| | Robot-Assisted Vitreoretinal Surgery: Development of a Prototype and Feasibility Studies in an Animal ModelReceived 27 October 2008; received in revised form 2 February 2009; accepted 3 March 2009. published online 22 June 2009. PurposeTo develop a prototype robotic system designed to assist vitreoretinal surgery and to evaluate its accuracy and maneuverability. DesignExperimental study. ParticipantsThis study used harvested porcine eyes. MethodsAfter development of a prototype robotic system, pointing accuracy tests of the system were performed on graph paper and in harvested porcine eyes. The average maximal deviation from the aiming point to the actual position of the tip of the instrument was compared between manually conducted procedures and those conducted with robotic assistance. The feasibility of creating posterior vitreous detachment (PVD), retinal vessel sheathotomy (RVS), and retinal vessel microcannulation also were evaluated in porcine eye models, and the success rates of 4 consecutive attempts for each kind of procedure were evaluated. Main Outcome MeasuresThe average maximum deviation in pointing accuracy tests both on graph paper and in animal eye models was a main outcome measure. The success rate of making PVD, RVS, and retinal vessel microcannulation was the other primary outcome measure. ResultsThe pointing accuracy was superior with robotic assistance both on graph paper (327.0 μm vs. 32.3 μm) and in animal eye models (140.8 μm vs. 33.5 μm). Creating PVD, RVS, and retinal vessel microcannulation was feasible in 4 of 4 attempts, 4 of 4 attempts, and 2 of 4 attempts, respectively. The 2 failures in microcannulation were considered to be the result of difficulty in visual differentiation between the retinal vessel and retina in harvested porcine eyes. ConclusionsImproved accuracy and desirable feasibility of a prototype robotic system to assist vitreoretinal surgery were shown in this study. Research for wider implementation of robot-assisted surgery should be continued; there are some hurdles to overcome. Financial Disclosure(s)The author(s) have no proprietary or commercial interest in any materials discussed in this article. Available online: June 21, 2009. For approximately a decade, robot-assisted surgery has been used on human patients in various surgical fields.1, 2, 3, 4, 5 Several studies have reported that its performance was as efficient and safe as conventional endoscopic surgeries.3, 6, 7 Generally reported advantages of robot-assisted surgery are (1) improved dexterity and accuracy, (2) rapid learning curve, and (3) telesurgery.5, 8, 9, 10, 11 However, application of robot-assisted surgery has been limited mainly to the field of laparoscopic and thoracoscopic procedures, and to the authors' knowledge, robot-assisted ocular surgery has yet to be performed on a patient. Vitreoretinal surgery, in particular, procedures where manipulations of tissues adjacent to retina of the posterior pole are required, requires some of the most sophisticated skills in ophthalmic surgery. These delicate surgical procedures are candidates for the application of robot-assisted surgery because the improved accuracy and dexterity may lead to better surgical results as well as fewer surgical complications. Moreover, to learn these sophisticated procedures requires surgical experience obtained from a large case volume over a long period. That is, the availability of such surgeons who can perform sophisticated procedures at the desirable level is limited. Robot-assisted surgery can be helpful because more surgeons can learn difficult surgical procedures with less surgical experience; that is, there is a rapid learning curve, which can increase the availability of high-quality surgical service for patients. Finally, telesurgery is an idea with increased availability for difficult surgeries. The aim of this study was to develop a prototype robotic system that performs sophisticated vitreoretinal procedures and to examine its accuracy and feasibility. Materials and Methods  Robotic System The robotic system consisted of the master controller and the slave manipulator set on different tables in the same experiment room. The master and slave communicate through real-time computer systems (operating system, VX Works by Wind River Systems; Alameda, CA) every 10 ms via the local area network. Input to the master system is presented at the slave system, scaled down 40:1 to increase the accuracy of the surgical procedures. An overview of the total system is shown in Figure 1. The micromanipulator of the slave system is designed to move along a pair of spherical guides (α-axis and β-axis), plus an inserting and pulling axis (γ-axis). As shown in Figure 2 (available at http://aaojournal.org), the range of motion of the manipulator is determined by the range of the α-axis, β-axis, and γ-axis. In the current study, those ranges are designed to cover the entire