Dual-Scale Transformer-Guided Attention Network for Efficient Multi-OAR Segmentation in Head and Neck Radiotherapy

Uzma  Nawaz; Hafiz Muhammad Ubaidullah; Zubair Saeed; Chaudhry Muhammad Ali Nawaz

doi:10.58190/imiens.2025.126

Authors

Uzma Nawaz Knowledge and Data Science Research Centre, Department of Computer and Software Engineering, College of Electrical and Mechanical Engineering, National University of Science and Technology, Islamabad, Pakistan https://orcid.org/0009-0003-3805-8101
Hafiz Muhammad Ubaidullah Department of Computer Information Systems Engineering, NED University of Engineering and Technology, Karachi, Pakistan https://orcid.org/0009-0009-9166-8949
Zubair Saeed Department of Electrical & Computer Engineering, Texas A&M University, College Station, TX, USA https://orcid.org/0000-0001-5302-7133
Chaudhry Muhammad Ali Nawaz Department of Creative Technologies, Software Engineering, Air University, Islamabad, Pakistan https://orcid.org/0009-0007-1256-0629

DOI:

https://doi.org/10.58190/imiens.2025.126

Keywords:

Deep Learning, segmentation, Organ-at-risk, CT images, convolutional neural networks, dual-scale transformer-guided attention network

Abstract

Accurate segmentation of organ-at-risk (OARs) in head and neck CT images is crucial for radiotherapy planning, but it remains a challenging task due to anatomical complexity, low soft-tissue contrast, and the presence of small, variable structures. We propose DSTANet, a novel dual-scale transformer-guided attention network that integrates multi-resolution encoding, transformer-based global context fusion, and anatomically guided attention refinement to deliver precise multi-OAR segmentation. Unlike traditional CNN-based methods, DSTANet effectively models long-range spatial dependencies while preserving high-resolution boundary detail. On the HNSCC-3DCT-RT dataset, DSTANet achieved a mean Dice Score of 97.5% and a mean 95^th percentile Hausdorff Distance (HD95) of 2.32 mm, while on the MICCAI 2015 benchmark dataset, it achieved 90.0% Dice, which surpasses several state-of-the-art approaches both in terms of overlap and geometric accuracy. These results, combined with a sub-20-second inference time, establish DSTANet as a robust and clinically viable solution for automated head and neck OAR segmentation.

Downloads

Download data is not yet available.

References

[1] Jin, D., Guo, D., Ge, J., Ye, X., & Lu, L. (2022). Towards automated organs at risk and target volumes contouring: Defining precision radiation therapy in the modern era. Journal of the National Cancer Center, 2(4), 306-313.

[2] Bose, P., Brockton, N. T., & Dort, J. C. (2013). Head and neck cancer: from anatomy to biology. International journal of cancer, 133(9), 2013-2023.

[3] Jaffray, D. A., Lindsay, P. E., Brock, K. K., Deasy, J. O., & Tomé, W. A. (2010). Accurate accumulation of dose for improved understanding of radiation effects in normal tissue. International Journal of Radiation Oncology* Biology* Physics, 76(3), S135-S139.

[4] Savenije, M. H., Maspero, M., Sikkes, G. G., van der Voort van Zyp, J. R., TJ Kotte, A. N., Bol, G. H., & T. van den Berg, C. A. (2020). Clinical implementation of MRI-based organs-at-risk auto-segmentation with convolutional networks for prostate radiotherapy. Radiation oncology, 15, 1-12.

[5] Han, X., Hoogeman, M. S., Levendag, P. C., Hibbard, L. S., Teguh, D. N., Voet, P., ... & Wolf, T. K. (2008, September). Atlas-based auto-segmentation of head and neck CT images. In International Conference on Medical Image Computing and Computer-assisted Intervention (pp. 434-441). Berlin, Heidelberg: Springer Berlin Heidelberg.

[6] Jeffery, G. (2001). Architecture of the optic chiasm and the mechanisms that sculpt its development. Physiological Reviews, 81(4), 1393-1414.

[7] King, A. D. (2017). Imaging Society (ICIS) 17th Annual Teaching Course. Cancer Imaging, 17(1), O1.

[8] Nawaz, U., Saeed, Z., & Atif, K. (2025). A novel framework for efficient dominance-based rough set approximations using K-dimensional (KD) tree partitioning and adaptive recalculations techniques. Engineering Applications of Artificial Intelligence, 154, 110993.

[9] Nawaz, U., Anees-ur-Rahaman, M., & Saeed, Z. (2025). A review of neuro-symbolic AI integrating reasoning and learning for advanced cognitive systems. Intelligent Systems with Applications, 200541.

[10] Nawaz, U., Anees-ur-Rahaman, M., & Saeed, Z. (2025). A Survey of Deep Learning Approaches for the Monitoring and Classification of Seagrass. Ocean Science Journal, 60(2), 19.

