{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T20:41:45Z","timestamp":1792010505088,"version":"4.4.0"},"reference-count":26,"publisher":"Oxford University Press (OUP)","issue":"7","license":[{"start":{"date-parts":[[2023,3,27]],"date-time":"2023-03-27T00:00:00Z","timestamp":1679875200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/pages\/standard-publication-reuse-rights"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2023,7,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Background<\/jats:title>\n                  <jats:p>The distinction between in situ melanoma (MIS) and invasive melanoma is challenging even for expert dermatologists. The use of pretrained convolutional neural networks (CNNs) as ancillary decision systems needs further research.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Aim<\/jats:title>\n                  <jats:p>To develop, validate and compare three deep transfer learning (DTL) algorithms to predict MIS vs. invasive melanoma and melanoma with a Breslow thickness (BT) of &amp;lt; 0.8\u2005mm vs. \u2265 0.8\u2005mm.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>A dataset of 1315 dermoscopic images of histopathologically confirmed melanomas was created from Virgen del Rocio University Hospital and open repositories of the International Skin Imaging Collaboration archive and Polesie\u00a0S et al.\u00a0(Dermatol Pract Concept 2021; 11:e2021079). The images were labelled as MIS or invasive melanoma and &amp;lt; 0.8\u2005mm\u00a0or \u2265 0.8\u2005mm of BT. We conducted three trainings, and overall means for receiver operating characteristic (ROC) curves, sensitivity, specificity, positive and negative predictive value, and balanced diagnostic accuracy outcomes were evaluated on the test set with ResNetV2, EfficientNetB6 and InceptionV3. The results of 10 dermatologists were compared with the algorithms. Grad-CAM gradient maps were generated, highlighting relevant areas considered by the CNNs within the images.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>EfficientNetB6 achieved the highest diagnostic accuracy for the comparison between MIS vs. invasive melanoma (61%) and BT &amp;lt; 0.8\u2005mm vs. \u2265 0.8\u2005mm (75%). For the BT comparison, ResNetV2 with an area under the ROC curve of 0.76 and InceptionV3 with an area under the ROC curve of 0.75, outperformed the results obtained by the dermatologist group with an area under the ROC curve of 0.70.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>EfficientNetB6 recorded the best prediction results, outperforming the dermatologists for the comparison of 0.8\u2005mm of BT. DTL could be an ancillary aid to support dermatologists\u2019 decisions in the near future.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/ced\/llad107","type":"journal-article","created":{"date-parts":[[2023,3,27]],"date-time":"2023-03-27T08:31:58Z","timestamp":1679905918000},"page":"752-758","source":"Crossref","is-referenced-by-count":13,"title":["Prediction of melanoma Breslow thickness using deep transfer learning algorithms"],"prefix":"10.1093","volume":"48","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-2525-4069","authenticated-orcid":false,"given":"Juan-Carlos","family":"Hern\u00e1ndez-Rodr\u00edguez","sequence":"first","affiliation":[{"name":"Department of Dermatology, Virgen del Rocio University Hospital , Seville , Spain"},{"name":"Department of Medicine, Faculty of Medicine"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5849-8003","authenticated-orcid":false,"given":"Lourdes","family":"Dur\u00e1n-L\u00f3pez","sequence":"additional","affiliation":[{"name":"Robotics and Technology of Computers Laboratory, University of Seville , Seville , Spain"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5474-107X","authenticated-orcid":false,"given":"Juan P","family":"Dom\u00ednguez-Morales","sequence":"additional","affiliation":[{"name":"Robotics and Technology of Computers Laboratory, University of Seville , Seville , Spain"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2134-3902","authenticated-orcid":false,"given":"Juan","family":"Ortiz-\u00c1lvarez","sequence":"additional","affiliation":[{"name":"Department of Dermatology, Virgen del Rocio University Hospital , Seville , Spain"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9108-9538","authenticated-orcid":false,"given":"Juli\u00e1n","family":"Conejo-Mir","sequence":"additional","affiliation":[{"name":"Department