AI-Powered Classification of Oral Lesions: Improving Early Detection and Diagnosis

Abstract

Oral malignancies pose significant global health challenges, with oral squamous cell carcinoma (OSCC) being the most prevalent form. Early detection of potentially malignant oral disorders (OPMDs) such as leukoplakia and oral submucous fibrosis is crucial for improving patient prognosis. Traditional diagnostic approaches often face limitations like subjective interpretation and potential delays. This study aimed to develop and evaluate a deep learning-based model for the classification of oral lesions as benign or malignant using publicly available image datasets. Utilizing a modified VGG16 architecture and optimized preprocessing techniques, the model was trained on 330 annotated intraoral images and achieved an overall accuracy of 94.79%. Key performance metrics included a precision of 95.11%, sensitivity and specificity of 94.58%, and an F1 score of 94.74%. The model’s performance was comparable to or exceeded existing models with larger datasets, demonstrating its capability for effective feature extraction and reliable classification. The high area under the curve (AUC) value of 0.96 reinforced its potential for clinical application. While the model showed strong diagnostic capability, expanding the dataset size and incorporating a broader range of cases could further enhance generalizability. Future work should also consider integrating real-time image acquisition and optimizing computational processes for practical deployment. The findings underscore the promise of AI-driven diagnostic tools in supporting healthcare professionals by enabling timely, accurate, and scalable detection of oral malignancies, thereby contributing to improved patient care and outcomes. This study represents a significant step toward the practical application of AI in oral health diagnostics.

Keywords

• OSCC detection
• Oral lesions
• OPMD classification
• VGG16 model

Yapay Zeka Destekli Oral Lezyon Sınıflandırması: Erken Tespit ve Tanıyı İyileştirme

Abstract

"Ağız maligniteleri, oral skuamöz hücreli karsinomun (OSCC) en yaygın form olmasıyla birlikte önemli küresel sağlık sorunları oluşturmaktadır. Lökoplaki ve oral submuköz fibrozis gibi potansiyel olarak malign oral bozuklukların (OPMB) erken teşhisi, hasta prognozunu iyileştirmek açısından kritik öneme sahiptir. Geleneksel tanı yaklaşımları genellikle öznel yorumlama ve olası gecikmeler gibi sınırlamalarla karşı karşıya kalmaktadır. Bu çalışma, kamuya açık görüntü veri setlerini kullanarak oral lezyonların benign veya malign olarak sınıflandırılması için derin öğrenme tabanlı bir model geliştirmeyi ve değerlendirmeyi amaçlamıştır. Modifiye edilmiş bir VGG16 mimarisi ve optimize edilmiş ön işleme teknikleri kullanılarak model, 330 anotasyonlu intraoral görüntü üzerinde eğitilmiş ve %94,79 genel doğruluk elde etmiştir. Temel performans ölçütleri arasında %95,11 hassasiyet (precision), %94,58 duyarlılık (sensitivity) ve özgüllük (specificity) ile %94,74 F1 skoru bulunmaktadır. Modelin performansı, daha büyük veri setlerine sahip mevcut modellerle karşılaştırılabilir veya daha üstün olarak değerlendirilmiş ve etkili özellik çıkarımı ile güvenilir sınıflandırma yapma yeteneğini göstermiştir. 0,96’lık yüksek ROC eğrisi altındaki alan (AUC) değeri, klinik uygulama potansiyelini güçlendirmiştir. Model güçlü bir tanısal yetenek sergilemekle birlikte, veri setinin boyutunun genişletilmesi ve daha geniş bir vaka yelpazesinin dahil edilmesi, genelleştirilebilirliği daha da artırabilir. Gelecekteki çalışmalar, gerçek zamanlı görüntü alımını entegre etmeyi ve pratik uygulama için hesaplama süreçlerini optimize etmeyi de dikkate almalıdır. Bulgular, ağız malignitelerinin zamanında, doğru ve ölçeklenebilir bir şekilde tespit edilmesine olanak tanıyan AI tabanlı tanı araçlarının sağlık profesyonellerine destek sağlama vaadini vurgulamakta ve hasta bakımı ile sonuçlarının iyileştirilmesine katkıda bulunmaktadır. Bu çalışma, yapay zekânın ağız sağlığı tanısındaki pratik uygulamasına yönelik önemli bir adımı temsil etmektedir."

