深度学习在肌骨超声中的研究进展

刘伟杰 ,  葛春璐 ,  闫玉梅 ,  李姝君 ,  吕欣阳

大连医科大学学报 ›› 2026, Vol. 48 ›› Issue (3) : 252 -258.

PDF (588KB)
大连医科大学学报 ›› 2026, Vol. 48 ›› Issue (3) : 252 -258. DOI: 10.11724/jdmu.2026.03.09
综述

深度学习在肌骨超声中的研究进展

作者信息 +

Research progress of deep learning in musculoskeletal ultrasound

Author information +
文章历史 +
PDF (601K)

摘要

肌骨超声在肌肉骨骼系统疾病诊疗中发挥着重要作用,但其诊断效能受限于操作者经验与主观视觉评估,而深度学习技术通过图像高维特征的自动提取、像素级分割及目标智能检测,有效降低了对主观经验的依赖,为肌骨超声的客观化、定量化分析提供了重要技术支撑。本文充分阐述深度学习的基本原理,并重点围绕骨骼、关节与软骨、骨骼肌与肌腱、外周神经等关键解剖结构,详细评述深度学习模型在病灶自动分割、目标识别、疾病分类与参数定量测量中展现出的优异效能,有效促进肌肉骨骼疾病诊疗的精准化与智能化。

Abstract

Musculoskeletal ultrasound plays a crucial role in the diagnosis and treatment of musculoskeletal system diseases. However, its diagnostic performance is limited by operator's experience and subjective visual assessment. Deep learning technology, through the automatic extraction of high-dimensional image features, pixel-level segmentation, and intelligent detection of targets, effectively reduces dependence on subjective experience and provides important technical support for the Objective and quantitative analysis of musculoskeletal ultrasound. This article comprehensively describes the basic principles of deep learning and focuses on key anatomical structures, such as bones, joints and cartilage, muscles and tendons, peripheral nerves, etc., to provide a detailed review of the excellent performance of deep learning models in automatic lesion segmentation, target recognition, disease classification, and parameter quantitative measurement, effectively promoting the precision and intelligence of musculoskeletal disease diagnosis and treatment. Finally, this article looks forward to future research directions based on existing limitations.

关键词

人工智能 / 超声检查 / 肌肉骨骼 / 关节 / 神经

Key words

artificial intelligence / ultrasonography / musculoskeletal / joints / nerve

引用本文

引用格式 ▾
刘伟杰,葛春璐,闫玉梅,李姝君,吕欣阳. 深度学习在肌骨超声中的研究进展[J]. 大连医科大学学报, 2026, 48(3): 252-258 DOI:10.11724/jdmu.2026.03.09

登录浏览全文

4963

注册一个新账户 忘记密码

参考文献

[1]

Rana M, Bhushan M. Machine learning and deep learning ap-proach for medical image analysis: diagnosis to detection[J]. Multimed Tools Appl, 2023, 82(17): 26731-26769. DOI: 10.1007/s11042-022-14305-w.

[2]

刘艳东, 程晓光, 于爱红. 基于深度学习的人工智能在肌骨影像中的研究进展[J]. 医学影像学杂志, 2022, 32(2): 339-341.

[3]

Yi J, Kang HK, Kwon JH, et al. Technology trends and applica-tions of deep learning in ultrasonography: image quality en-hancement, diagnostic support, and improving workflow effi-ciency[J]. Ultrasonography, 2021, 40(1): 7-22. DOI: 10.14366/usg.20102.

[4]

Lin TM, Lee HY, Chang CK, et al. Identification of tophi in ul-trasound imaging based on transfer learning and clinical prac-tice[J]. Sci Rep, 2023, 13: 12507. DOI: 10.1038/s41598-023-39508-5.

[5]

Shen YT, Chen L, Yue WW, et al. Artificial intelligence in ultra-sound[J]. Eur J Radiol, 2021, 139: 109717. DOI: 10.1016/j.ejrad.2021.109717.

[6]

Zhu FC, Liu MY, et al. Automatic measurement of fetal femur length in ultrasound images: a comparison of random forest re-gression model and SegNet[J]. Math Biosci Eng, 2021, 18(6): 7790-7805. DOI: 10.3934/mbe.2021387.

[7]

袁琳, 张雨, 丁炎, 等. 基于优化 U-Net 神经网络模型在乳腺肿瘤超声图像分割中的应用[J]. 医学影像学杂志, 2023, 33(6): 1081-1085. DOI: 10.20258/j.cnki.1006-9011.2023.06.033.

[8]

Tang SY, Yang X, Shajudeen P, et al. A CNN-based method to reconstruct 3-D spine surfaces from US images in vivo [J]. Med Image Anal, 2021, 74: 102221. DOI: 10.1016/j.media.2021.102221.

