Vision-LSTM模型在甲状腺影像报告与数据系统4b类甲状腺结节超声影像诊断中的应用与评估

doi:10.6040/j.issn.1671-7554.0.2025.0696

摘要/Abstract

摘要： 目的探讨基于Vision-LSTM的人工智能(artificial intelligence, AI)技术对甲状腺影像报告与数据系统 4b(Thyroid Imaging Reporting and Data System Category 4b, TI-RADS 4b)类甲状腺结节的超声诊断准确性,评估其辅助临床决策的可行性。方法收集我院401例TI-RADS 4b类甲状腺结节的超声影像数据,并利用这些数据对Vision-LSTM模型进行训练和验证。将AI模型的诊断结果与初级医生及高级医生的诊断结果进行对比,评估其在诊断准确性、稳定性等方面的表现;采用曲线下面积(area under the curve, AUC)、精确率-召回率(precision-recall, PR)曲线等指标对模型性能进行量化分析。结果在独立验证中,Vision-LSTM模型的AUC(0.88)与准确率(89.4%)均显著高于初级医生(AUC: 0.624),并达到与高级医生(AUC: 0.787)相当的水平,证明了其辅助诊断的应用潜力。AI模型能够准确识别超声影像中的复杂特征,稳定输出一致的诊断结果,展现出较高的准确性和可靠性。结论基于Vision-LSTM模型的AI技术可显著提升TI-RADS 4b类甲状腺结节的诊断效率与准确性,为医生提供有效辅助,减轻工作负担。

关键词: 甲状腺影像报告与数据系统, 甲状腺结节, Vision-LSTM模型, 诊断准确性, 人工智能

Abstract: Objective To evaluate the diagnostic accuracy and clinical utility of a Vision-LSTM-based artificial intelligence(AI)model in classifying Thyroid Imaging Reporting and Data System Category 4b(TI-RADS 4b)thyroid nodules on ultrasound. Methods This study utilized ultrasound imaging data from 401 TI-RADS 4b thyroid nodules. A Vision-LSTM model was developed and validated. The AIs diagnostic performance was compared against that of junior and senior physicians using key metrics, including the area under the curve(AUC)and the precision-recall curve(PRC). Results On an independent validation set, the Vision-LSTM model achieved an AUC of 0.88 and an accuracy of 89.4%, significantly outperforming junior physicians(AUC: 0.624)and performing on par with senior physicians(AUC: 0.787). The model demonstrated a high capability for identifying complex sonographic features and delivering consistent diagnostic outcomes.The AI model was able to accurately identify the complex features in ultrasound images, and consistently produced consistent diagnostic results, demonstrating a high degree of accuracy and reliability. Conclusion The Vision-LSTM-based AI model significantly improves the diagnostic efficiency and accuracy for TI-RADS 4b thyroid nodules, showing great potential as an effective tool to support clinical decision-making and reduce physician workload.

Key words: Thyroid Imaging Reporting and Data System, Thyroid nodules, Vision-LSTM model, Diagnostic accuracy, Artificial intelligence

中图分类号:

R736.1

张鑫茹,李扬,孙萌,聂玮,马喆. Vision-LSTM模型在甲状腺影像报告与数据系统4b类甲状腺结节超声影像诊断中的应用与评估[J]. 山东大学学报 (医学版), 2025, 63(11): 68-74.

ZHANG Xinru, LI Yang, SUN Meng, NIE Wei, MA Zhe. Application and evaluation of Vision-LSTM model in diagnostic ultrasound imaging of Thyroid Imaging Reporting and Data System Category 4b thyroid nodules[J]. Journal of Shandong University (Health Sciences), 2025, 63(11): 68-74.

