<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE root>
<article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:ali="http://www.niso.org/schemas/ali/1.0/" article-type="research-article" dtd-version="1.2" xml:lang="en"><front><journal-meta><journal-id journal-id-type="publisher-id">Russian Journal of Skin and Venereal Diseases</journal-id><journal-title-group><journal-title xml:lang="en">Russian Journal of Skin and Venereal Diseases</journal-title><trans-title-group xml:lang="ru"><trans-title>Российский журнал кожных и венерических болезней</trans-title></trans-title-group></journal-title-group><issn publication-format="print">1560-9588</issn><issn publication-format="electronic">2412-9097</issn><publisher><publisher-name xml:lang="en">Eco-Vector</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="publisher-id">642028</article-id><article-id pub-id-type="doi">10.17816/dv642028</article-id><article-categories><subj-group subj-group-type="toc-heading" xml:lang="en"><subject>DERMATOONCOLOGY</subject></subj-group><subj-group subj-group-type="toc-heading" xml:lang="ru"><subject>ДЕРМАТООНКОЛОГИЯ</subject></subj-group><subj-group subj-group-type="article-type"><subject>Research Article</subject></subj-group></article-categories><title-group><article-title xml:lang="en">Stages of training neural networks for classification and detection of skin neoplasms</article-title><trans-title-group xml:lang="ru"><trans-title>Этапы обучения нейросетей классификации и детекции новообразований кожи</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-1000-9848</contrib-id><contrib-id contrib-id-type="spin">1408-3490</contrib-id><name-alternatives><name xml:lang="en"><surname>Uskova</surname><given-names>Kseniia A.</given-names></name><name xml:lang="ru"><surname>Ускова</surname><given-names>Ксения Александровна</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><email>k_balyasova@bk.ru</email><xref ref-type="aff" rid="aff1"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-1473-6241</contrib-id><name-alternatives><name xml:lang="en"><surname>Dardyk</surname><given-names>Veniamin I.</given-names></name><name xml:lang="ru"><surname>Дардык</surname><given-names>Вениамин Иосифович</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><email>ben@aimedpro.ru</email><xref ref-type="aff" rid="aff2"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-7326-7553</contrib-id><contrib-id contrib-id-type="spin">6758-5913</contrib-id><name-alternatives><name xml:lang="en"><surname>Garanina</surname><given-names>Oxana E.</given-names></name><name xml:lang="ru"><surname>Гаранина</surname><given-names>Оксана Евгеньевна</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><bio xml:lang="en"><p>MD, Cand. Sci. (Medicine), Assistant Professor</p></bio><bio xml:lang="ru"><p>канд. мед. наук, доцент</p></bio><email>oksanachekalkina@yandex.ru</email><xref ref-type="aff" rid="aff1"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-1015-472X</contrib-id><contrib-id contrib-id-type="spin">3123-9969</contrib-id><name-alternatives><name xml:lang="en"><surname>Sinelnikov</surname><given-names>Igor E.</given-names></name><name xml:lang="ru"><surname>Синельников</surname><given-names>Игорь Евгеньевич</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><bio xml:lang="en"><p>MD, Cand. Sci. (Medicine)</p></bio><bio xml:lang="ru"><p>канд. мед. наук</p></bio><email>sinelnikov.igor@gmail.com</email><xref ref-type="aff" rid="aff3"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-0223-0753</contrib-id><contrib-id contrib-id-type="spin">9828-9522</contrib-id><name-alternatives><name xml:lang="en"><surname>Gamayunov</surname><given-names>Sergey V.</given-names></name><name xml:lang="ru"><surname>Гамаюнов</surname><given-names>Сергей Викторович</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><bio xml:lang="en"><p>MD, Dr. Sci. (Medicine)</p></bio><bio xml:lang="ru"><p>д-р мед. наук</p></bio><email>gamajnovs@mail.ru</email><xref ref-type="aff" rid="aff4"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0001-7150-5071</contrib-id><contrib-id contrib-id-type="spin">3691-8923</contrib-id><name-alternatives><name xml:lang="en"><surname>Samoylenko</surname><given-names>Igor V.