摘要
在血管外科领域,血管介入手术是一种高效的微创治疗方法。然而,传统的介入手术方法需要医生长时间穿戴防护铅衣,并且存在暴露于辐射的风险,这不仅为医生带来健康隐患,还可能对手术效率产生不利影响。随着血管腔内介入手术机器人(EIR)的研发应用,可在减少医生辐射暴露的同时,提供相较于传统方法更高的操作精度和稳定性。介入器械的自动识别和实时跟踪两大关键技术可以使得EIR在复杂的介入手术中拥有对器械方位的把控能力和判断能力,从而保证治疗的质量和安全。目前研究大多倾向利用介入影像实现EIR在血管结构中实时、准确地检测和定位器械,即令机器人观察图像信息进行手术。与此同时,介入器械识别技术与EIR的协同工作具有重大潜在价值。这要求机器人不仅要能精准执行命令,更需能够理解和预测医生的操作意图。随着人工智能技术的发展,有望在辅助机器人更精确地识别和跟踪器械并修改定位误差方面提供支持,实现真正的协同手术。在此,笔者分析介入器械的自动识别和实时跟踪两大技术在EIR领域的应用情况,综合讨论它们的临床应用前景,并对国内外相关技术的发展进行对比和总结。
2023年6月发布的《中国心血管健康与疾病报告2022》指出,推算我国心血管病现患人数约3.3亿,城乡居民疾病死亡构成比中,心血管病占首位,心血管病发病率仍处于持续上升阶
介入器械识别是指在医学影像中突出显示并定位术中介入器械的能力。借助这一自动化技术,EIR得以在连续或单一的医学影像中准确地识别并区别各种器械。尤其当应用于X线影像时,该技术可以对介入器械的近端和躯干进行高精度的视觉定位和可视化。
在数字化医学影像技术普及之前,研究者通过改进器械材料和开发专门的识别设备来实现器械识别。根据其工作机制,分为主动识别和被动识别:⑴ 主动识别依赖于器械内部的反馈元件,由器械内部发出信号,然后被设备接收成像。Collins
数字成像技术进步带动了视觉图像识别在医学领域的发展。这项技术利用AI视觉算法在介入影像中直接识别器械,省去对传统器械特殊改装的需求,更便于融入常规手术流程。Useche Murillo
在EIR手术中,提高手术的准确度与实时性至关重要。为了达到这一目标,混合模式识别受到了广泛的关注。它不仅是单一技术的简单应用,而是多种技术的融合,旨在为器械定位提供最高的精确度。通过结合二维和三维图像信息的配准方法已被提
形状检测专注于精确获取介入器械近端的姿态及其朝向,用于在手术中选择入路血管。通过自动识别与形态检测,EIR能够获取更多关于器械近端的反馈信息,为医生和EIR本身在执行手术时提供更加准确的指导和支持,从而确保手术的精准性和效果。
通过将FBG传感器嵌入到介入器械中,可以有效估测血管腔的内部形状。形状测定的过程如下:⑴ 利用FBG传感器测量应变;⑵ 依据测得的应变计算曲率和扭矩;⑶ 参照纤维的数据,推断纤维中心的曲率和扭矩;⑷ 根据Frenet-Serret方程(描述曲线在三维空间中运动的计算公式),计算中心曲率和扭矩来确定器械的具体形
器械力形变检测本质上是一种基于力传感器的改装结构。这一结构由多个灵活的力传感器组成,嵌入介入器械内部,特别集中于器械远端和血管腔内组织相互作用的弯曲
图像形状检测提供了一种从视觉角度评估器械姿态的途径。这种方法依托AI技术,特别是利用基于CNN的U-net分割算法,它能够从介入影像中分析介入器械的近端姿态,并准确获取形状与位
腔内器械追踪是指在成功识别器械后,能够确定和追踪介入器械关键位置,并反馈位置信息的技术。追踪的主要目的在于提供器械近端在血管腔内的具体位置和方向,以实时更新当前介入器械的位置,进而输出这些位置信息以供EIR使用。根据这些位置信息,可以实现图像引导下的EIR手术导
在EIR手术中,光学追踪是一种传统追踪技术,它能通过监控手术设备的物理位置和方向实时地更新并提供这些信息。这一技术通常使用高精度的光学摄像头,配合一种或多种被附加在手术设备上的反光标记或红外发射器,这些标记或发射器可以发出特定频率的光线,被光学摄像头捕捉并解析,从而在三维空间中追踪和定位器械的运动。Langsch
电磁跟踪利用磁场控制和引导介入器械至病灶位置。使用这种追踪方式需配备额外硬件,如场发生器、电磁传感器和跟踪模块等。在Rogers
基于图像的器具近端跟踪技术,类似于图像视觉识别,避免了对额外硬件设备的需求,大幅提高EIR操作的便捷性和灵活性。当前,研究者们正倾向于构建适用于介入器械追踪的通用跟踪方法,需要预先收集的图像样本进行模型训练。Peng
在EIR实施介入器械的识别和跟踪过程中,应充分考虑实现的复杂性、成本、医师培训以及在临床环境中的安全性、实用性和适应性。对于识别技术而言,反馈式识别和混合式识别均需额外硬件支持,基于图像信息识别虽无需额外硬件或改变器械内部结构,但仍面临器械在图像中被遮挡的技术挑
综上所述,腔内介入器械的识别和跟踪技术可以为临床医师和EIR提供自动、精准且实时的手术视野,使手术过程更加精确和安全。为了使EIR更好地应用于临床手术,识别和跟踪技术需要具备强大的泛化手术能力,并结合术前其他影像学信息。未来的技术趋势可能更偏向基于图像视觉技术,构建端到端的介入器械图像识别和跟踪系统,以提升手术质量。研究如何能够避免事前采集大量数据集的无监督训练方法,以及在不改变精度和实时性能前提下建立手术预警机制或许是处理视觉算法的AI未来的方向。此外,需要建立能使EIR与医师有效协作的机制,这可能需要设计更直观和人性化的器械交互界面,让医师能快速、精确地了解手术过程中的实时信息,并将识别模型的结果与医师的经验和判断相结合,实现更精确、高效的手术操作。应用于EIR手术中的识别与追踪技术各有挑战和发展趋势,未来技术的更新一定要以临床为出发点,借鉴各种技术,继续克服现有的挑战,推动技术创新和临床应用的发展。
作者贡献声明
伍尚至参与本文设计实验和参考论文数据采集,起草论文和统计分析;陆清声参与对文章的知识性内容作批评性审阅及指导。
利益冲突
所有作者均声明不存在利益冲突。
参考文献
马丽媛, 王增武, 樊静, 等. 《中国心血管健康与疾病报告2022》要点解读[J]. 中国全科医学, 2023, 26(32):3975-3994. doi: 10.12114/j.issn.1007-9572.2023.0408. [百度学术]
Ma LY, Wang ZW, Fan J, et al. Interpretation of report on cardiovascular health and diseases in China 2022[J]. Chinese General Practice, 2023, 26(32):3975-3994. doi: 10.12114/j.issn.1007-9572.2023.0408. [百度学术]