参考文献

[1] Malhi HS, Grant EG. Ultrasound of thyroid nodules and the thyroid imaging reporting and data system[J]. Neuroimaging Clin N Am, 2021, 31(3): 285-300.
[2] Liang FP, Li X, Ji Q, et al. Revised thyroid imaging reporting and data system(TIRADS): imitating the American college of radiology TIRADS, a single-center retrospective study[J]. Quant Imaging Med Surg, 2023, 13(6): 3862-3872.
[3] Chen ZG, Du Y, Cheng LG, et al. Diagnostic perfor-mance of simplified TI-RADS for malignant thyroid nodules: comparison with 2017 ACR-TI-RADS and 2020 C-TI-RADS[J]. Cancer Imaging, 2022, 22(1): 41. doi: 10.1186/s40644-022-00478-y
[4] 田文, 孙辉. 超声引导下甲状腺结节和颈部淋巴结细针穿刺活检中国专家共识及操作指南(2025版)[J]. 中国实用外科杂志, 2025, 45(1): 34-41. TIAN Wen, SUN Hui. Chinese expert consensus and operational guidelines for ultrasound-guided fine needle aspiration biopsy of thyroid nodules and cervical lymph nodes(2025 edition)[J]. Chinese Journal of Practical Surgery, 2025, 45(1): 34-41.
[5] Velez Torres JM, Vaickus LJ, Kerr DA. Thyroid fine-needle aspiration: the current and future landscape of cytopathology[J]. Surg Pathol Clin, 2024, 17(3): 371-381.
[6] 钟迪, 唐棣, 高小强, 等. 多模态超声与超声引导下细针穿刺抽吸活检鉴别C-TIRADS 4类甲状腺良、恶性结节[J]. 中国医学影像技术, 2024, 40(2): 182-185. ZHONG Di, TANG Di, GAO Xiaoqiang, et al. Differentiation of benign and malignant C-TIRADS category 4 thyroid nodules by multimodal ultrasound and ultrasound-guided fine-needle aspiration biopsy[J]. Chinese Journal of Medical Imaging Technology, 2024, 40(2): 182-185.
[7] 刘燕, 曹广磊, 陈丽. 超声引导下细针穿刺和BRAFV600E分子检测在甲状腺癌诊断中的价值[J]. 山东大学学报(医学版), 2022, 60(10): 57-61. LIU Yan, CAO Guanglei, CHEN Li. Ultrasound-guided fine needle aspiration biopsy and BRAFV600E molecular detection in the diagnosis of thyroid cancer[J]. Journal of Shandong University(Health Science), 2022, 60(10): 57-61.
[8] 徐子良, 郑敏文. 影像人工智能在医学领域的时代创新与挑战[J]. 山东大学学报(医学版), 2023, 61(12): 7-12. XU Ziliang, ZHENG Minwen. Era innovation and challenges of imaging artificial intelligence in the medical field[J]. Journal of Shandong University(Health Sciences), 2023, 61(12): 7-12.
[9] Wang J, Jiang J, Zhang D, et al. An integrated AI model to improve diagnostic accuracy of ultrasound and output known risk features in suspicious thyroid nodules[J]. Eur Radiol, 2022, 32(3): 2120-2129.
[10] Shen YT, Chen L, Yue WW, et al. Artificial intelligence in ultrasound[J]. Eur J Radiol, 2021, 139: 109717. doi: 10.1016/j.ejrad.2021.109717
[11] Donahue J, Hendricks LA, Rohrbach M, et al. Long-term recurrent convolutional networks for visual recognition and description[J]. IEEE Trans Pattern Anal Mach Intell, 2017, 39(4): 677-691.
[12] 陈莉军, 王琳. 影像组学在甲状腺结节定性诊断中的研究进展[J]. 磁共振成像, 2025, 16(2): 165-171. CHEN Lijun, WANG Lin. Research progress of radiomics in the qualitative diagnosis of thyroid nodules[J]. Magnetic Resonance Imaging, 2025, 16(2): 165-171.
[13] 陈冲, 陈俊, 夏黎明. 人工智能促进医学影像临床应用与研究[J]. 放射学实践, 2024, 39(1): 12-16. CHEN Chong, CHEN Jun, XIA Liming. Artificial intelligence promotes clinical application and research in medical imaging[J]. Journal of Radiology Practice, 2024, 39(1): 12-16.
[14] Kumari S, Singh P. Deep learning for unsupervised domain adaptation in medical imaging: recent advancements and future perspectives[J]. Comput Biol Med, 2024, 170: 107912. doi: 10.1016/j.compbiomed.2023.107912
[15] 汪洋, 王永仁, 陈雯, 等. 人工智能在医学影像学辅助诊疗中的发展及应用研究新进展[J]. 影像研究与医学应用, 2024, 8(11): 9-11. WANG Yang, WANG Yongren, CHEN Wen, et al. New progress in the development and application of artificial intelligence in assisted diagnosis and treatment of medical imaging[J]. Journal of Imaging Research and Medical Applications, 2024, 8(11): 9-11.
[16] Giger ML. Computer-aided detection and diagnosis/radiomics/machine learning/deep learning in medical imaging[J]. Med Phys, 2023, 50(supp1): 50-53.
[17] Shafiq M, Gu ZQ. Deep residual learning for image recognition: a survey[J]. Appl Sci, 2022, 12(18): 8972. doi: 10.3390/app12188972
[18] Webb JM, Meixner DD, Adusei SA, et al. Automatic deep learning semantic segmentation of ultrasound thyroid cineclips using recurrent fully convolutional networks[J]. IEEE Access, 2021, 9: 5119-5127. doi: 10.1109/access.2020.3045906
[19] 王玮. 面向医学图像的神经网络模型鲁棒性研究[D]. 贵阳: 贵州大学, 2024.
[20] Ong JCL, Chang SY, William W, et al. Ethical and regulatory challenges of large language models in medicine[J]. Lancet Digit Health, 2024, 6(6): e428-e432.
[21] Belle V, Papantonis I. Principles and practice of explainable machine learning[J]. Front Big Data, 2021, 4: 688969. doi: 10.3389/fdata.2021.688969
[22] 杨军洁, 周程. 医学人工智能的算法黑箱问题: 伦理挑战与化解进路[J]. 科学通报, 2023, 68(13): 1604-1610. YANG Junjie, ZHOU Cheng. Algorithmic black-box problem in medical artificial intelligence: ethical challenges and solution approach[J]. Chinese Science Bulletin, 2023, 68(13): 1604-1610.
[23] Wadden JJ. Defining the undefinable: the black box problem in healthcare artificial intelligence[J]. J Med Ethics, 2021: medethics-medet2021-107529. doi: 10.1136/medethics-2021-107529
[24] Welchowski T, Maloney KO, Mitchell R, et al. Techniques to improve ecological interpretability of black-box machine learning models[J]. J Agric Biol Environ Stat, 2022, 27(1): 175-197.
[25] 肖丽. 人工智能时代的营销伦理问题及其决策模型的重构[D]. 广州: 暨南大学, 2019.
[26] 黄键, 张平, 王志刚, 等. 大模型时代的数据安全与伦理问题研究[J]. 信息安全与通信保密, 2025(3): 46-53. HUANG Jian, ZHANG Ping, WANG Zhigang, et al. Research on data security and ethical issues in the era of large models[J]. Information Security and Communications Privacy, 2025(3): 46-53.
[27] Alom MR, Al Farid F, Rahaman MA, et al. An explainable AI-driven deep neural network for accurate breast cancer detection from histopathological and ultrasound images[J]. Sci Rep, 2025, 15: 17531. doi: 10.1038/s41598-025-97718-5

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed