JAMA Dermatol 151(8):899–902. https://doi.org/10.1001/jamadermatol.2015.0173
doi: 10.1001/jamadermatol.2015.0173
pubmed: 25806897
Nehal KS, Bichakjian CK (2018) Update on keratinocyte carcinomas. N Engl J Med 379(4):363–374. https://doi.org/10.1056/NEJMRA1708701
doi: 10.1056/NEJMRA1708701
pubmed: 30044931
Pacheco AGC, Lima GR, Salomão AS, Krohling B, Biral IP, de Angelo GG, Alves FCR et al (2020) PAD-UFES-20: a skin lesion dataset composed of patient data and clinical images collected from smartphones. Data Brief 32(8):106221. https://doi.org/10.1016/j.dib.2020.106221
doi: 10.1016/j.dib.2020.106221
pubmed: 32939378
pmcid: 7479321
Papageorgiou V, Apalla Z, Sotiriou E, Papageorgiou C, Lazaridou E, Vakirlis S, Ioannides D, Lallas A (2018) The Limitations of dermoscopy: false-positive and false-negative tumours. J Eur Acad Dermat Venereol 32(6):879–888. https://doi.org/10.1111/jdv.14782
doi: 10.1111/jdv.14782
Privalle A, Havighurst T, Kim KM, Bennett DD, Xu YG (2020) Number of skin biopsies needed per malignancy: comparing the use of skin biopsies among dermatologists and nondermatologist clinicians. J Am Acad Dermatol 82(1):110–116. https://doi.org/10.1016/j.jaad.2019.08.012
doi: 10.1016/j.jaad.2019.08.012
pubmed: 31408683
Rogers HW, Weinstock MA, Feldman SR, Coldiron BM (2015) Incidence estimate of nonmelanoma skin cancer (keratinocyte carcinomas) in the us population. JAMA Dermatol 151(10):1081–1086. https://doi.org/10.1001/jamadermatol.2015.1187
doi: 10.1001/jamadermatol.2015.1187
pubmed: 25928283
Schmults CD, Karia PS, Carter JB, Han J, Qureshi AA (2013) Factors predictive of recurrence and death from cutaneous squamous cell carcinoma: a 10-year, single-institution cohort study. JAMA Dermatol 149(5):541–547. https://doi.org/10.1001/jamadermatol.2013.2139
doi: 10.1001/jamadermatol.2013.2139
pubmed: 23677079
Simon AB, Vitzthum LK, Mell LK (2020) Challenge of directly comparing imaging-based diagnoses made by machine learning algorithms with those made by human clinicians. J Clinl Oncol 38(16):1868–1869. https://doi.org/10.1200/JCO.19.03350
doi: 10.1200/JCO.19.03350
Topol EJ (2019) High-performance medicine: the convergence of human and artificial intelligence. Nat Med 25(1):44–56
doi: 10.1038/s41591-018-0300-7
Tschandl P, Rosendahl C, Kittler H (2018) Data descriptor: the HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Sci Data 5:180161. https://doi.org/10.1038/sdata.2018.161
doi: 10.1038/sdata.2018.161
pubmed: 30106392
pmcid: 6091241
Tschandl P, Rosendahl C, Akay BN, Argenziano G, Blum A, Braun RP, Cabo H et al (2019) Expert-level diagnosis of nonpigmented skin cancer by combined convolutional neural networks. JAMA Dermatol 155(1):58–65. https://doi.org/10.1001/jamadermatol.2018.4378
doi: 10.1001/jamadermatol.2018.4378
pubmed: 30484822
Waldmann A, Nolte S, Geller AC, Katalinic A, Weinstock MA, Volkmer B, Greinert R, Breitbart EW (2012) Online first frequency of excisions and yields of malignant skin tumors in a population-based screening intervention of 360 288 whole-body examinations. Arch Dermatol 148(8):903–910. https://doi.org/10.1001/archdermatol.2012.893
doi: 10.1001/archdermatol.2012.893
pubmed: 22911184
Walker BN, Nees MA (2011) Theory of sonification. In: Hermann T, Hunt A and Neuhoff JG (Eds) The Sonification Handbook. Logos Publishing House, Berlin: 9–38. https://sonification.de/handbook/download/TheSonificationHandbook-chapter2.pdf .
Walker BN, Rehg JM, Kalra A, Winters RM, Drews P, Dascalu J, David EO, Dascalu A (2019) Dermoscopy diagnosis of cancerous lesions utilizing dual deep learning algorithms via visual and audio (sonification) outputs: laboratory and prospective observational studies. EBioMedicine 40(2):176–183. https://doi.org/10.1016/j.ebiom.2019.01.028
doi: 10.1016/j.ebiom.2019.01.028
pubmed: 30674442
pmcid: 6413349
Wehner MR, Dalma N, Landefeld C, Pare-Anastasiadou A, Koutelidas I, Chren MM, Aji N et al (2018) Natural history of lesions suspicious for basal cell carcinoma in older adults in Ikaria. Greece Br J Dermatol 179(3):767–768. https://doi.org/10.1111/bjd.16730
doi: 10.1111/bjd.16730
pubmed: 29741766