A Hybrid U-Net and Vision Transformer Framework with Dimensionality Reduction for Automated Detection of Throat Cancer Metastasis
DOI:
https://doi.org/10.3991/ijoe.v22i08.62322Keywords:
Throat cancer, metastasis detection, U-Net, Vision Transformer, t-SNE, Random Forest, medical image analysis, lymph node classification, automated oncology diagnosticsAbstract
Ensuring the early and accurate detection of lymph node metastasis continues to underpin prognostication and treatment for throat cancer patients. Histopathological whole-slide images (WSIs) are manually analysed by pathologists in clinical practice; however, this process is labour-intensive and requires a significant amount of time, leading to poor reproducibility, inter-observer variability, and delays/inconsistencies in diagnostics. Addressing these limitations, we propose a U-Net segmentation, Vision Transformer (ViT)-based feature extraction, t-Distributed Stochastic Neighbour Embedding (t-SNE) for dimensionality reduction, and a random forest (RF) classification-driven hybrid deep learning framework for fully automated patient-level detection. The U-Net module segments lymph node regions from multimodal scans (MRI and PET), which makes the localisation of somatic lesions clearer. Features extracted from the ViT both represent high and low-spatial-frequency histological signatures and long-range relations that are then mapped via t-SNE to minimise within-cluster distance while maximising between-cluster distances both locally and globally. This then results in robust discrimination between metastatic and non-metastatic lymph nodes with a RF classifier. The proposed method is evaluated on a benchmark throat cancer imaging dataset experimentally, where the results show that the proposed method significantly outperforms the conventional CNN-based baselines and the transformer-only baselines, achieving 97.6%, 96.8% and 95.9% accuracy, sensitivity and specificity, respectively. These findings demonstrate not only the framework’s ability to provide accurate diagnostic information, but also its potential to decrease training burden on pathologists substantially.
References
[1] Zhong, Y. W., Jiang, Y., Dong, S., Wu, W. J., Wang, L. X., Zhang, J., & Huang, M. W. (2022). Tumor radiomics signature for artificial neural network-assisted detection of neck metastasis in patient with tongue cancer. Journal of Neuroradiology, 49(2), 213-218.
[2] Zhou, Z., Xue, J., Wu, Y., Mao, J., Li, C., Yu, X., ... & Zhao, G. (2025). Automated detection of metastatic lymph nodes in head and neck malignant tumors on high− resolution MRI images using an improved convolutional neural network. International Journal of Medical Informatics, 200, 105904.
[3] Alix-Panabieres, C., Magliocco, A., Cortes-Hernandez, L. E., Eslami-S, Z., Franklin, D., & Messina, J. L. (2022). Detection of cancer metastasis: past, present and future. Clinical & Experimental Metastasis, 39(1), 21-28.
[4] Kim, G. H., Sung, E. S., & Nam, K. W. (2021). Automated laryngeal mass detection algorithm for home-based self-screening test based on convolutional neural network. BioMedical Engineering Online, 20(1), 51.
[5] Banavar, G., Ogundijo, O., Julian, C., Toma, R., Camacho, F., Torres, P. J., & Punyadeera, C. (2023). Detecting salivary host and microbiome RNA signature for aiding diagnosis of oral and throat cancer. Oral Oncology, 145, 106480.
[6] Kato, S., Amemiya, S., Takao, H., Yamashita, H., Sakamoto, N., & Abe, O. (2021). Automated detection of brain metastases on non-enhanced CT using single-shot detectors. Neuroradiology, 63(12), 1995-2004.
[7] Eida, S., Fukuda, M., Katayama, I., Takagi, Y., Sasaki, M., Mori, H., ... & Sumi, M. (2024). Metastatic lymph node detection on ultrasound images using YOLOv7 in patients with head and neck squamous cell carcinoma. Cancers, 16(2), 274.
[8] Xu, X., Xi, L., Wei, L., Wu, L., Xu, Y., Liu, B., ... & Shang, Z. (2023). Deep learning assisted contrast-enhanced CT–based diagnosis of cervical lymph node metastasis of oral cancer: a retrospective study of 1466 cases. European Radiology, 33(6), 4303-4312.
[9] Wu, I. C., Chen, Y. C., Karmakar, R., Mukundan, A., Gabriel, G., Wang, C. C., & Wang, H. C. (2024). Advancements in hyperspectral imaging and computer-aided diagnostic methods for the enhanced detection and diagnosis of head and neck cancer. Biomedicines, 12(10), 2315.
[10] Chang, C. Y., Buckless, C., Yeh, K. J., & Torriani, M. (2022). Automated detection and segmentation of sclerotic spinal lesions on body CTs using a deep convolutional neural network. Skeletal Radiology, 51(2), 391-399.
[11] Chen, K., Lai, Y. C., Vanniarajan, B., Wang, P. H., Wang, S. C., Lin, Y. C., ... & Lin, G. (2022). Clinical impact of a deep learning system for automated detection of missed pulmonary nodules on routine body computed tomography including the chest region. European Radiology, 32(5), 2891-2900.