Nicholls M. Mark Nicholls speaks to Professor Eric Durand and Professor Rémi Sabatier about Europe's first remote robotic-assisted angioplasty procedure[J]. Eur Heart J, 2021, 42(32):3033-3035. doi: 10.1093/eurheartj/ehab085. [百度学术]
陈政, 沈毓, 陆清声. 机器人辅助血管介入治疗研究进展[J]. 介入放射学杂志, 2018, 27(1):1-4. doi: 10.3969/j.issn.1008-794X.2018.01.001. [百度学术]
Chen Z, Shen Y, Lu QS. Research progress in robot-assisted vascular interventional therapy[J]. Journal of Interventional Radiology, 2018, 27(1):1-4. doi: 10.3969/j.issn.1008-794X.2018.01.001. [百度学术]
童静, 储呈晨, 李斌. 血管介入手术机器人及其力反馈技术研究进展[J]. 中国普通外科杂志, 2023, 32(6):915-922. doi: 10.7659/j.issn.1005-6947.2023.06.013. [百度学术]
Tong J, Chu CC, Li B. Vascular interventional surgery robot and its force feedback technology[J]. China Journal of General Surgery, 2023, 32(6):915-922. doi: 10.7659/j.issn.1005-6947.2023.06.013. [百度学术]
何昊, 叶子健, 舒畅. 血管介入手术机器人系统关键技术及研发现状[J]. 中国普通外科杂志, 2021, 30(12):1477-1484. doi: 10.7659/j.issn.1005-6947.2021.12.012. [百度学术]
He H, Ye ZJ, Shu C. Essential techniques and current research situation in robotic endovascular systems[J]. China Journal of General Surgery, 2021, 30(12):1477-1484. doi: 10.7659/j.issn.1005-6947.2021.12.012. [百度学术]
Collins GC, Brumfiel TA, Bercu ZL, et al. Dual-resonance (16/32 MHz) piezoelectric transducer with a single electrical connection for forward-viewing robotic guidewire[J]. IEEE Trans Ultrason Ferroelectr Freq Control, 2022, 69(4):1428-1441. doi: 10.1109/TUFFC.2022.3150746. [百度学术]
Chen XY, Zhang YN, Shen LY, et al. Fabrication and shape detection of a catheter using fiber Bragg grating[J]. Adv Manuf, 2020, 8(1):107-118. doi: 10.1007/s40436-019-00284-z. [百度学术]
Al-Ahmad O, Ourak M, Vlekken J, et al. Local one-dimensional motion estimation using FBG-based shape sensing for cardiac applications[J]. IEEE Robotics Autom Lett, 2022, 7(3):8122-8129. doi: 10.1109/LRA.2022.3186761. [百度学术]
Anfinogentov V, Karimov K, Kuznetsov A, et al. Algorithm of FBG spectrum distortion correction for optical spectra analyzers with CCD elements[J]. Sensors (Basel), 2021, 21(8):2817. doi: 10.3390/s21082817. [百度学术]
Ramadani A, Maier H, Bourier F, et al. Feature-based electromagnetic tracking registration using bioelectric sensing[J]. IEEE Robotics Autom Lett, 2023, 8(6):3286-3293. doi: 10.1109/LRA.2023.3262988. [百度学术]
Piskarev Y, Shintake J, Chautems C, et al. A variable stiffness magnetic catheter made of a conductive phase-change polymer for minimally invasive surgery[J]. Adv Funct Mater, 2022, 32(20):2107662. doi: 10.1002/adfm.202107662. [百度学术]
Magnusson P, Johansson E, Månsson S, et al. Passive catheter tracking during interventional MRI using hyperpolarized 13C[J]. Magn Reson Med, 2007, 57(6):1140-1147. doi: 10.1002/mrm.21239. [百度学术]