area of the vitreous body below the scleral port plane. The slave system has a structurally fixed point, which is designed to be consistent with a scleral port at the pars plana, through which micromanipulators are introduced into the vitreous body. By this design, the slave manipulator is expected to exert no mechanical stress on the ocular tissue (Video 1, available at http://aaojournal.org). Several kinds of 20- or 25-gauge micromanipulators can be attached at the tip of the slave system and introduced into the vitreous body through the scleral pars plana ports. In the intraocular space, the manipulator has 5 degrees of freedom that consist of x, y, and z directions; rotation around the instrument axis; and gripping and releasing or cutting movements. The attached micromanipulators are microscissors, microforceps (Alcon, Inc., Fort Worth, TX), microneedle, and microcannula. These micromanipulators are disposable and dealt with as full-sterilization parts. The tip of the slave system, where the micromanipulators are attached, is dealt with as the semisterilization part. Visual System The 3-dimensional visual system used in the present study (Fig 3, available at http://aaojournal.org; NHK Engineering Services, Inc., Tokyo, Japan) is described in the reference.12 Briefly, the high-definition video camera with a 2010 × 1086-pixel resolution captures left-lens and right-lens images through a beam splitter. The image then is projected on a 6-inch, high-definition liquid crystal display. Pixels on this display measure 138 μm, and 500 000 pixels are projected on the screen. The resolution is 5 times that of conventional monitors. A surgeon has a clear 3-dimensional view of the vitreoretinal operative field through a prism lens viewer. Pointing Accuracy Tests on Graph Paper To evaluate the capacity of the system, we conducted pointing accuracy tests using graph paper ruled in 1-mm squares. The task was to try to keep the instrument tip as stable as possible just above the crossing point of the graph paper for 2 minutes. In surgical procedures such as retinal vessel microcannulation, it is important to hold the instrument stable to prevent iatrogenic vessel laceration or retinal break during drug infusion through the microcannula. At first, the task was conducted manually; that is, the instrument was held by hand. Second, the task was conducted through the robotic system; that is, the instrument was held by the slave manipulator. Third, the same task was conducted again manually. The tasks were recorded by the same visual system. Each frame of the film was evaluated for the distance between the aiming point and the actual position of the tip of the instrument. The average of the 3 largest distances during the 2-minute task was considered the extent of accuracy in the current study. Two graduate students from the department of mechanical engineering and 2 ophthalmologists with respective clinical experience of 6 (TU) and 20 (YT) years performed the test. Animal Model This test was conducted in harvested porcine eyes attached to the orbital fossa of a facial model. The tests consisted of 2 parts: pointing accuracy tests and more clinically associated tasks, including creating posterior vitreous detachment (PVD), retinal vessel sheathotomy, and microcannulation in the retinal vessels. Procedures of pointing accuracy tests in the animal model were the same as those using graph paper, except that the instrument held by hand or by the slave system was introduced into the vitreous body through the scleral port and the aiming point was just above the crossing or branching point of the retinal vessels. The porcine eyes were slaughterhouse materials, and this study began after the Institutional Animal Care and Use Committee of The University of Tokyo approved the experimental protocol. Results Pointing Accuracy Tests on Graph Paper The average (standard deviation) of the 3 largest distances between the aiming point and the tip of the instrument during the task was 327.0 (121.1) μm by hand (Video 2, available at http://aaojournal.org) and 32.3 (4.5) μm by robotic control (Video 3, available at http://aaojournal.org). Depending on the examinees, the range was 185 to 575 μm when manually performed, whereas it was 30 to 39 μm through the robotic system. By introducing robotic assistance, not only the accuracy was improved, but also the variance among the examinees was diminished. Pointing Accuracy Tests in Animal Model The average (standard deviation) of the 3 largest distances between the aiming point and the tip of the instrument during the task was 140.8 (28.7) μm by hand (Video 4, available at http://aaojournal.org) and 33.5 (10.6) μm by the robotic control (Video 5, available at http://aaojournal.org). The range was 115 to 170 μm through manual control and 26 to 40 μm through the robotic control. Manually performed tasks were more accurate in animal models than on