[11] Nawaz, U., Saeed, Z., & Atif, K. (2025). A Novel Transformer-based approach for adult’s facial emotion recognition. IEEE Access.

[12] Mirza, F., & Zhao, H. (2024, August). Hybrid Attention Mechanisms and Bio-Inspired Optimization for Enhanced Breast Cancer Diagnosis from Ultrasound Images. In 2024 7th International Conference on Pattern Recognition and Artificial Intelligence (PRAI) (pp. 786-792). IEEE.

[13] Shelhamer, E., Long, J., & Darrell, T. (2016). Fully convolutional networks for semantic segmentation. IEEE transactions on pattern analysis and machine intelligence, 39(4), 640-651.

[14] Nemoto, T., Futakami, N., Yagi, M., Kumabe, A., Takeda, A., Kunieda, E., & Shigematsu, N. (2020). Efficacy evaluation of 2D, 3D U-Net semantic segmentation and atlas-based segmentation of normal lungs excluding the trachea and main bronchi. Journal of radiation research, 61(2), 257-264.

[15] Ma, Y. (2021, November). PANet: parallel attention network for remote sensing image semantic segmentation. In ISCTT 2021; 6th International Conference on Information Science, Computer Technology and Transportation (pp. 1-4). VDE.

[16] Smelyanskiy, M., Holmes, D., Chhugani, J., Larson, A., Carmean, D. M., Hanson, D., ... & Robb, R. (2009). Mapping high-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures. IEEE transactions on visualization and computer graphics, 15(6), 1563-1570.

[17] Liu, C., Zhang, X., Si, W., & Ni, X. (2021). Multiview Self‐Supervised Segmentation for OARs Delineation in Radiotherapy. Evidence‐Based Complementary and Alternative Medicine, 2021(1), 8894222.

[18] The Cancer Imaging Archive. (n.d.). HNSCC-3DCT-RT. https://www.cancerimagingarchive.net/collection/hnscc-3dct-rt/

[19] Papers With Code. (n.d.). MICCAI 2015 Head and Neck Auto Segmentation Challenge. https://paperswithcode.com/dataset/miccai-2015-head-and-neck-challenge

[20] Zhu, W., Huang, Y., Zeng, L., Chen, X., Liu, Y., Qian, Z., ... & Xie, X. (2019). AnatomyNet: deep learning for fast and fully automated whole‐volume segmentation of head and neck anatomy. Medical physics, 46(2), 576-589.

[21] Gao, Y., Huang, R., Chen, M., Wang, Z., Deng, J., Chen, Y., ... & Li, H. (2019). FocusNet: imbalanced large and small organ segmentation with an end-to-end deep neural network for head and neck CT images. In Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Conference, Shenzhen, China, October 13–17, 2019, Proceedings, Part III 22 (pp. 829-838). Springer International Publishing.

[22] Liang, S., Thung, K. H., Nie, D., Zhang, Y., & Shen, D. (2020). Multi-view spatial aggregation framework for joint localization and segmentation of organs at risk in head and neck CT images. IEEE Transactions on Medical Imaging, 39(9), 2794-2805.

[23] Chen, H., Huang, D., Lin, L., Qi, Z., Xie, P., Wei, J., ... & Lu, Y. (2020). Prior attention enhanced convolutional neural network based automatic segmentation of organs at risk for head and neck cancer radiotherapy. IEEE Access, 8, 179018-179027.

[24] Wang, T., Lei, Y., Roper, J., Ghavidel, B., Beitler, J. J., McDonald, M., ... & Yang, X. (2021). Head and neck multi-organ segmentation on dual-energy CT using dual pyramid convolutional neural networks. Physics in Medicine & Biology, 66(11), 115008.

[25] Dai, X., Lei, Y., Wang, T., Zhou, J., Rudra, S., McDonald, M., ... & Yang, X. (2022). Multi-organ auto-delineation in head-and-neck MRI for radiation therapy using regional convolutional neural network. Physics in Medicine & Biology, 67(2), 025006.

[26] Dai, X., Lei, Y., Wang, T., Tian, Z., Zhou, J., McDonald, M., ... & Yang, X. (2022, April). Automated CT segmentation for rapid assessment of anatomical variations in head-and-neck radiation therapy. In Medical Imaging 2022: Image-Guided Procedures, Robotic Interventions, and Modeling (Vol. 12034, pp. 306-311). SPIE.

[27] Gaikwad, U., & Shah, K. (2024). Hidden Markov Random Field Model Based VGG-16 for Segmentation and Classification of Head and Neck Cancer. International Journal of Intelligent Engineering & Systems, 17(1).

[28] Müller, D., Voran, J. C., Macedo, M., Hartmann, D., Lind, C., Frank, D., ... & Ulrich, H. (2024). Assessing Patient Health Dynamics by Comparative CT Analysis: An Automatic Approach to Organ and Body Feature Evaluation. Diagnostics, 14(23), 2760.