</given-names></name><name xml:lang="ru"><surname>Самойленко</surname><given-names>Игорь Вячеславович</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><bio xml:lang="en"><p>MD, Cand. Sci. (Medicine)</p></bio><bio xml:lang="ru"><p>канд. мед. наук</p></bio><email>i.samoylenko@ronc.ru</email><xref ref-type="aff" rid="aff5"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-4482-1252</contrib-id><contrib-id contrib-id-type="spin">7623-7151</contrib-id><name-alternatives><name xml:lang="en"><surname>Luchinina</surname><given-names>Daria G.</given-names></name><name xml:lang="ru"><surname>Лучинина</surname><given-names>Дарья Григорьевна</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><email>luchininadg@mail.ru</email><xref ref-type="aff" rid="aff6"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-7535-3025</contrib-id><contrib-id contrib-id-type="spin">3431-7447</contrib-id><name-alternatives><name xml:lang="en"><surname>Mironycheva</surname><given-names>Anna M.</given-names></name><name xml:lang="ru"><surname>Миронычева</surname><given-names>Анна Михайловна</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><email>mironychevann@gmail.com</email><xref ref-type="aff" rid="aff1"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0004-9228-7770</contrib-id><contrib-id contrib-id-type="spin">3368-8554</contrib-id><name-alternatives><name xml:lang="en"><surname>Stepanova</surname><given-names>Yana L.</given-names></name><name xml:lang="ru"><surname>Степанова</surname><given-names>Яна Леонидовна</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><email>stepanova.ya09@yandex.ru</email><xref ref-type="aff" rid="aff1"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0003-1042-8425</contrib-id><contrib-id contrib-id-type="spin">8119-2480</contrib-id><name-alternatives><name xml:lang="en"><surname>Klemenova</surname><given-names>Irina A.</given-names></name><name xml:lang="ru"><surname>Клеменова</surname><given-names>Ирина Александровна</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><bio xml:lang="en"><p>MD, Dr. Sci. (Medicine), Professor</p></bio><bio xml:lang="ru"><p>д-р мед. наук, профессор</p></bio><email>iklemenova@mail.ru</email><xref ref-type="aff" rid="aff1"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0001-7253-7091</contrib-id><contrib-id contrib-id-type="spin">8301-4815</contrib-id><name-alternatives><name xml:lang="en"><surname>Shlivko</surname><given-names>Irena L.</given-names></name><name xml:lang="ru"><surname>Шливко</surname><given-names>Ирена Леонидовна</given-names></name></name-alternatives><address><country country="RU">Russian Federation</country></address><bio xml:lang="en"><p>MD, Dr. Sci. (Medicine), Assistant Professor</p></bio><bio xml:lang="ru"><p>д-р мед. наук, доцент</p></bio><email>irshlivko@gmail.com</email><xref ref-type="aff" rid="aff1"/></contrib></contrib-group><aff-alternatives id="aff1"><aff><institution xml:lang="en">Privolzhsky Research Medical University</institution></aff><aff><institution xml:lang="ru">Приволжский исследовательский медицинский университет</institution></aff></aff-alternatives><aff-alternatives id="aff2"><aff><institution xml:lang="en">AIMED Limited liability company</institution></aff><aff><institution xml:lang="ru">ООО «АИМЕД»</institution></aff></aff-alternatives><aff-alternatives id="aff3"><aff><institution xml:lang="en">Melanoma Unit Limited liability company</institution></aff><aff><institution xml:lang="ru">ООО «Меланома Юнит»</institution></aff></aff-alternatives><aff-alternatives id="aff4"><aff><institution xml:lang="en">Nizhny Novgorod Regional Clinical Oncological Dispensary</institution></aff><aff><institution xml:lang="ru">Нижегородский областной клинический онкологический диспансер</institution></aff></aff-alternatives><aff-alternatives id="aff5"><aff><institution xml:lang="en">N.N. Blokhin National Medical Research Center of Oncology</institution></aff><aff><institution xml:lang="ru">Национальный медицинский исследовательский центр онкологии имени Н.