Wegner F, Lüdtke-Buzug K, Cremers S, et al. Bimodal interventional instrument markers for magnetic particle imaging and magnetic resonance imaging-a proof-of-concept study[J]. Nanomaterials (Basel), 2022, 12(10):1758. doi: 10.3390/nano12101758. [百度学术]
Lu P, Xia J, Li ZC, et al. A vessel segmentation method for multi-modality angiographic images based on multi-scale filtering and statistical models[J]. Biomed Eng Online, 2016, 15(1):120. doi: 10.1186/s12938-016-0241-7. [百度学术]
Useche Murillo PC, Moreno RJ, Pinzon Arenas JO. Comparison between CNN and Haar classifiers for surgical instrumentation classification[J]. Ces, 2017, 10(28):1351-1363. doi: 10.12988/ces.2017.711157. [百度学术]
Li RQ, Xie XL, Zhou XH, et al. A unified framework for multi-guidewire endpoint localization in fluoroscopy images[J]. IEEE Trans Biomed Eng, 2022, 69(4):1406-1416. doi: 10.1109/TBME.2021.3118001. [百度学术]
Song HS, Yi BJ, Won JY, et al. Learning-based catheter and guidewire-driven autonomous vascular intervention robotic system for reduced repulsive force[J]. J Comput Des Eng, 2022, 9(5):1549-1564. doi: 10.1093/jcde/qwac074. [百度学术]
Jiang X, Zeng Y, Xiao S, et al. Automatic detection of coronary metallic stent struts based on YOLOv3 and R-FCN[J]. Comput Math Methods Med, 2020, 2020:1793517. doi: 10.1155/2020/1793517. [百度学术]
Zhao Y, Xing H, Guo S, et al. A novel noncontact detection method of surgeon's operation for a master-slave endovascular surgery robot[J]. Med Biol Eng Comput, 2020, 58(4):871-885. doi: 10.1007/s11517-020-02143-7. [百度学术]
Vlontzos A, Mikolajczyk K. Deep segmentation and registration in x-ray angiography video[DB]. arXiv preprint arXiv, 2018, (1):1-8. doi:10.48550/arXiv.1805.06406. [百度学术]
Guan S, Li T, Meng C, et al. Multi-mode information fusion navigation system for robot-assisted vascular interventional surgery[J]. BMC Surg, 2023, 23(1):51. doi: 10.1186/s12893-023-01944-5. [百度学术]
Ma Y, James Housden R, Fazili A, et al. Real-time registration of 3D echo to X-ray fluoroscopy based on cascading classifiers and image registration[J]. Phys Med Biol, 2021, 66(5):055019. doi: 10.1088/1361-6560/abe420. [百度学术]
Vöth T, König T, Eulig E, et al. Real-time 3D reconstruction of multiple guidewires at dose values of conventional 2D fluoroscopy[C]//SPIE Medical Imaging. Proc SPIE 12031, Medical Imaging 2022: Physics of Medical Imaging, San Diego, California, USA. 2022, 12031:405-409. doi: 10.1117/12.2606579. [百度学术]
Abdulhafiz I, Janabi-Sharifi F. A hybrid approach to 3D shape estimation of catheters using ultrasound images[J]. IEEE Robotics Autom. Lett., 2023, 8(4):1912-1919. doi: 10.1109/LRA.2023.3244415. [百度学术]
Peng W, Wu W, Zhang J, et al. An automatic framework for estimating the pose of the catheter distal section using a coarse-to-fine network[J]. Comput Methods Programs Biomed, 2022, 225:107036. doi: 10.1016/j.cmpb.2022.107036. [百度学术]