Keywords

• OSCC tespiti
• Oral lezyonlar
• OPMD sınıflandırması
• VGG16 modeli

References

1. [1] S. Warnakulasuriya et al., ‘Oral Potentially Malignant Disorders: A Consensus Report From an International Seminar on Nomenclature and Classification, Convened by the WHO Collaborating Centre for Oral Cancer’, Oral Diseases, vol. 27, no. 8, pp. 1862–1880, 2020, doi: 10.1111/odi.13704.
2. [2] S. Yang, Y. Lee, L. Chang, C. Yang, C. Luo, and P. Wu, ‘Oral Tongue Leukoplakia: Analysis of Clinicopathological Characteristics, Treatment Outcomes, and Factors Related to Recurrence and Malignant Transformation’, Clinical Oral Investigations, vol. 25, no. 6, pp. 4045–4058, 2021, doi: 10.1007/s00784-020-03735-1.
3. [3] C. B. More and N. R. Rao, ‘Proposed Clinical Definition for Oral Submucous Fibrosis’, Journal of Oral Biology and Craniofacial Research, vol. 9, no. 4, pp. 311–314, 2019, doi: 10.1016/j.jobcr.2019.06.016.
4. [4] S. Abati, C. Bramati, S. Bondi, A. Lissoni, and M. Trimarchi, ‘Oral Cancer and Precancer: A Narrative Review on the Relevance of Early Diagnosis’, International Journal of Environmental Research and Public Health, vol. 17, no. 24, p. 9160, 2020, doi: 10.3390/ijerph17249160.
5. [5] K. Matsuoka, ‘Oral Malignant Melanoma Detected After Resection of Amelanotic Pulmonary Metastasis’, International Journal of Surgery Case Reports, vol. 4, no. 12, pp. 1169–1172, 2013, doi: 10.1016/j.ijscr.2013.10.004.
6. [6] L. Cigic, ‘Increased Prevalence of Oral Potentially Malignant Lesions Among Croatian War Invalids, a Cross-Sectional Study’, Journal of Clinical and Experimental Dentistry, pp. e734-741, 2023, doi: 10.4317/jced.60715.
7. [7] F. M. Ghanaei, F. Joukar, M. Rabiei, A. Dadashzadeh, and A. K. Valeshabad, ‘Prevalence of Oral Mucosal Lesions in an Adult Iranian Population’, Iranian Red Crescent Medical Journal, vol. 15, no. 7, pp. 600–604, 2013, doi: 10.5812/ircmj.4608.
8. [8] A. M. Kavarodi, M. Thomas, and J. Kannampilly, ‘Prevalence of Oral Pre-Malignant Lesions and Its Risk Factors in an Indian Subcontinent Low Income Migrant Group in Qatar’, Asian Pacific Journal of Cancer Prevention, vol. 15, no. 10, pp. 4325–4329, 2014, doi: 10.7314/apjcp.2014.15.10.4325.