Н. Блохина</institution></aff></aff-alternatives><aff-alternatives id="aff6"><aff><institution xml:lang="en">Republican Dermatovenerologic Dispensary</institution></aff><aff><institution xml:lang="ru">Республиканский кожно-венерологический диспансер</institution></aff></aff-alternatives><pub-date date-type="preprint" iso-8601-date="2025-03-30" publication-format="electronic"><day>30</day><month>03</month><year>2025</year></pub-date><pub-date date-type="pub" iso-8601-date="2025-02-06" publication-format="electronic"><day>06</day><month>02</month><year>2025</year></pub-date><volume>28</volume><issue>1</issue><issue-title xml:lang="en"/><issue-title xml:lang="ru"/><fpage>5</fpage><lpage>15</lpage><history><date date-type="received" iso-8601-date="2024-11-19"><day>19</day><month>11</month><year>2024</year></date><date date-type="accepted" iso-8601-date="2025-01-31"><day>31</day><month>01</month><year>2025</year></date></history><permissions><copyright-statement xml:lang="en">Copyright ©; 2025, Eco-Vector</copyright-statement><copyright-statement xml:lang="ru">Copyright ©; 2025, Эко-Вектор</copyright-statement><copyright-year>2025</copyright-year><copyright-holder xml:lang="en">Eco-Vector</copyright-holder><copyright-holder xml:lang="ru">Эко-Вектор</copyright-holder><ali:free_to_read xmlns:ali="http://www.niso.org/schemas/ali/1.0/" start_date="2028-01-01"/><license><ali:license_ref xmlns:ali="http://www.niso.org/schemas/ali/1.0/">https://eco-vector.com/for_authors.php#07</ali:license_ref></license></permissions><self-uri xlink:href="https://rjsvd.com/1560-9588/article/view/642028">https://rjsvd.com/1560-9588/article/view/642028</self-uri><abstract xml:lang="en"><p><bold>BACKGROUND:</bold> In recent years, neural networks have become an integral part of many fields, including medicine. However, the effectiveness of these models directly depends on the quality of the training data on which they are trained. Creating and maintaining a high-quality training dataset is a critical step in the development process of neural networks.</p> <p><bold>AIM:</bold><italic> </italic>The aim of the research is to identify the key characteristics of the training database for the neural network that influence its subsequent sensitivity and specificity.</p> <p><bold>MATERIALS AND METHODS:</bold> A database of verified images of skin neoplasms was created to train a neural network to implement it in large-scale screening examinations. In the first phase of the study, a database was created to train a neural network to classify images of skin neoplasms (NSCa). Between 2017 and 2019, 7,680 digital images were collected from 6,892 patients with verified diagnoses: 5,316 (69,22%) confirmed by pathological examination, and 2,364 (30,78%)) confirmed clinically and dermatoscopically. A dataset containing 7,680 verified clinical images of skin neoplasms was created, and 1,680 images constituted the test sample for analyzing the model's effectiveness. The performance indicators of NSCa were as follows: sensitivity (Se): 70.47%; specificity (Sp): 79.86%; diagnostic accuracy (Ac): 74.68%. Due to the low sensitivity and specificity rates, the following steps were taken: (1) an additional round of training was conducted; (2) image quality control methods were developed; (3) a detection neural network was created, and (4) a new neural (NSCb) was established.</p> <p><bold>RESULTS:</bold> The neural network, trained on a verified dataset of clinical images of benign and malignant skin neoplasms and having undergone multiple rounds of training, operates with a sensitivity of 85.32–86.97% and a specificity of 87.59–88.92%. These rates exceed the sensitivity and specificity of skin neoplasm diagnoses made by non-oncological specialists using the naked eye, allowing for the use of this method in population screening. Following the retraining of the neural network and the establishment of NSCb, the creation of neural network, and the development of image quality control methods, an increase in the sensitivity and specificity of the neural network's performance was observed.