Stages of training neural networks for classification and detection of skin neoplasms

Cover Page


Cite item

Full Text

Open Access Open Access
Restricted Access Access granted
Restricted Access Subscription or Fee Access

Abstract

BACKGROUND: 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.

AIM: 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.

MATERIALS AND METHODS: 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.

RESULTS: 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.

CONCLUSION: 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.

Full Text

Restricted Access

About the authors

Kseniia A. Uskova

Privolzhsky Research Medical University

Author for correspondence.
Email: k_balyasova@bk.ru
ORCID iD: 0000-0002-1000-9848
SPIN-code: 1408-3490
Russian Federation, Nizhny Novgorod

Veniamin I. Dardyk

AIMED Limited liability company

Email: ben@aimedpro.ru
ORCID iD: 0000-0002-1473-6241
Russian Federation, Moscow

Oxana E. Garanina

Privolzhsky Research Medical University

Email: oksanachekalkina@yandex.ru
ORCID iD: 0000-0002-7326-7553
SPIN-code: 6758-5913

MD, Cand. Sci. (Medicine), Assistant Professor

Russian Federation, Nizhny Novgorod

Igor E. Sinelnikov

Melanoma Unit Limited liability company

Email: sinelnikov.igor@gmail.com
ORCID iD: 0000-0002-1015-472X
SPIN-code: 3123-9969

MD, Cand. Sci. (Medicine)

Russian Federation, Moscow

Sergey V. Gamayunov

Nizhny Novgorod Regional Clinical Oncological Dispensary

Email: gamajnovs@mail.ru
ORCID iD: 0000-0002-0223-0753
SPIN-code: 9828-9522

MD, Dr. Sci. (Medicine)

Russian Federation, Nizhny Novgorod

Igor V. Samoylenko

N.N. Blokhin National Medical Research Center of Oncology

Email: i.samoylenko@ronc.ru
ORCID iD: 0000-0001-7150-5071
SPIN-code: 3691-8923

MD, Cand. Sci. (Medicine)

Russian Federation, Moscow

Daria G. Luchinina

Republican Dermatovenerologic Dispensary

Email: luchininadg@mail.ru
ORCID iD: 0000-0002-4482-1252
SPIN-code: 7623-7151
Russian Federation, Yoshkar-Ola

Anna M. Mironycheva

Privolzhsky Research Medical University

Email: mironychevann@gmail.com
ORCID iD: 0000-0002-7535-3025
SPIN-code: 3431-7447
Russian Federation, Nizhny Novgorod

Yana L. Stepanova

Privolzhsky Research Medical University

Email: stepanova.ya09@yandex.ru
ORCID iD: 0009-0004-9228-7770
SPIN-code: 3368-8554
Russian Federation, Nizhny Novgorod

Irina A. Klemenova

Privolzhsky Research Medical University

Email: iklemenova@mail.ru
ORCID iD: 0000-0003-1042-8425
SPIN-code: 8119-2480

MD, Dr. Sci. (Medicine), Professor

Russian Federation, Nizhny Novgorod

Irena L. Shlivko

Privolzhsky Research Medical University

Email: irshlivko@gmail.com
ORCID iD: 0000-0001-7253-7091
SPIN-code: 8301-4815

MD, Dr. Sci. (Medicine), Assistant Professor

Russian Federation, Nizhny Novgorod

References

  1. 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
  2. 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
  3. 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
  4. Harskamp RE, de Vijlder HC, Bekkenk MW. [Smartphone apps for self-diagnosis of skin cancer. (In Dutch)]. Ned Tijdschr Geneeskd. 2022;166:D5986.
  5. 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
  6. 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
  7. 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
  8. 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
  9. 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
  10. 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
  11. 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
  12. 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
  13. El-Azhary RA. The inevitability of change. Clin Dermatol. 2019;37(1):4–11. doi: 10.1016/j.clindermatol.2018.09.003
  14. 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
  15. 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

Supplementary files

Supplementary Files
Action
1. JATS XML
2. Fig. 1. Study design. Stages of training and analysis of the performance of neural networks for classification and detection of skin neoplasms. НС ― neural network. Source: Uskova K.A. et al., 2025.

Download (912KB)
3. Fig. 2. Classification of the training dataset. Source: Uskova K.A. et al., 2025.

Download (270KB)
4. Fig. 3. Examples of clinical images of basal cell skin cancer, melanoma, nevi, and seborrheic keratosis. Source: Uskova K.A. et al., 2025.

Download (532KB)
5. Fig. 4. Example of the neural network’s detection performance when more than one neoplasm is present in the image. Source: Uskova K.A. et al., 2025.

Download (386KB)

Copyright (c) 2025 Eco-Vector

License URL: https://eco-vector.com/for_authors.php#07

СМИ зарегистрировано Федеральной службой по надзору в сфере связи, информационных технологий и массовых коммуникаций (Роскомнадзор).
Регистрационный номер и дата принятия решения о регистрации СМИ: серия ПИ № ФС 77 - 86501 от 11.12.2023 г
СМИ зарегистрировано Федеральной службой по надзору в сфере связи, информационных технологий и массовых коммуникаций (Роскомнадзор).
Регистрационный номер и дата принятия решения о регистрации СМИ: серия ЭЛ № ФС 77 - 80653 от 15.03.2021 г
.