Zhao Z, Li R, Xu LJ, et al. Angular needle tracker and stabilizer for image-guided interventions[J]. Minim Invasive Ther Allied Technol, 2022, 31(3):410-417. doi: 10.1080/13645706.2020.1832122. [百度学术]
Sahu SK, Sozer C, Rosa B, et al. Shape reconstruction processes for interventional application devices: state of the art, progress, and future directions[J]. Front Robot AI, 2021, 8:758411. doi: 10.3389/frobt.2021.758411. [百度学术]
Wang Y, Lam HK, Hou ZG, et al. High-resolution feature based central venous catheter tip detection network in X-ray images[J]. Med Image Anal, 2023, 88:102876. doi: 10.1016/j.media.2023.102876. [百度学术]
Ha XT, Ourak M, Al-Ahmad O, et al. Robust catheter tracking by fusing electromagnetic tracking, fiber Bragg grating and sparse fluoroscopic images[J]. IEEE Sens J, 2021, 21(20):23422-23434. doi: 10.1109/JSEN.2021.3107036. [百度学术]
Wang K, Wu Z, Wu R, et al. Direct fabrication of flexible strain sensor with adjustable gauge factor on medical catheters[J]. J Sci Adv Mater Devices, 2023, 8(3):100558. doi: 10.1016/j.jsamd.2023.100558. [百度学术]
Han F, He Y, Zhu H, et al. A novel catheter shape-sensing method based on deep learning with a multi-core optical fiber[J]. Sensors (Basel), 2023, 23(16):7243. doi: 10.3390/s23167243. [百度学术]
Lezcano DA, ⅡIordachita, Kim JS. Lie-group theoretic approach to shape-sensing using FBG-sensorized needles including double-layer tissue and S-shape insertions[J]. IEEE Sens J, 2022, 22(22):22232-22243. doi: 10.1109/JSEN.2022.3212209. [百度学术]
李天梁, 宋珍珍, 陈发银, 等. 光纤光栅与人工智能融合的形状自感知穿刺针[J]. 光学精密工程, 2023, 31(2):160-167. doi: 10.37188/OPE.20233102.0160. [百度学术]
Li TL, Song ZZ, Chen FY, et al. Fiber Bragg grating and artificial intelligence fusion for shape self-sensing puncture needle[J]. Optics and Precision Engineering, 2023, 31(2):160-167. doi: 10.37188/OPE.20233102.0160. [百度学术]
Pourafzal M, Talebi A, Rabenorosoa K. Stochastic model-based contact force estimation for concentric tube robots[J]. Mech Mach Theory, 2023, 180:105135. doi: 10.1016/j.mechmachtheory.2022.105135. [百度学术]
Ha XT, Wu D, Lai CF, et al. Contact localization of continuum and flexible robot using data-driven approach[J]. IEEE Robotics Autom Lett, 2022, 7(3):6910-6917. doi: 10.1109/LRA.2022.3176723. [百度学术]
Deaton NJ, Brumfiel TA, Sarma A, et al. Simultaneous shape and tip force sensing for the COAST guidewire robot[J]. IEEE Robotics Autom Lett, 2023, 8(6):3725-3731. doi: 10.1109/LRA.2023.3267008. [百度学术]
Xiang P, Zhang J, Sun D, et al. Learning-based high-precision force estimation and compliant control for small-scale continuum robot[J]. IEEE Trans Autom Sci Eng, 2023, PP(99):1-13. doi: 10.1109/TASE.2023.3311179. [百度学术]
Alkayas AY, Feliu-Talegon D, Mathew AT, et al. Shape and tip force estimation of concentric tube robots based on actuation readings alone[C]//2023 IEEE International Conference on Soft Robotics (RoboSoft). April 3-7, 2023, Singapore, Singapore. IEEE, 2023:1-8. doi: 10.1109/RoboSoft55895.2023.10121920. [百度学术]