[29] Chen, Q., Bernard, M. E., Duan, J., & Feng, X. (2021). A transfer learning approach for improving OAR segmentation in the adaptive therapy or retreatment of head and neck cancer. International Journal of Radiation Oncology, Biology, Physics, 111(3), e125-e126.

[30] Raza, A., Khan, M. U., Saeed, Z., Samer, S., Mobeen, A., & Samer, A. (2021, December). Classification of eye diseases and detection of cataract using digital fundus imaging (DFI) and inception-V4 deep learning model. In 2021 International Conference on Frontiers of Information Technology (FIT) (pp. 137-142). IEEE.

[31] Saeed, Z., Khan, M. U., Raza, A., Khan, H., Javed, J., & Arshad, A. (2021, October). Classification of pulmonary viruses X-ray and detection of COVID-19 based on invariant of inception-V 3 deep learning model. In 2021 International Conference on Computing, Electronic and Electrical Engineering (ICE Cube) (pp. 1-6). IEEE.

[32] Khan, M. U., Abbasi, M. A., Saeed, Z., Asif, M., Raza, A., & Urooj, U. (2021, December). Deep learning based intelligent emotion recognition and classification system. In 2021 International Conference on Frontiers of Information Technology (FIT) (pp. 25-30). IEEE.

[33] Saeed, Z., Bouhali, O., Ji, J. X., Hammoud, R., Al-Hammadi, N., Aouadi, S., & Torfeh, T. (2024). Cancerous and non-cancerous MRI classification using dual DCNN approach. Bioengineering, 11(5), 410

[34] Khan, M. U., Saeed, Z., Raza, A., Abbasi, Z., Ali, S. Z. E. Z., & Khan, H. (2022). Deep Learning-based Decision Support System for classification of COVID-19 and Pneumonia patients. JAREE (Journal on Advanced Research in Electrical Engineering), 6(1).

[35] Naqvi, S. Z. H., Khan, M. U., Raza, A., Saeed, Z., Abbasi, Z., & Ali, S. Z. E. Z. (2021, November). Deep Learning Based Intelligent Classification of COVID-19 & Pneumonia Using Cough Auscultations. In 2021 6th International Multi-Topic ICT Conference (IMTIC) (pp. 1-6). IEEE.

[36] Saeed, Z., Torfeh, T., Aouadi, S., Ji, X., & Bouhali, O. (2024). An efficient ensemble approach for brain tumors classification using magnetic resonance imaging. Information, 15(10), 641

[37] Saeed, Z., Torfeh, T., Aouadi, S., Ji, X., & Bouhali, O. (2024). An efficient ensemble approach for brain tumors classification using magnetic resonance imaging. Information, 15(10), 641.

[38] Nawaz, U., Anees-ur-Rahaman, M., & Saeed, Z. (2025). A review of neuro-symbolic AI integrating reasoning and learning for advanced cognitive systems. Intelligent Systems with Applications, 200541.

[39] Saeed, Z., Raza, A., Qureshi, A. H., & Yousaf, M. H. (2021, October). A multi-crop disease detection and classification approach using cnn. In 2021 International Conference on Robotics and Automation in Industry (ICRAI) (pp. 1-6). IEEE.

[40] Saeed, Z., Khan, M. U., Raza, A., Sajjad, N., Naz, S., & Salal, A. (2021, December). Identification of leaf diseases in potato crop using Deep Convolutional Neural Networks (DCNNs). In 2021 16th International conference on emerging technologies (icet) (pp. 1-6). IEEE.

[41] Saeed, Z., Yousaf, M. H., Ahmed, R., Velastin, S. A., & Viriri, S. (2023). On-board small-scale object detection for unmanned aerial vehicles (UAVs). Drones, 7(5), 310.

[42] Ishtiaq, A., Saeed, Z., Khan, M. U., Samer, A., Shabbir, M., & Ahmad, W. (2022). Fall detection, wearable sensors & artificial intelligence: A short review. JAREE (Journal on Advanced Research in Electrical Engineering), 6(2).

[43] Raza, A., Saeed, Z., Aslam, A., Nizami, S. M., Habib, K., & Malik, A. N. (2024, February). Advances, application and challenges of lithography techniques. In 2024 5th International Conference on Advancements in Computational Sciences (ICACS) (pp. 1-6). IEEE.

[44] Saeed, Z., Awan, M. N. M., & Yousaf, M. H. (2022, November). A Robust Approach for Small-Scale Object Detection From Aerial-View. In 2022 International Conference on Digital Image Computing: Techniques and Applications (DICTA) (pp. 1-7). IEEE.

[45] Nawaz, U., Saeed, Z., & Atif, K. (2025). A novel framework for efficient dominance-based rough set approximations using K-dimensional (KD) tree partitioning and adaptive recalculations techniques. Engineering Applications of Artificial Intelligence, 154, 110993.

[46] Nawaz, U., Saeed, Z., & Atif, K. (2025). A Novel Transformer-based approach for adult’s facial emotion recognition. IEEE Access.