of Dermatology, Virgen del Rocio University Hospital , Seville , Spain"},{"name":"Department of Medicine, Faculty of Medicine"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6843-5877","authenticated-orcid":false,"given":"Jose-Juan","family":"Pereyra-Rodriguez","sequence":"additional","affiliation":[{"name":"Department of Dermatology, Virgen del Rocio University Hospital , Seville , Spain"},{"name":"Department of Medicine, Faculty of Medicine"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2023,3,27]]},"reference":[{"key":"2023101015310033700_llad107-B1","doi-asserted-by":"crossref","first-page":"236","DOI":"10.1016\/j.ejca.2022.03.008","article-title":"European consensus-based interdisciplinary guideline for melanoma. Part 1: diagnostics: update 2022","volume":"170","author":"Garbe","year":"2022","journal-title":"Eur J Cancer"},{"key":"2023101015310033700_llad107-B2","doi-asserted-by":"crossref","first-page":"902","DOI":"10.1097\/00000658-197011000-00017","article-title":"Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma","volume":"172","author":"Breslow","year":"1970","journal-title":"Ann Surg"},{"key":"2023101015310033700_llad107-B3","doi-asserted-by":"crossref","first-page":"635","DOI":"10.1016\/j.clindermatol.2021.03.009","article-title":"Dermatoscopy","volume":"39","author":"Ring","year":"2021","journal-title":"Clin Dermatol"},{"key":"2023101015310033700_llad107-B4","article-title":"Can dermoscopy be used to predict if a melanoma is in situ or invasive?","volume":"11","author":"Polesie","year":"2021","journal-title":"Dermatol Pract Concept"},{"issue":"S1","key":"2023101015310033700_llad107-B5","doi-asserted-by":"crossref","first-page":"4","DOI":"10.1684\/ejd.2019.3538","article-title":"Computational neural network in melanocytic lesions diagnosis: artificial intelligence to improve diagnosis in dermatology?","volume":"29","author":"Aractingi","year":"2019","journal-title":"Eur J Dermatol"},{"key":"2023101015310033700_llad107-B6","doi-asserted-by":"crossref","first-page":"420","DOI":"10.1007\/s42979-021-00815-1","article-title":"Deep learning: a comprehensive overview on techniques, taxonomy, applications and research directions","volume":"2","author":"Sarker","year":"2021","journal-title":"SN Comput Sci"},{"key":"2023101015310033700_llad107-B7","doi-asserted-by":"crossref","DOI":"10.3389\/fmed.2021.723914","article-title":"Discrimination between invasive and in situ melanomas using clinical close-up images and a de novo convolutional neural network","volume":"8","author":"Polesie","year":"2021","journal-title":"Front Med"},{"key":"2023101015310033700_llad107-B8","doi-asserted-by":"crossref","first-page":"647","DOI":"10.1016\/j.jaad.2021.02.012","article-title":"Discrimination between invasive and in situ melanomas using a convolutional neural network","volume":"86","author":"Gillstedt","year":"2022","journal-title":"J Am Acad Dermatol"},{"key":"2023101015310033700_llad107-B9","doi-asserted-by":"crossref","DOI":"10.2340\/actadv.v102.2681","article-title":"Evaluation of melanoma thickness with clinical close-up and dermoscopic images using a convolutional neural network","volume":"102","author":"Gillstedt","year":"2022","journal-title":"Acta Derm Venereol"},{"key":"2023101015310033700_llad107-B10","doi-asserted-by":"crossref","first-page":"2002","DOI":"10.1111\/jdv.18436","article-title":"Assessment of melanoma thickness based on dermoscopy images: an open, web-based, international, diagnostic study","volume":"36","author":"Polesie","year":"2022","journal-title":"J Eur Acad Dermatol Venereol"},{"key":"2023101015310033700_llad107-B11","doi-asserted-by":"crossref","first-page":"2268","DOI":"10.1016\/j.jid.2021.12.033","article-title":"Deep learning algorithms for predicting Breslow thickness from dermoscopic images of acral lentiginous melanomas","volume":"142","author":"Chu","year":"2022","journal-title":"J Invest Dermatol"},{"key":"2023101015310033700_llad107-B12","doi-asserted-by":"crossref","first-page":"202","DOI":"10.1016\/j.ejca.2021.06.049","article-title":"Skin cancer classification via convolutional neural networks: systematic review of studies involving human experts","volume":"156","author":"Haggenm\u00fcller","year":"2021","journal-title":"Eur J Cancer"},{"key":"2023101015310033700_llad107-B13","article-title":"Can dermoscopy be used