9. [9] G. Guan and N. Firth, ‘Oral Manifestations as an Early Clinical Sign of Acute Myeloid Leukaemia: A Case Report’, Australian Dental Journal, vol. 60, no. 1, pp. 123–127, 2015, doi: 10.1111/adj.12272.
10. [10] H. Mawardi et al., ‘Oral Epithelial Dysplasia and Squamous Cell Carcinoma Following Allogeneic Hematopoietic Stem Cell Transplantation: Clinical Presentation and Treatment Outcomes’, Bone Marrow Transplantation, vol. 46, no. 6, pp. 884–891, 2011, doi: 10.1038/bmt.2011.77.
11. [11] N. Al-Rawi et al., ‘The Effectiveness of Artificial Intelligence in Detection of Oral Cancer’, International Dental Journal, vol. 72, no. 4, pp. 436–447, Aug. 2022, doi: 10.1016/j.identj.2022.03.001.
12. [12] M. García-Pola, E. Pons-Fuster, C. Suárez-Fernández, J. Seoane-Romero, A. Romero-Méndez, and P. López-Jornet, ‘Role of Artificial Intelligence in the Early Diagnosis of Oral Cancer. A Scoping Review’, Cancers, vol. 13, no. 18, p. 4600, Sep. 2021, doi: 10.3390/cancers13184600.
13. [13] S. Nath, R. Raveendran, and S. Perumbure, ‘Artificial Intelligence and Its Application in the Early Detection of Oral Cancers’, Clin Cancer Investig J, vol. 11, no. 1, pp. 5–9, 2022, doi: 10.51847/h7wa0UHoIF.
14. [14] K. Warin, W. Limprasert, S. Suebnukarn, S. Jinaporntham, P. Jantana, and S. Vicharueang, ‘AI-based analysis of oral lesions using novel deep convolutional neural networks for early detection of oral cancer’, PLoS ONE, vol. 17, no. 8, p. e0273508, Aug. 2022, doi: 10.1371/journal.pone.0273508.
15. [15] F. Jubair, O. Al‐karadsheh, D. Malamos, S. Al Mahdi, Y. Saad, and Y. Hassona, ‘A novel lightweight deep convolutional neural network for early detection of oral cancer’, Oral Diseases, vol. 28, no. 4, pp. 1123–1130, May 2022, doi: 10.1111/odi.13825.
16. [16] Q. Huang, H. Ding, and N. Razmjooy, ‘Optimal deep learning neural network using ISSA for diagnosing the oral cancer’, Biomedical Signal Processing and Control, vol. 84, p. 104749, Jul. 2023, doi: 10.1016/j.bspc.2023.104749.
17. [17] Q. Fu et al., ‘A deep learning algorithm for detection of oral cavity squamous cell carcinoma from photographic images: A retrospective study’, EClinicalMedicine, vol. 27, p. 100558, Oct. 2020, doi: 10.1016/j.eclinm.2020.100558.
18. [18] S. Bansal, R. S. Jadon, and S. K. Gupta, ‘Lips and Tongue Cancer Classification Using Deep Learning Neural Network’, in 2023 6th International Conference on Information Systems and Computer Networks (ISCON), Mathura, India: IEEE, Mar. 2023, pp. 1–3. doi: 10.1109/ISCON57294.2023.10112158.
19. [19] H. Lin, H. Chen, L. Weng, J. Shao, and J. Lin, ‘Automatic detection of oral cancer in smartphone-based images using deep learning for early diagnosis’, J. Biomed. Opt., vol. 26, no. 08, Aug. 2021, doi: 10.1117/1.JBO.26.8.086007.
20. [20] G. Tanriver, M. Soluk Tekkesin, and O. Ergen, ‘Automated Detection and Classification of Oral Lesions Using Deep Learning to Detect Oral Potentially Malignant Disorders’, Cancers, vol. 13, no. 11, p. 2766, 2021.
21. [21] M. Bai and M. Li, ‘A Presentation of Structures and Applications of Convolutional Neural Networks’, Highlights in Science Engineering and Technology, vol. 61, pp. 180–187, 2023, doi: 10.54097/hset.v61i.10291.
