RTS-Net: thyroid nodule segmentation network integrating dual-path attention and graph convolution
Frontiers in Medicine · 13.04.2026 · 13
DOI: 10.3389/fmed.2026.1785796 ↗Kurz gefasst
Fachartikel in der Zeitschrift Frontiers in Medicine, online erschienen am 13.04.2026. Verfasst von Xiaojie Sun und 11 weiteren Autor:innen. Beteiligte Einrichtung: Department of Health Management, The Sixth Hospital of Shanxi Medical University, Taiyuan, China. Die Arbeit nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Introduction
Thyroid ultrasound is the primary imaging modality for nodule detection, but manual interpretation suffers from subjectivity and inefficiency due to speckle noise, low contrast, and operator dependence. Deep learning-based segmentation methods often overlook anatomical prior information, leading to suboptimal performance on atypical nodules and complex backgrounds.
Methods
We propose RTS-Net, a novel segmentation network that integrates a dual-path attention enhancement mechanism (combining spatial and channel attention) and a cascaded graph convolution decoding architecture to leverage multi-scale feature pyramid fusion. A deep supervision strategy is also employed to accelerate convergence. The model is trained and evaluated on the TN3K, DDTI, and a large-scale clinical dataset.
Results
Extensive experiments demonstrate that RTS-Net achieves superior performance on both in-distribution and cross-dataset settings. On the TN3K dataset, it attains 81.66% F1-score and 71.87% IoU; on the DDTI dataset, it achieves 71.10% F1-score and 60.09% IoU, outperforming state-of-the-art methods including UNet, DeepLabv3+, TransUNet, and recent foundation-model-based approaches. Ablation studies confirm the effectiveness of each proposed component.
Discussion
The proposed dual-path attention and graph convolution modules effectively enhance feature representation and boundary integrity, particularly for small nodules and blurred edges. While RTS-Net shows strong generalization, failure cases reveal challenges in heterogeneous backgrounds and acoustic artifacts, suggesting future integration with foundation models like SAM to further improve robustness.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Health Management, The Sixth Hospital of Shanxi Medical University, Taiyuan, China.
- Department of Health Management, General Hospital of Tisco, Taiyuan, China.
- Department of Interventional Ultrasound, The Fifth People's Hospital of Shanxi Province, Taiyuan, China.
- Department of Computer Science and Technology, Taiyuan University of Technology, Taiyuan, China.
- Department of Ultrasound, Shanxi Province Cancer Hospital / Shanxi Hospital Affiliated to Cancer Hospital, Chinese Academy of Medical Sciences / Cancer Hospital Affiliated to Shanxi Medical University, Taiyuan, China.
- Translational Medicine Research Center, Shanxi Medical University, Taiyuan, China.
Zitieren
APA
Sun, X., Li, X., Yang, Z., Zhang, B., Liu, Y., Liang, H., Guo, W., Liang, Y., Chen, X., Li, X., Wang, B., & Wang, X. (2026). RTS-Net: thyroid nodule segmentation network integrating dual-path attention and graph convolution. Frontiers in Medicine, 13. https://doi.org/10.3389/fmed.2026.1785796
Vancouver
Sun X, Li X, Yang Z, Zhang B, Liu Y, Liang H, et al. RTS-Net: thyroid nodule segmentation network integrating dual-path attention and graph convolution. Frontiers in Medicine. 2026;13. doi:10.3389/fmed.2026.1785796
BibTeX
@article{Sun2026,
author = {Sun, Xiaojie and Li, Xiaohong and Yang, Zhou and Zhang, Bo and Liu, Yiqian and Liang, Heng and Guo, Wanli and Liang, Yiping and Chen, Xianchuan and Li, Xiuhui and Wang, Bin and Wang, Xinghua},
title = {RTS-Net: thyroid nodule segmentation network integrating dual-path attention and graph convolution},
journal = {Frontiers in Medicine},
year = {2026},
volume = {13},
doi = {10.3389/fmed.2026.1785796}
}
EUVIMED