graph paper because there was a stabilizing force at the pars plana ports for the micromanipulator. Meanwhile, robotically performed tasks tended to be less accurate in animal models than on graph paper. This was considered because the aiming point was not small enough to measure accuracy of robotically performed tasks. The crossing or branching points of retinal vessels have several dozens of micrometers in diameter. Still, the robotically performed tasks were approximately 5 times more accurate than those performed manually. Feasibility Tests in Animal Model Enucleated porcine eyes were attached on the facial model, and core vitrectomy was performed before the robotic system was introduced onto the operative table. Posterior vitreous detachment was created with the aid of triamcinolone acetate for visualization. The procedure used 25-gauge microforceps (Alcon, Inc.). In the enucleated porcine eyes, the posterior vitreous membrane was attached firmly to the retina, whereas the retina lost its physiologic strength to attach to the retinal pigment epithelium. Despite these difficulties, PVD was created in all 4 of the eyes by the robotic system (Fig 4; Video 6, available at http://aaojournal.org). The second feasibility test was retinal vessel sheathotomy. There was no difficulty in inserting the 25-gauge microscissors (Alcon, Inc.) into the space between the retinal artery and vein. The microscissors also could be inserted into the space between the retinal vessel and retina. The procedure was feasible in 4 consecutive tests (Fig 5; Video 7, available at http://aaojournal.org). The third procedure was microcannulation to the retinal vessels. The microcannula used in this experiment was made of glass using commercially available products (Narishige, Inc., Tokyo, Japan). The tip of the microcannula was made with an outer diameter of 20 μm and was angled at 30° for easier cannulation. Targeted retinal vessels had an inner diameter of approximately 100 μm. Microcannulation was feasible in 2 of 4 vessels (Fig 6; Video 8, available at http://aaojournal.org). The main cause of failure was the difficulty of visualization. Because both the retina and collapsed retinal vessels of the enucleated porcine eyes has white color, it was often difficult to differentiate visually between the 2 tissues during the experiment. Discussion Current robot-assisted surgeries on human patients largely are performed via the da Vinci system (Intuitive Surgical Corporation, Sunnyvale, CA). Using the da Vinci system, repair of a corneal laceration in a harvested porcine eye model was reported.13 However, because the da Vinci system has 10-mm wide arm shafts, development of a different robotic system is a prerequisite to achieve robot-assisted vitreoretinal surgery. Several designs of robot-assisted vitreoretinal surgery have been pursued to date. A handheld manipulator to prevent hand tremor during vitreoretinal surgery is a type of proposed design.14, 15 The design is unique because it focuses on improving surgeon accuracy by eliminating hand tremor. As for telerobotic systems, Charles et al16 constructed a system to remove a 380-μm diameter particle from an artificial eye model. To the best of the authors' knowledge, this report is the first to demonstrate delicate vitreoretinal surgical procedures via robot-assisted telesurgery in an animal model. This system revealed that robot assistance increased accuracy. The tasks performed by the robotic system consistently were 5 to 10 times more accurate than those performed manually based on the results of the pointing accuracy tests. Moreover, its accuracy in the pointing tasks was independent of the person controlling the robotic system, whereas there was a significant variance of accuracy in manually performed tasks, depending on the person. According to previous reports, the maximum hand tremor of vitreoretinal surgeons during intraocular procedures exceeds 100 μm.17, 18, 19 These reports are in line with the results of the authors' manually performed pointing accuracy studies. By using robotic assistance, the hand tremor was minimized and accuracy was improved to a point unattainable by conventional manual control. For reliable procedures such as microcannulation into retinal vessels with a diameter of approximately 100 μm, accuracy of less than 50 μm would be necessary. By introducing robotic assistance to vitreoretinal surgery, a novel treatment may be developed in the future. To date, the popularity of robot-assisted surgery has been limited primarily because of cost–benefit imbalance. Several studies have revealed that robotic systems can perform as speedily and safely as experienced hands.3, 6, 7 However, evidence has been scarce that certain procedures are impossible without robot assistance or that robot assistance determined the patient's prognosis. Meanwhile, robot-assisted surgery is more expensive than conventional endoscopic