[9]

Lyu J, Bi XJ, Banerjee S, et al. Dual-task ultrasound spine trans-verse vertebrae segmentation network with contour regulariza-tion[J]. Comput Med Imaging Graph, 2021, 89: 101896. DOI: 10.1016/j.compmedimag.2021.101896.

[10]

Wu M, Wu H, Wu LL, et al. A deep learning classification of metacarpophalangeal joints synovial proliferation in rheuma-toid arthritis by ultrasound images[J]. J Clin Ultrasound, 2022, 50(2): 296-301. DOI: 10.1002/jcu.23143.

[11]

Yi J, Shin Y, Hahn S, et al. Deep learning based sarcopenia pre-diction from shear-wave ultrasonographic elastography and gray scale ultrasonography of rectus femoris muscle[J]. Sci Rep, 2022, 12: 3596. DOI: 10.1038/s41598-022-07683-6.

[12]

Pandey PU, Quader N, Guy P, et al. Ultrasound bone segmenta-tion: a scoping review of techniques and validation practices[J]. Ultrasound Med Biol, 2020, 46(4): 921-935. DOI: 10.1016/j.ul-trasmedbio.2019.12.014.

[13]

Jin C, Wang SR, Yang GD, et al. A review of the methods on Cobb angle measurements for spinal curvature[J]. Sensors, 2022, 22(9): 3258. DOI: 10.3390/s22093258.

[14]

Wong J, Reformat M, Parent E, et al. Convolutional neural net-work to segment laminae on 3D ultrasound spinal images to as-sist Cobb angle measurement[J]. Ann Biomed Eng, 2022, 50(4): 401-412. DOI: 10.1007/s10439-022-02925-0.

[15]

Chen YP, Fan TY, Chu CC, et al. Automatic and human level Graf's type identification for detecting developmental dysplasia of the hip[J]. Biomed J, 2024, 47(2): 100614. DOI: 10.1016/j.bj.2023.100614.

[16]

Huang BX, Xia B, Qian JK, et al. Artificial intelligence-assisted ultrasound diagnosis on infant developmental dysplasia of the hip under constrained computational resources[J]. J Ultra-sound Med, 2023, 42(6): 1235-1248. DOI: 10.1002/jum.16133.

[17]

Oelen D, Kaiser P, Baumann T, et al. Accuracy of trained physi-cians is inferior to deep learning-based algorithm for determin-ing angles in ultrasound of the newborn hip[J]. Ultraschall Med, 2022, 43(5): e49-e55. DOI: 10.1055/a-1177-0480.

[18]

Sezer A, Sezer HB. Deep convolutional neural network-based automatic classification of neonatal hip ultrasound images: a novel data augmentation approach with speckle noise reduc-tion[J]. Ultrasound Med Biol, 2020, 46(3): 735-749. DOI: 10.1016/j.ultrasmedbio.2019.09.018.

[19]

Li RX, Davoodi A, Cai YY, et al. Robot-assisted ultrasound re-construction for spine surgery: from bench-top to pre-clinical study[J]. Int J Comput Assist Radiol Surg, 2023, 18(9): 1613-1623. DOI: 10.1007/s11548-023-02932-z.

[20]

Gueziri HE, Santaguida C, Collins DL. The state-of-the-art in ultrasound-guided spine interventions[J]. Med Image Anal, 2020, 65: 101769. DOI: 10.1016/j.media.2020.101769.

[21]

Zhao YP, Zheng SQ, Cai N, et al. Utility of artificial intelligence for real-time anatomical landmark identification in ultrasound-guided thoracic paravertebral block[J]. J Digit Imag, 2023, 36(5): 2051-2059. DOI: 10.1007/s10278-023-00851-8.

[22]

Tang J, Jin ZB, Zhou X,et al. Enhancing convolutional neural network scheme for rheumatoid arthritis grading with limited clinical data[J]. Chin Phys B, 2019, 28(3): 038701. DOI: 10.1088/1674-1056/28/3/038701.

[23]

Hemalatha RJ, Vijaybaskar V, Thamizhvani TR. Automatic lo-calization of anatomical regions in medical ultrasound images of rheumatoid arthritis using deep learning[J]. Proc Inst Mech Eng H, 2019, 233(6): 657-667. DOI: 10.1177/0954411919845747.

[24]

He XL, Wang M, Zhao CY, et al. Deep learning-based automat-ic scoring models for the disease activity of rheumatoid arthri-tis based on multimodal ultrasound images[J]. Rheumatology (Oxford), 2024, 63(3): 866-873. DOI: 10.1093/rheumatology/kead366.

[25]

Fiorentino MC, Cipolletta E, Filippucci E, et al. A deep-learn-ing framework for metacarpal-head cartilage-thickness estima-tion in ultrasound rheumatological images[J]. Comput Biol Med, 2022, 141: 105117. DOI: 10.1016/j.compbiomed.2021.105117.

[26]

Hossain MB, Lai KW, Pingguan-Murphy B, et al. Contrast en-hancement of ultrasound imaging of the knee joint cartilage for early detection of knee osteoarthritis[J]. Biomed Signal Process Control, 2014, 13: 157-167. DOI: 10.1016/j.bspc.2014.04.008.

[27]

du Toit C, Orlando N, Papernick S, et al. Automatic femoral ar-ticular cartilage segmentation using deep learning in three-di-mensional ultrasound images of the knee[J]. Osteoarthr Cartil Open, 2022, 4(3): 100290. DOI: 10.1016/j.ocarto.2022.100290.

[28]

Antico M, Sasazawa F, Dunnhofer M, et al. Deep learning-based femoral cartilage automatic segmentation in ultrasound imag-ing for guidance in robotic knee arthroscopy[J]. Ultrasound Med Biol, 2020, 46(2): 422-435. DOI: 10.1016/j.ultrasmedbio.2019.10.015.

[29]

Burlina P, Billings S, Joshi N, et al. Automated diagnosis of myositis from muscle ultrasound: exploring the use of machine learning and deep learning methods[J]. PLoS One, 2017, 12(8): e0184059. DOI: 10.1371/journal.pone.0184059.

[30]

Cunningham RJ, Loram ID. Estimation of absolute states of hu-man skeletal muscle via standard B-mode ultrasound imaging and deep convolutional neural networks [J]. J R Soc Interface, 2020, 17(162): 20190715. DOI: 10.1098/rsif.2019.0715.

[31]

Saleh A, Laradji IH, Lammie C, et al. A deep learning localiza-tion method for measuring abdominal muscle dimensions in ul-trasound images[J]. IEEE J Biomed Health Inform, 2021, 25(10): 3865-3873. DOI: 10.1109/JBHI.2021.3085019.

[32]

Ritsche P, Wirth P, Cronin NJ, et al. DeepACSA: automatic seg-mentation of cross-sectional area in ultrasound images of lower limb muscles using deep learning[J]. Med Sci Phys Exerc, 2022, 54(12): 2188-2195. DOI: 10.1249/mss.0000000000003010.

[33]

Nhat PTH, Van Hao N, Yen LM, et al. Clinical evaluation of AI-assisted muscle ultrasound for monitoring muscle wasting in ICU patients[J]. Sci Rep, 2024, 14: 14798. DOI: 10.1038/s41598-024-64564-w.

[34]

Chiu PH, Boudier-Revéret M, Chang SW, et al. Deep learning for detecting supraspinatus calcific tendinopathy on ultrasound images[J]. J Med Ultrasound, 2022, 30(3): 196-202. DOI: 10.4103/jmu.jmu_182_21.

[35]

Kuok CP, Yang TH, Tsai BS, et al. Segmentation of finger ten-don and synovial sheath in ultrasound image using deep convo-lutional neural network[J]. BioMedical Eng OnLine, 2020, 19(1): 24. DOI: 10.1186/s12938-020-00768-1.

[36]

Smerilli G, Cipolletta E, Sartini G, et al. Development of a con-volutional neural network for the identification and the mea-surement of the Median nerve on ultrasound images acquired at carpal tunnel level[J]. Arthritis Res Ther, 2022, 24(1): 38. DOI: 10.1186/s13075-022-02729-6.

[37]

Festen RT, Schrier VJMM, Amadio PC. Automated segmenta-tion of the Median nerve in the carpal tunnel using U-net[J]. Ultrasound Med Biol, 2021, 47(7): 1964-1969. DOI: 10.1016/j.ultrasmedbio.2021.03.018.

[38]

Shinohara I, Inui A, Mifune Y, et al. Diagnosis of cubital tunnel syndrome using deep learning on ultrasonographic images[J]. Diagnostics, 2022, 12(3): 632. DOI: 10.3390/diagnostics12030632.

[39]

Berggreen J, Johansson A, Jahr J, et al. Deep learning on ultra-sound images visualizes the femoral nerve with good precision[J]. Healthcare (Basel), 2023, 11(2): 184. DOI: 10.3390/healthcare11020184.

[40]

Huang CY, Zhou Y, Tan WL, et al. Applying deep learning in recognizing the femoral nerve block region on ultrasound im-ages[J]. Ann Transl Med, 2019, 7(18): 453. DOI: 10.21037/atm.2019.08.61.

[41]

Gungor I, Gunaydin B, Oktar SO, et al. A real-time anatomy ıdentification via tool based on artificial ıntelligence for ultra-sound-guided peripheral nerve block procedures: an accuracy study [J]. J Anesth, 2021, 35(4): 591-594. DOI: 10.1007/s00540-021-02947-3.

基金资助

辽宁省研究生教育教学改革研究项目(LNYJG2024223)

AI Summary AI Mindmap
PDF (588KB)

0

访问

0

被引

详细

导航
相关文章

AI思维导图

/

〈 〉