[12] Takao, H., Amemiya, S., Kato, S., Yamashita, H., Sakamoto, N., & Abe, O. (2022). Deep-learning 2.5-dimensional single-shot detector improves the performance of automated detection of brain metastases on contrast-enhanced CT. Neuroradiology, 64(8), 1511-1518.
[13] Nikulin, P., Zschaeck, S., Maus, J., Cegla, P., Lombardo, E., Furth, C., ... & van den Hoff, J. (2023). A convolutional neural network with self-attention for fully automated metabolic tumor volume delineation of head and neck cancer in [18 F] FDG PET/CT. European Journal of Nuclear Medicine and Molecular Imaging, 50(9), 2751-2766.
[14] Lee, T. L., Chen, T. H., Kuo, Y. J., Lan, H. Y., Yang, M. H., & Chu, P. Y. (2023). Tumor-associated tissue eosinophilia promotes angiogenesis and metastasis in head and neck squamous cell carcinoma. Neoplasia, 35, 100855.
[15] Iuga, A. I., Carolus, H., Höink, A. J., Brosch, T., Klinder, T., Maintz, D., ... & Püsken, M. (2021). Automated detection and segmentation of thoracic lymph nodes from CT using 3D foveal fully convolutional neural networks. BMC Medical Imaging, 21(1), 69.
[16] Mohamed, N., Almutairi, R. L., Abdelrahim, S., Alharbi, R., Alhomayani, F. M., Elamin Elnaim, B. M., & Dhakal, R. (2023). Automated Laryngeal cancer detection and classification using dwarf mongoose optimization algorithm with deep learning. Cancers, 16(1), 181.
[17] Wen, X., Wang, Y., Zhang, D., Xiu, Y., Sun, L., Zhao, B., ... & Xing, D. (2025). U2-Attention-Net: a deep learning automatic delineation model for parotid glands in head and neck cancer organs at risk on radiotherapy localization computed tomography images. Physica Medica, 135, 105024.
[18] Ding, H., Cen, Q., Si, X., Pan, Z., & Chen, X. (2022). Automatic glottis segmentation for laryngeal endoscopic images based on U-Net. Biomedical Signal Processing and Control, 71, 103116.
[19] Chen, C., Zhang, T., Teng, Y., Yu, Y., Shu, X., Zhang, L., & Xu, J. (2023). Automated segmentation of craniopharyngioma on MR images using U-Net-based deep convolutional neural network. European Radiology, 33(4), 2665-2675.
[20] Taku, N., Wahid, K. A., van Dijk, L. V., Sahlsten, J., Jaskari, J., Kaski, K.,& Naser, M. A. (2022). Auto-detection and segmentation of involved lymph nodes in HPV-associated oropharyngeal cancer using a convolutional deep learning neural network. Clinical and Translational Radiation Oncology, 36, 47-55.
[21] Lin, Y. C., Lin, G., Pandey, S., Yeh, C. H., Wang, J. J., Lin, C. Y., ... & Ng, S. H. (2023). Fully automated segmentation and radiomics feature extraction of hypopharyngeal cancer on MRI using deep learning. European Radiology, 33(9), 6548-6556.
[22] Gandikota, H. P., Abirami, S., & Kumar, M. S. (2023). Bottleneck Feature-Based U-Net for Automated Detection and Segmentation of Gastrointestinal Tract Tumors from CT Scans. Traitement du Signal, 40(6), 2789.
[23] Ariji, Y., Kise, Y., Fukuda, M., Kuwada, C., & Ariji, E. (2022). Segmentation of metastatic cervical lymph nodes from CT images of oral cancers using deep-learning technology. Dentomaxillofacial Radiology, 51(4), 20210515.
[24] Rao, D., K, P., & Singh, R. (2022). Automated segmentation of the larynx on computed tomography images: a review. Biomedical Engineering Letters, 12(2), 175-183.
[25] Tomita, H., Yamashiro, T., Heianna, J., Nakasone, T., Kobayashi, T., Mishiro, S., ... & Kobayashi, Y. (2021). Deep learning for the preoperative diagnosis of metastatic cervical lymph nodes on contrast-enhanced computed tomography in patients with oral squamous cell carcinoma. Cancers, 13(4), 600.
[26] Rokhshad, R., Salehi, S. N., Yavari, A., Shobeiri, P., Esmaeili, M., Manila, N., ... & Mohammad-Rahimi, H. (2024). Deep learning for diagnosis of head and neck cancers through radiographic data: a systematic review and meta-analysis. Oral radiology, 40(1), 1-20.
[27] Caldonazzi, N., Rizzo, P. C., Eccher, A., Girolami, I., Fanelli, G. N., Naccarato, A. G., ... & Marletta, S. (2023). Value of artificial intelligence in evaluating lymph node metastases. Cancers, 15(9), 2491.
[28] Puladi, B., Ooms, M., Kintsler, S., Houschyar, K. S., Steib, F., Modabber, A., ... & Braunschweig, T. (2021). Automated PD-L1 scoring using artificial intelligence in head and neck squamous cell carcinoma. Cancers, 13(17), 4409.
