HYBRID AI FOR GASTRIC DIAGNOSTICS: MULTI-MODAL FUSION AND INTERPRETABILITY FOR EARLY ULCER AND CANCER DETECTION

Iskandarova Sayyora Nurmamatovna; Abdurashidova Kamola Turg’unbayevna; Omonov Sanjarbek G‘anisher o’g’li; Sabitova Nazokat Qobuljon qizi

Authors

Iskandarova Sayyora Nurmamatovna Tashkent University of Information Technologies named after Muhammad al-Khwarizmi
Abdurashidova Kamola Turg’unbayevna Tashkent University of Information Technologies named after Muhammad al-Khwarizmi
Omonov Sanjarbek G‘anisher o’g’li Tashkent State University of Economics
Sabitova Nazokat Qobuljon qizi Tashkent University of Information Technologies named after Muhammad al-Khwarizmi

Keywords:

Hybrid Artificial Intelligence, Gastric Cancer, Gastric Ulcer, Multi-Modal Fusion, Deep Learning, Explainable AI (XAI), Diagnostic Imaging, Fuzzy Logic, Early Detection

Abstract

This study investigates the development of hybrid Artificial Intelligence (AI) models for the early and accurate diagnosis of gastric ulcer and gastric cancer. Traditional diagnostic methods such as endoscopy and biopsy, though widely used, face limitations including inter-observer variability, sampling errors, and delays in diagnosis. To address these challenges, two novel hybrid AI architectures are proposed: the Deep Feature-Attention SVM (DFA-SVM) and the Multi-Modal Fuzzy-Fusion Network (M2F2-Net). These models integrate multi-modal data sources, including endoscopic images, histopathology slides, and structured clinical records, while employing attention mechanisms, fuzzy logic, and interpretable AI approaches to enhance diagnostic performance. Experimental evaluation using the conceptual GIMIC-12K dataset demonstrates that M2F2-Net achieves superior performance, particularly in early gastric cancer detection, by significantly reducing false negatives and improving sensitivity. The findings highlight the transformative potential of hybrid AI in improving patient outcomes, optimizing healthcare efficiency, and supporting explainable decision-making in clinical practice.

References

1. Ahmed, I.A.; Senan, E.M.; Shatnawi, H.S.A. "Hybrid Models for Endoscopy Image Analysis for Early Detection of Gastrointestinal Diseases Based on Fused Features." Diagnostics, vol. 13, no. 10, 1758, 2023.

https://doi.org/10.3390/diagnostics13101758.

2. Dahan, F., Shah, J.H., Saleem, R., Hasnain, M., Afzal, M. & Alfakih, T.M. "A hybrid XAI-driven deep learning framework for robust GI tract disease diagnosis." Scientific Reports, vol. 15, no. 1, 21139, 2025.

https://www.nature.com/articles/s41598-025-07690-3.

3. Dosovitskiy, A., et al. "An Image Is Worth 16x16 Words: Transformers for Image Recognition at Scale." International Conference on Learning Representations (ICLR), 2021.

4. Fadl Dahan, Jamal Hussain Shah, Rabia Saleem, Muhammad Hasnain, Maira Afzal & Taha M. Alfakih. "A hybrid XAIdriven deep learning framework for robust GI tract disease diagnosis." Scientific Reports, vol. 15, Article number: 21139, 2025.

https://www.nature.com/articles/s41598-025-07690-3.

5. Figueiredo, P., Figueiredo, I., Pinto, L., Kumar, S., Tsai, Y.-H., & Mamonov, A. "Polyp detection with computer-aided diagnosis in white light colonoscopy: comparison of three different methods." Endoscopic International Open, vol. 7, no. 2, E209-E215, 2019.

https://doi.org/10.1055/a-0808-4456.

6. He, K., et al. "Deep Residual Learning for Image Recognition." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, 770-778.

7. Huang, G., et al. "Densely Connected Convolutional Networks." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, 4700-4708.

8. Huang, Y., Shao, Y., Yu, X., Chen, C., Guo, J., & Ye, G. "Global progress and future prospects of early gastric cancer screening." Journal of Cancer, vol. 15, no. 10, 3045-3064, 2024.

https://www.jcancer.org/v15p3045.pdf.

9. Kim, S., Lee, N., Lee, J., Hyun, D., Park, C. "Heterogeneous Graph Learning for Multi-Modal Medical Data Analysis." Proceedings of the AAAI Conference on Artificial Intelligence, vol. 37, no. 12, 5141-5150, 2023.

https://ojs.aaai.org/index.php/AAAI/article/view/25643/25415.

10. Kim, W., et al. "Artificial Intelligence in Gastric Cancer Imaging With Emphasis on Diagnostic Imaging and Body Morphometry." Journal of Gastric Cancer, vol. 23, no. 3, 388-399, 2023.

https://pubmed.ncbi.nlm.nih.gov/37553127/.

11. Ijaz, M. F. "Artificial intelligence in disease diagnosis: a systematic literature review, synthesizing framework and future research agenda." Journal of Ambient Intelligence and Humanized Computing, vol. 14, no. 7, 8459-8486, 2022.

https://pmc.ncbi.nlm.nih.gov/articles/PMC8754556/.

12. Macenko, M., et al. "A method for normalizing histology slides for quantitative analysis." IEEE International Symposium on Biomedical Imaging (ISBI), 2009, 1107-1110.

13. Mohammad Amin Morid, Alireza Borjali, Guilherme Del Fiol. "A scoping review of transfer learning research on medical image analysis using ImageNet." arXiv, 2020, pp. 1-38.https://arxiv.org/pdf/2004.13175.

14. Pal, A.; Rai, H.M.; Frej, M.B.H.; Razaque, A. "Advanced Segmentation of Gastrointestinal (GI) Cancer Disease Using a Novel U-MaskNet Model." Life, vol. 14, no. 9, 1488, 2024.

https://pdfs.semanticscholar.org/eoa7/874cb7f2eac933309fb3f2b2ab2b19ed30da.pdf.

15. Parkash, O., Siddiqui, A. T. S., Jiwani, U., Rind, F., Padhani, Z. A., Rizvi, A., Hoodbhoy, Z., & Das, J. K. "Diagnostic accuracy of artificial intelligence for detecting gastrointestinal luminal pathologies: A systematic review and meta-analysis." Frontiers in Medicine, vol. 9, 1018937, 2022.

https://ecommons.aku.edu/pakistan_fhs_mc_med_gastroenterol/318.