</p> <p><bold>CONCLUSION:</bold> The use of artificial intelligence as a physician's assistant imposes quite high requirements on the performance parameters of the neural network. Mechanical learning, even on a large volume of verified data, did not achieve the desired results. The sequential work aimed at improving the parameters involved conducting an additional round of training, developing image quality control methods, and creating a detection neural network and a classification neural network. As a result, the trained neural network operates with a sensitivity of 85.32% to 86.97% and a specificity of 87.59% to 88.92%, which has enabled the use of the trained neural network as a tool for population screening.</p></abstract><trans-abstract xml:lang="ru"><p><bold>Обоснование.</bold> В последние годы нейросети стали неотъемлемой частью многих областей, включая медицину, однако эффективность этих моделей напрямую зависит от качества базы, на которой они обучаются. Создание и поддержание качественной обучающей базы является критически важным этапом в процессе разработки нейросетей.</p> <p><bold>Цель исследования</bold> ― определить этапы обучения нейросетей для достижения высокой чувствительности и специфичности их работы в области распознавания новообразований кожи.</p> <p><bold>Материалы и методы.</bold> На первом этапе исследования сформирована база данных фотографических изображений новообразований кожи для первичного обучения нейросети классификации изображений. С этой целью в период с 2017 по 2019 год собрано 7680 цифровых изображений от 6892 пациентов с диагнозами, верифицированными с помощью патоморфологического исследования в 5316 (69,22%) случаях, с помощью дерматоскопического и клинического исследования ― в 2364 (30,78%). Создан массив данных (датасет) из 7680 верифицированных клинических изображений новообразований кожи, из них 1680 изображений составили тестовую выборку для анализа эффективности модели. Показатели эффективности нейросети классификации «а» (НСКа): чувствительность (Se) 70,47%; специфичность (Sp) 79,86%; диагностическая точность (Ac) 74,68%. В связи с низкими показателями чувствительности и специфичности был проведён дополнительный раунд обучения; разработаны методы контроля качества изображения; созданы нейросеть детекции и нейросеть классификации «б» (НСКб) после второго раунда обучения нейросети.</p> <p><bold>Результаты.</bold> Нейросеть, обученная на верифицированном наборе клинических изображений доброкачественных и злокачественных новообразований кожи и прошедшая раунды обучения, работает с чувствительностью 85,32–86,97% и специфичностью 87,59–88,92%, что превышает чувствительность и специфичность диагностики новообразований кожи врачами неонкологических специальностей при обследовании невооружённым глазом и позволяет предложить данный инструмент для проведения популяционного скрининга. После дообучения нейросети, создания НСКб и нейросети детекции, разработки методов контроля качества изображения наблюдался рост показателей чувствительности и специфичности работы нейросети.</p> <p><bold>Заключение.</bold> Использование искусственного интеллекта в качестве помощника врача предъявляет достаточно высокие требования к параметрам работы нейросети. Механическое обучение даже на большом объёме верифицированных данных не позволило достичь желаемых результатов. Для улучшения параметров выполнен дополнительный раунд обучения, разработаны методы контроля качества изображения, созданы нейросети детекции и второй НСКб. В результате обученная нейросеть имеет чувствительность 85,32–86,97% и специфичность 87,59–88,92%, что позволяет предложить её в качестве инструмента для популяционного скрининга.</p></trans-abstract><kwd-group xml:lang="en"><kwd>neural network</kwd><kwd>artificial intelligence</kwd><kwd>dataset</kwd><kwd>skin neoplasms</kwd><kwd>skin tumors</kwd></kwd-group><kwd-group xml:lang="ru"><kwd>нейросеть</kwd><kwd>искусственный интеллект</kwd><kwd>датасет</kwd><kwd>новообразования кожи</kwd><kwd>опухоли кожи</kwd></kwd-group><funding-group><award-group><funding-source><institution-wrap><institution xml:lang="ru">Министерство науки и высшего образования РФ</institution></institution-wrap><institution-wrap><institution xml:lang="en">Ministry of Science and Higher Education of the Russian Federation</institution></institution-wrap></funding-source></award-group></funding-group></article-meta></front><body></body><back><ref-list><ref id="B1"><label>1.