[29] Liu, X., Han, C., Cui, Y., Xie, T., Zhang, X., & Wang, X. (2021). Detection and segmentation of pelvic bones metastases in MRI images for patients with prostate cancer based on deep learning. Frontiers in oncology, 11, 773299.
[30] Hegde, S., Ajila, V., Zhu, W., & Zeng, C. (2022). Artificial intelligence in early diagnosis and prevention of oral cancer. Asia-Pacific journal of oncology nursing, 9(12), 100133.
[31] Al Hasan, M. M., Ghazimoghadam, S., Tunlayadechanont, P., Mostafiz, M. T., Gupta, M., Roy, A., ... & Forghani, R. (2024). Automated segmentation of lymph nodes on neck ct scans using deep learning. Journal of Imaging Informatics in Medicine, 37(6), 2955-2966.
[32] Wang, R., Chen, X., Zhang, X., He, P., Ma, J., Cui, H., ... & Wu, Y. (2024). Automatic segmentation of esophageal cancer, metastatic lymph nodes and their adjacent structures in CTA images based on the UperNet Swin network. Cancer Medicine, 13(18), e70188.
[33] Iantsen, A. (2024). Improving the U-Net configuration for automated delineation of head and neck cancer on MRI. In Challenge on Head and Neck Tumor Segmentation for MRI-Guided Applications (pp. 230-240). Cham: Springer Nature Switzerland.
[34] Weikert, T., Jaeger, P. F., Yang, S., Baumgartner, M., Breit, H. C., Winkel, D. J., ... & Sauter, A. W. (2023). Automated lung cancer assessment on 18F-PET/CT using Retina U-Net and anatomical region segmentation. European radiology, 33(6), 4270-4279.
[35] Kumar, A., Mehta, R., Reddy, B. R., & Singh, K. K. (2024). Vision transformer based effective model for early detection and classification of lung cancer. SN Computer Science, 5(7), 839.
[36] Chilet-Martos, E., Vila-Francés, J., Bagan, J. V., & Vives-Gilabert, Y. (2025). Automated classification of oral cancer lesions: Vision transformers vs radiomics. Computers in Biology and Medicine, 189, 109913.
[37] Song, B., Kc, D. R., Yang, R. Y., Li, S., Zhang, C., & Liang, R. (2024). Classification of mobile-based oral cancer images using the vision transformer and the Swin transformer. Cancers, 16(5), 987.
[38] Asif, S., Wang, V. Y., & Xu, D. (2025). OralTransNet: A novel hybrid model integrating transformer attention and CNN features for accurate diagnosis of mouth and oral diseases. Engineering Applications of Artificial Intelligence, 159, 111609.
[39] Ayana, G., Barki, H., & Choe, S. W. (2024). Pathological insights: Enhanced vision transformers for the early detection of colorectal cancer. Cancers, 16(7), 1441.
[40] S. Jayapal and R. Annamalai, "Graph Neural Network With Hessian-Based Locally Linear Embedding for Cancer Metastasis Analysis in Lymph Nodes Using DeepLab Segmentation," in IEEE Access, vol. 13, pp. 46448-46458, 2025, doi: 10.1109/ACCESS.2025.3546716.
[41] B. Ganesh, R. Annamalai and S. Jayapal, "Enhancing Gastrointestinal Image Classification: A Fusion of CNN and Vision Trasformers," 2024 IEEE International Conference on Contemporary Computing and Communications (InC4), Bangalore, India, 2024, pp. 1-6, doi: 10.1109/InC460750.2024.10649204.
[42] Palanisamy, R., Kanagala, S., & Govardhan Rao, K. V. (2025). Deep learning for insulator and binding fault diagnosis in smart grids with YOLOv11 instance segmentation. In 2025, IEEE 6th International Conference on Robotics and Manufacturing Automation (ROMA) (pp. 1–6). IEEE. https://doi.org/10.1109/ROMA66616.2025.11155726
[43] S. Xavier P., P. K. Sathish, and G. Raju, “Advancing Brain Tumor Segmentation in MRI Scans Using Hybrid Attention-Residual UNET with Transformer Blocks (HART-UNet),” International Journal of Online and Biomedical Engineering (iJOE), Vol. 20 No. 6, 2024.https://doi.org/10.3991/ijoe.v20i06.46979
[44] Ghali et al., “Hybrid Attention and Learnable Thresholding for 3D Brain Tumor Segmentation,” International Journal of Online and Biomedical Engineering (iJOE), vol. 22, no. 3, 2026.https://doi.org/10.3991/ijoe.v22i03.58359
[45] D. A. Cadillo-Laurentt and E. A. Paiva-Peredo, “Histopathological Image Classification Using Convolutional Neural Networks for Detection of Metastatic Breast Cancer in Lymph Nodes,” International Journal of Online and Biomedical Engineering (iJOE), vol. 20, no. 2, 2024.https://doi.org/10.3991/ijoe.v20i02.46789
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Senthil Jayapal, R. Annamalai
This work is licensed under a Creative Commons Attribution 4.0 International License.