22. [22] K. O’Shea and R. R. Nash, ‘An Introduction to Convolutional Neural Networks’, 2015, doi: 10.48550/arxiv.1511.08458.
23. [23] N. Zakaria, ‘Improved Image Classification Task Using Enhanced Visual Geometry Group of Convolution Neural Networks’, Joiv International Journal on Informatics Visualization, vol. 7, no. 4, p. 2498, 2023, doi: 10.30630/joiv.7.4.1752.
24. [24] D. Kwiatkowska, P. Kluska, and A. Reich, ‘Convolutional Neural Networks for the Detection of Malignant Melanoma in Dermoscopy Images’, Advances in Dermatology and Allergology, vol. 38, no. 3, pp. 412–420, 2021, doi: 10.5114/ada.2021.107927.
25. [25] Y. Tian, ‘Artificial Intelligence Image Recognition Method Based on Convolutional Neural Network Algorithm’, Ieee Access, vol. 8, pp. 125731–125744, 2020, doi: 10.1109/access.2020.3006097.
26. [26] K. Simonyan and A. Zisserman, ‘Very Deep Convolutional Networks for Large-Scale Image Recognition’, 2014, arXiv. doi: 10.48550/ARXIV.1409.1556.
27. [27] I. Fawwaz, T. Candra, D. A. M. Marpaung, A. Dinis, and M. R. Fachrozi, ‘Classification of Beetle Type Using the Convolutional Neural Network Algorithm’, Sinkron, vol. 7, no. 4, pp. 2340–2348, 2022, doi: 10.33395/sinkron.v7i4.11673.
28. [28] Akshitha and M. Veena, ‘Melanoma Detection Using CNN’, International Research Journal of Modernization in Engineering Technology and Science, 2023, doi: 10.56726/irjmets43733.
29. [29] H. Yılmaz, ‘AI-Powered Healthcare Innovations: Rehabilitation, Education, And Early Diagnosis’, Sep. 2024, Serüven Yayınevi. doi: 10.5281/ZENODO.13885904.
30. [30] M. Alehegn, ‘Application of Machine Learning and Deep Learning for the Prediction of HIV/AIDS’, Hiv & Aids Review, vol. 21, no. 1, pp. 17–23, 2022, doi: 10.5114/hivar.2022.112852.
31. [31] M. Sangeetha, ‘Heart Disease Prediction Using ML’, International Journal of Innovative Science and Research Technology, pp. 2630–2633, 2024, doi: 10.38124/ijisrt/ijisrt24mar2016.
32. [32] P. S. Mattas and I. Nadaan, ‘Optimizing Cardiovascular Disease Diagnosis With Machine Learning: An Analysis’, International Journal of Research Publication and Reviews, vol. 04, no. 02, pp. 430–434, 2023, doi: 10.55248/gengpi.2023.4217.
33. [33] C. S. Anita, P. Nagarajan, G. A. Sairam, P. Ganesh, and G. Deepakkumar, ‘Fake Job Detection and Analysis Using Machine Learning and Deep Learning Algorithms’, Revista Gestão Inovação E Tecnologias, vol. 11, no. 2, pp. 642–650, 2021, doi: 10.47059/revistageintec.v11i2.1701.
34. [34] U. Ramasamy and S. Santhoshkumar, ‘Benchmark Datasets and Real-Time Autoimmune Disease Dataset Analysis Using Machine Learning Algorithms With Implementation, Analysis and Results’, Journal of Intelligent & Fuzzy Systems, vol. 45, no. 2, pp. 2449–2463, 2023, doi: 10.3233/jifs-224115.
35. [35] H. T. Sihotang, M. K. Albert, F. Riandari, and L. A. Rendell, ‘Efficient Optimization Algorithms for Various Machine Learning Tasks, Including Classification, Regression, and Clustering’, Idea, vol. 1, no. 1, pp. 14–24, 2023, doi: 10.35335/idea.v1i1.3.
36. [36] S. A. Pane and F. M. Sihombing, ‘Classification of Rock Mineral in Field X Based on Spectral Data (SWIR &Amp; TIR) Using Supervised Machine Learning Methods’, Iop Conference Series Earth and Environmental Science, vol. 830, no. 1, p. 012042, 2021, doi: 10.1088/1755-1315/830/1/012042.