任龙飞, 孟偲, 刘博, 等. 基于图像的介入器械前端变形建模与受力估计[J]. 中国图象图形学报, 2022, 27(3):1001-1008. doi: 10.11834/jig.210653. [百度学术]
Ren LF, Meng C, Liu B, et al. Image-based deformation modeling and force estimation of the front end of vascular interventional surgical apparatus[J]. Journal of Image and Graphics, 2022, 27(3):1001-1008. doi: 10.11834/jig.210653. [百度学术]
Kong Y, Wang J, Zhang N, et al. Dexterity analysis and motion optimization of In-situ torsionally-steerable flexible surgical robots[J]. IEEE Robotics Autom Lett, 2022, 7(3):8347-8354. doi: 10.1109/LRA.2022.3183531. [百度学术]
Chaillou P, Shi J, Kruszewski A, et al. Reduced finite element modelling and closed-loop control of pneumatic-driven soft continuum robots[C]//2023 IEEE International Conference on Soft Robotics (RoboSoft). April 3-7, 2023, Singapore, Singapore. IEEE, 2023:1-8. doi: 10.1109/RoboSoft55895.2023.10122081. [百度学术]
Qiao Q, Borghesan G, De Schutter J, et al. Force from shape—estimating the location and magnitude of the external force on flexible instruments[J]. IEEE Trans Robotics, 2021, 37(5):1826-1833. doi: 10.1109/TRO.2021.3062504. [百度学术]
Wei W, Yang D, Li L, et al. An intravascular catheter bending recognition method for interventional surgical robots[J]. Machines, 2022, 10(1):42. doi: 10.3390/machines10010042. [百度学术]
Holm Andersen JK, Schwaner KL, Savarimuthu TR. Real-time segmentation of surgical tools and needle using a mobile-U-net[C]//2021 20th International Conference on Advanced Robotics (ICAR). December 6-10, 2021, Ljubljana, Slovenia. IEEE, 2022:148-154. doi: 10.1109/ICAR53236.2021.9659326. [百度学术]
Gherardini M, Mazomenos E, Menciassi A, et al. Catheter segmentation in X-ray fluoroscopy using synthetic data and transfer learning with light U-nets[J]. Comput Methods Programs Biomed, 2020, 192:105420. doi: 10.1016/j.cmpb.2020.105420. [百度学术]
Zheng Y, Er MJ, Shen S, et al. An improved image segmentation model based on U-net for interventional intravascular robots[C]//2021 4th International Conference on Intelligent Autonomous Systems (ICoIAS). May 14-16, 2021, Wuhan, China. IEEE, 2021:84-90. doi: 10.1109/ICoIAS53694.2021.00023. [百度学术]
Zhou YJ, Xie XL, Zhou XH, et al. Pyramid attention recurrent networks for real-time guidewire segmentation and tracking in intraoperative X-ray fluoroscopy[J]. Comput Med Imaging Graph, 2020, 83:101734. doi: 10.1016/j.compmedimag.2020.101734. [百度学术]
Tsoumakidou G, Saltiel S, Villard N, et al. Image-guided marking techniques in interventional radiology: a review of current evidence[J]. Diagn Interv Imaging, 2021, 102(12):699-707. doi: 10.1016/j.diii.2021.07.002. [百度学术]
Heemeyer F, Choudhary A, Desai JP. Pose-aware C-arm calibration and image distortion correction for guidewire tracking and image reconstruction[C]//2020 International Symposium on Medical Robotics (ISMR). November 18-20, 2020, Atlanta, GA, USA. IEEE, 2021:181-187. doi: 10.1109/ISMR48331.2020.9312944. [百度学术]
陆清声. 血管腔内介入手术机器人的特点和未来发展方向[J]. 机器人外科学杂志: 中英文, 2020, 1(4):231-235. doi: 10.12180/j.issn.2096-7721.2020.04.001. [百度学术]
Lu QS. Characteristics and future development of endovascular intervention robot[J]. Chinese Journal of Robotic Surgery, 2020, 1(4):231-235. doi: 10.12180/j.issn.2096-7721.2020.04.001. [百度学术]
Langsch F, Virga S, Esteban J, et al. Robotic ultrasound for catheter navigation in endovascular procedures[C]//2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). November 3-8, 2019, China. IEEE, 2020:5404-5410. doi: 10.1109/IROS40897.2019.8967652. [百度学术]
Rogers T, Campbell-Washburn AE, Ramasawmy R, et al. Interventional cardiovascular magnetic resonance: state-of-the-art[J]. J Cardiovasc Magn Reson, 2023, 25(1):48. doi: 10.1186/s12968-023-00956-7. [百度学术]
Alexander Jaeger H, O'Donoghue K, Cantillon-Murphy P. Electromagnetic tracking in guided medical interventions[C]//2022 3rd URSI Atlantic and Asia Pacific Radio Science Meeting (AT-AP-RASC). May 30-June 4, 2022, Gran Canaria, Spain. IEEE, 2022:1-4. DOI: 10.23919/at-ap-rasc54737.2022.9814313 [百度学术]
Peng W, Wang Z, Xie H, et al. Design, development and evaluation of an ergonomically designed dual-use mechanism for robot-assisted cardiovascular intervention[J]. Int J Comput Assist Radiol Surg, 2023, 18(2):205-216. doi: 10.1007/s11548-022-02755-4. [百度学术]
Ullah I, Chikontwe P, Park SH. Real-time tracking of guidewire robot tips using deep convolutional neural networks on successive localized frames[J]. IEEE Access, 2019, 7:159743-159753. doi: 10.1109/ACCESS.2019.2950263. [百度学术]
周国敬, 梁世超, 李振锋, 等. 基于图像识别的血管介入器械追踪方法[J]. 北京生物医学工程, 2022, 41(1):90-96. doi: 10.3969/j.issn.1002-3208.2022.01.015. [百度学术]
Zhou GJ, Liang SC, Li ZF, et al. Instrument tracking methods of vascular intervention based on image recognition[J]. Beijing Biomedical Engineering, 2022, 41(1):90-96. doi: 10.3969/j.issn.1002-3208.2022.01.015. [百度学术]
Peng J, Chen Q, Kang L, et al. Autonomous recognition of multiple surgical instruments tips based on arrow OBB-YOLO network[J]. IEEE Trans Instrum Meas, 2022, 71:1-13. doi: 10.1109/TIM.2022.3162596. [百度学术]
Karstensen L, Ritter J, Hatzl J, et al. Learning-based autonomous vascular guidewire navigation without human demonstration in the venous system of a porcine liver[J]. Int J Comput Assist Radiol Surg, 2022, 17(11):2033-2040. doi: 10.1007/s11548-022-02646-8. [百度学术]
Gumbs AA, Grasso V, Bourdel N, et al. The advances in computer vision that are enabling more autonomous actions in surgery: a systematic review of the literature[J]. Sensors (Basel), 2022, 22(13):4918. doi: 10.3390/s22134918. [百度学术]