</label><mixed-citation>Abbasov IB, Deshmukh RR. Application of artificial intelligence for medical imaging. Mezhdunarodnyi nauchno-issledovatel’skii zhurnal. 2021;(12-1):43–49. doi: 10.23670/IRJ.2021.114.12.005 EDN: QGKKCU</mixed-citation></ref><ref id="B2"><label>2.</label><mixed-citation>De A, Sarda A, Gupta S, Das S. Use of artificial intelligence in dermatology. Indian J Dermatol. 2020;65(5):352–357. doi: 10.4103/ijd.IJD_418_20</mixed-citation></ref><ref id="B3"><label>3.</label><mixed-citation>Uskova KA, Garanina OE, Ukharov AO, et al. Artificial intelligence as a tool for population screening of skin tumors. Effective pharmacotherapy. 2024;20(1):62–71. doi: 10.33978/2307-3586-2024-20-1-62-71 EDN: RIDZDV</mixed-citation></ref><ref id="B4"><label>4.</label><mixed-citation>Harskamp RE, de Vijlder HC, Bekkenk MW. [Smartphone apps for self-diagnosis of skin cancer. (In Dutch)]. Ned Tijdschr Geneeskd. 2022;166:D5986.</mixed-citation></ref><ref id="B5"><label>5.</label><mixed-citation>Li Z, Koban KC, Schenck TL, et al. Artificial intelligence in dermatology image analysis: Current developments and future trends. J Clin Med. 2022;11(22):6826. doi: 10.3390/jcm11226826 EDN: PFUTYM</mixed-citation></ref><ref id="B6"><label>6.</label><mixed-citation>Caffery LJ, Janda M, Miller R, et al. Informing a position statement on the use of artificial intelligence in dermatology in Australia. Australas J Dermatol. 2023;64(1):e11–e20. doi: 10.1111/ajd.13946 EDN: TXYOWT</mixed-citation></ref><ref id="B7"><label>7.</label><mixed-citation>Meshcheryakova AM, Akopyan EA, Slinin AS. Artificial intelligence in medical imaging. Main objectives and development scenarios. Russian journal of telemedicine and E-Health. 2018;(3):98–102. EDN: YWZVLN</mixed-citation></ref><ref id="B8"><label>8.</label><mixed-citation>Erkal EY, Akpınar A, Erkal HŞ. Ethical evaluation of artificial intelligence applications in radiotherapy using the Four Topics Approach. Artif Intell Med. 2021;115:102055. doi: 10.1016/j.artmed.2021.102055 EDN: VCFHPD</mixed-citation></ref><ref id="B9"><label>9.</label><mixed-citation>Nogales A, García-Tejedor ÁJ, Monge D, et al. A survey of deep learning models in medical therapeutic areas. Artif Intell Med. 2021;112:102020. doi: 10.1016/j.artmed.2021.102020 EDN: PWJSKR</mixed-citation></ref><ref id="B10"><label>10.</label><mixed-citation>Pai VV, Pai RB. Artificial intelligence in dermatology and healthcare: An overview. Indian J Dermatol Venereol Leprol. 2021;87(4):457–467. doi: 10.25259/IJDVL_518_19 EDN: QTHEZX</mixed-citation></ref><ref id="B11"><label>11.</label><mixed-citation>Muthukrishnan N, Maleki F, Ovens K, et al. Brief history of artificial intelligence. Neuroimaging Clin N Am. 2020;30(4):393–399. doi: 10.1016/j.nic.2020.07.004 EDN: UWNULY</mixed-citation></ref><ref id="B12"><label>12.</label><mixed-citation>Meldo AA, Utkin LV, Trofimova TN, et al. Novel approaches to development of artificial intelligence algorithms in the lung cancer diagnostics. Diagnostic radiology and radiotherapy. 2019;(1):8–18. doi: 10.22328/2079-5343-2019-10-1-8-18 EDN: ZHUVML</mixed-citation></ref><ref id="B13"><label>13.</label><mixed-citation>El-Azhary RA. The inevitability of change. Clin Dermatol. 2019;37(1):4–11. doi: 10.1016/j.clindermatol.2018.09.003</mixed-citation></ref><ref id="B14"><label>14.</label><mixed-citation>Database RU 2021620654/09.02.2021. Application number 2021620162. Burdakov AV, Ukharov AO, Dardyk VI, Shlivko IL. Database of images and results of diagnostics of skin neoplasms. Available from: https://www.elibrary.ru/item.asp?id=45805269. Accessed: Jan 15, 2025. (In Russ.) EDN: OFLCNP</mixed-citation></ref><ref id="B15"><label>15.</label><mixed-citation>Piccolo D, Ferrari A, Peris K, et al. Dermoscopic diagnosis by a trained clinician vs. a clinician with minimal dermoscopy training vs. computer-aided diagnosis of 341 pigmented skin lesions: A comparative study. Br J Dermatol. 2002;147(3):481–486. doi: 10.1046/j.1365-2133.2002.04978.x</mixed-citation></ref></ref-list></back></article>
