E.O. Bryanskaya1, D.V. Gerasin2, A.V. Bakotina3, A.Yu. Ovchinnikov4, Yu.O. Nikolaeva5, V.V. Dremin6, A.V. Dunaev7

1, 2, 6, 7 Research and Development Center of Biomedical Photonics, Orel State University (Orel, Russia)
3–5 Russian University of Medicine of the Ministry of Health of the Russian Federation (Moscow, Russia)
1bryanskayae@mail.ru

The trend towards increasing the availability of early diagnosis of ENT pathology, as well as the accuracy of diagnosis, has created the prerequisites for the development of optical methods, among which is the method of digital diaphanoscopy. Machine learning methods, in particular, convolutional neural networks, represent a prospect for use in digital diaphanoscopy to significantly reduce the subjectivity of diagnosing maxillary sinus pathologies.

The purpose of the work is the development of a classification model based on the ResNet-50 convolutional neural network for differential diagnosis of HPV in digital diaphanoscopy into classes "absence of pathology", "sinusitis", "cystic change", indicating the side of the pathology "left-sided", "right-sided", "two-sided", with improved accuracy compared to previously proposed approaches.

The scientific substantiation of the possibility of using the convolutional neural network ResNet-50 for the differentiation of maxillary sinus pathologies has been carried out. A classification model has been developed that makes it possible to differentiate the conditions of the maxillary sinuses into classes "sinusitis" with sensitivity 0.9, specificity 0.95, and accuracy 0.88; "cystic change" with sensitivity 0.86, specificity 0.96, and accuracy 0.88.

The method proposed and described in the paper is the basis for creating a system to support medical decision-making in digital diaphanoscopy for the purpose of early detection of maxillary sinus pathologies, in particular in the framework of public health screening by the incidence class of ENT organs, and in telemedicine.

Bryanskaya E.O., Gerasin D.V., Bakotina A.V., Ovchinnikov A.Yu., Nikolaeva Yu.O., Dremin V.V., Dunaev A.V. System for differential diagnosis of maxillary sinus pathologies based on digital diaphanoscopy and convolutional neural networks. Biomedicine Radioengineering. 2026. V. 29. № 3. P. 18–22. DOI: https:// doi.org/10.18127/ j15604136-202603-03 (In Russian)

1. Pluzhnikov M.C., Ivanov B.S., Usanov A.A. Lazernaya diafanoskopiya pri vospalitel'nyh zabolevaniyah pridatochnyh pazuh nosa. Vestnik otorinolaringologii. 1991. T. 4. S.22 (In Russian).
2. Feldmann H. Die Geschichte der Diaphanoskopie. Laryngo-Rhino-Otologie. 1998. V. 77. № 5. P. 297–304.
3. Bryanskaya E.O., Dremin V.V., Shupletsov V.V., Kornaev A.V., Kirillin M.Yu., Bakotina A.V., Panchenkov D.N., Podmasteryev K.V., Artyushenko V.G., Dunaev A.V. Digital diaphanoscopy of maxillary sinus pathologies supported by machine learning. Journal of Biophotonics. 2023. V. 16. № 9. P. e202300138.
4. Bryanskaya E.O., Dunaev A.V. Metod i ustrojstvo cifrovoj diafanoskopii dlya diagnostiki patologij verhnechelyustnyh pazuh. Medicinskaya tekhnika. 2023. T. 339. №3. S. 5–7 (In Russian).
5. Chainansamit S., Chit-uea-ophat C., Reechaipichitkul W., Piromchai P. The Diagnostic Value of Traditional Nasal Examination Tools in an Endoscopic Era. Ear, Nose Throat J. 2021. V. 100. № 3. P. 167–171.
6. Jacques S.L. Optical properties of biological tissues: A review. Phys. Med. Biol. 2013. V. 58. № 11. P. R37–R61.
7. Valkov A., Nokolov G., Duhlenski B., Stoyanov T., Mladenov T., Yildiz M., Atanasova K., Mirchev S., Valcheva K. The application of ultrasound examination in the treatment of Acute Sinusitis. Comparing X-ray to ultrasound of paranasal sinuses. Int. Bull. Otorhinolaryngol. 2021. V. 17. № 1. P. 40–41.
8. Bryanskaya E.O., Novikova I.N., Dremin V.V., Gneushev R.Y., Bibikova O.A., Dunaev A.V., Artyushenko V.G. Optical Diagnostics of the Maxillary Sinuses by Digital Diaphanoscopy Technology. Diagnostics. 2021. V. 77. № 11. P.1–13.
9. Wu Q., Wang X., Liang G., Luo X., Zhou M., Deng H., Zhang Y., Huang X., Yang Q. Advances in Image-Based Artificial Intelligence in Otorhinolaryngology-Head and Neck Surgery: A Systematic Review. Otolaryngol. Head Neck Surg. 2023. V. 169. № 5. P. 1132–1142.
10. Gerasin D.V., Bryanskaya E.O., Dremin V.V., Dunaev A.V. The Use of Convolutional Neural Networks to Classify the States of the Maxillary Sinuses in Digital Diaphanoscopy. 2024 Intelligent Technologies and Electronic Devices in Vehicle and Road Transport Complex (TIRVED). Moscow, Russia. 2024. November 13–15. P. 1–4.