to predict if a melanoma is in situ or invasive?","volume":"11","author":"Polesie","year":"2021","journal-title":"Dermatol Pract Concept"},{"key":"2023101015310033700_llad107-B14","author":"International Skin Imaging Collaboration"},{"key":"2023101015310033700_llad107-B15","doi-asserted-by":"crossref","first-page":"256","DOI":"10.1016\/j.ejca.2022.04.018","article-title":"European consensus-based interdisciplinary guideline for melanoma. Part 2: treatment\u00a0\u2013 update 2022","volume":"170","author":"Garbe","year":"2022","journal-title":"Eur J Cancer"},{"key":"2023101015310033700_llad107-B16","article-title":"R-Drop: regularized dropout for neural networks","author":"Liang","year":"2021","journal-title":"ArXiv"},{"key":"2023101015310033700_llad107-B17","author":"Chollet"},{"key":"2023101015310033700_llad107-B18","doi-asserted-by":"crossref","first-page":"336","DOI":"10.1007\/s11263-019-01228-7","article-title":"Visual explanations from deep networks via gradient-based localization","volume":"128","author":"Selvaraju","year":"2020","journal-title":"Int J Comput Vis"},{"key":"2023101015310033700_llad107-B19","first-page":"55","article-title":"An R package for non-standard ROC curve analysis","volume-title":"R J","author":"P\u00e9rez-Fern\u00e1ndez","year":"2019"},{"key":"2023101015310033700_llad107-B20","doi-asserted-by":"crossref","first-page":"378","DOI":"10.1037\/h0031619","article-title":"Measuring nominal scale agreement among many raters","volume":"76","author":"Fleiss","year":"1971","journal-title":"Psychol Bull"},{"key":"2023101015310033700_llad107-B21","doi-asserted-by":"crossref","first-page":"414","DOI":"10.1001\/jamadermatol.2017.6447","article-title":"Accuracy of dermoscopic criteria for the diagnosis of melanoma in situ","volume":"154","author":"Lallas","year":"2018","journal-title":"JAMA Dermatol"},{"key":"2023101015310033700_llad107-B22","doi-asserted-by":"crossref","first-page":"529","DOI":"10.1111\/ced.14950","article-title":"\u2018Rainbow pattern\u2019: a dermoscopic sign of invasive melanoma","volume":"47","author":"Rodr\u00edguez-Lomba","year":"2022","journal-title":"Clin Exp Dermatol"},{"key":"2023101015310033700_llad107-B23","doi-asserted-by":"crossref","first-page":"1345","DOI":"10.1109\/TKDE.2009.191","article-title":"A survey on transfer learning","volume":"22","author":"Pan","year":"2010","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"2023101015310033700_llad107-B24","doi-asserted-by":"crossref","first-page":"542","DOI":"10.1111\/ced.14969","article-title":"Perceptions of the use of artificial intelligence in the diagnosis of skin cancer: an outpatient survey","volume":"47","author":"Lim","year":"2022","journal-title":"Clin Exp Dermatol"},{"key":"2023101015310033700_llad107-B25","doi-asserted-by":"crossref","first-page":"568","DOI":"10.1007\/s10278-017-0045-8","article-title":"Transforming dermatologic imaging for the digital era: metadata and standards","volume":"31","author":"Caffery","year":"2018","journal-title":"J Digit Imaging"},{"key":"2023101015310033700_llad107-B26","doi-asserted-by":"crossref","first-page":"1333","DOI":"10.1111\/ced.14726","article-title":"Unintentional consequences of artificial intelligence in dermatology for patients with skin of colour","volume":"46","author":"Butt","year":"2021","journal-title":"Clin Exp Dermatol"}],"container-title":["Clinical and Experimental Dermatology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/ced\/advance-article-pdf\/doi\/10.1093\/ced\/llad107\/50126801\/llad107.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/ced\/article-pdf\/48\/7\/752\/51988953\/llad107.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/ced\/article-pdf\/48\/7\/752\/51988953\/llad107.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,10,10]],"date-time":"2023-10-10T15:35:09Z","timestamp":1696952109000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/ced\/article\/48\/7\/752\/7086611"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,3,27]]},"references-count":26,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2023,3,27]]},"published-print":{"date-parts":[[2023,7,7]]}},"URL":"https:\/\/doi.org\/10.1093\/ced\/llad107","relation":{},"ISSN":["0307-6938","1365-2230"],"issn-type":[{"value":"0307-6938","type":"print"},{"value":"1365-2230","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2023,7]]},"published":{"date-parts":[[2023,3,27]]}}}