surgery because of its expensive instruments and the maintenance cost involved.6 Still, there is potential in robot-assisted surgery, and research should be continued. Looking at surgery in general, the case volume for surgeons is decreasing as nonsurgical treatments have been developed. This may lead to fewer opportunities to obtain the surgical education and experience essential for surgeons.8 Also, society demands safer and higher-quality care. Robot-assisted surgery has the potential to address these needs. In conclusion, a prototype robotic system was developed to assist sophisticated vitreoretinal procedures, and its accuracy and feasibility were demonstrated. Although there are some obstacles to overcome before application to clinical situations, the evidence strongly suggests that research should be continued. Supplementary data Video 1. VidClip of the slave manipulator system has a structurally fixed point that is designed to be consistent with a scleral port at the pars plana and around which the inserted micromanipulator moves pivotally. Video 2. VidClip of pointing accuracy test to try to keep the instrument tip as stable as possible just above the crossing point of the graph paper. This manually controlled instrument could not be held with the same stability as a robotically controlled instrument (Video 3). Video 3. VidClip of pointing accuracy test to try to keep the instrument tip as stable as possible just above the crossing point of the graph paper. This robotically controlled instrument could be held with the same stability compared with manual control (Video 2). Video 4. VidClip of pointing accuracy test in animal models. This manually controlled instrument could not be held with the same stability as a robotically controlled instrument (Video 5). Video 5. VidClip of pointing accuracy test in animal models. This robotically controlled instrument could be held with the same stability compared with a manually controlled instrument (Video 4). Video 6. VidClip of posterior vitreous detachment (PVD) was created with the aid of triamcinolone acetate by the robotic system in harvested porcine eyes. The procedure used 25-gauge microforceps (Alcon, Inc., Fort Worth, TX). In harvested porcine eyes, the posterior vitreous membrane was attached firmly to the retina, whereas the retina lost its physiologic strength to attach to the retinal pigment epithelium. Despite these difficulties, PVD was created in all 4 attempts by the robotic system. Video 7. VidClip of retinal vessel sheathotomy was performed by the robotic system in harvested porcine eyes. The 25-gauge microscissors (Alcon, Inc., Fort Worth, TX) was inserted into the space between the retinal vessel and retina or between the retinal artery and vein. The procedure was feasible in 4 consecutive attempts. Video 8. VidClip of microcannulation into the retinal vessel of approximately 80 μm inner diameter was performed by the robotic system in harvested porcine eyes. The outer diameter of the tip of the microcannula was 20 μm. After the microcannulation, indocyanine green dye was infused. References 1. 1Esposito M, Ilbeigi P, Ahmed M, Lanteri V. Use of fourth arm in Da Vinci robot-assisted extraperitoneal laparoscopic prostatectomy: novel technique. Urology. 2005;66:649–652. Abstract | Full Text |
Full-Text PDF (202 KB)
|
CrossRef
2. 2Falk V, Walther T, Autschbach R, et al. Robot-assisted minimally invasive solo mitral valve operation. J Thorac Cardiovasc Surg. 1998;115:470–471. Full Text |
Full-Text PDF (195 KB)
|
CrossRef
3. 3Fanning J, Fenton B, Purohit M. Robotic radical hysterectomy [online report]. Am J Obstet Gynecol. 2008;198:649.e1–649.e4. Abstract | Full Text |
Full-Text PDF (107 KB)
|
Summary PDF (91 KB)
|
CrossRef
4. 4Grossi EA, Lapietra A, Applebaum RM, et al. Case report of robotic instrument-enhanced mitral valve surgery. J Thorac Cardiovasc Surg. 2000;120:1169–1171. Full Text |
Full-Text PDF (90 KB)
|
CrossRef
5. 5Jacob BP, Gagner M. Robotics and general surgery. Surg Clin North Am. 2003;83:1405–1419. Full Text |
Full-Text PDF (302 KB)
|
CrossRef
6. 6Breitenstein S, Nocito A, Puhan M, et al. Robotic-assisted versus laparoscopic cholecystectomy: outcome and cost analyses of a case-matched control study. Ann Surg. 2008;247:987–993.
CrossRef
7. 7Zorn KC, Gofrit ON, Orvieto MA, et al. Da Vinci robot error and failure rates: single institution experience on a single three-arm robot unit of more than 700 consecutive robot-assisted laparoscopic radical prostatectomies. J Endourol. 2007;21:1341–1344.
CrossRef
8. 8Morita A, Sora S, Mitsuishi M, et al. Microsurgical robotic system for the deep surgical field: development of a prototype and feasibility studies in animal and cadaveric models. J Neurosurg. 2005;103:320–327. MEDLINE |
CrossRef
9. 9Marescaux J, Leroy J, Gagner M, et al. Transatlantic robot-assisted telesurgery. Nature. 2001;413:379–380. MEDLINE |
CrossRef
10. 10Marescaux J, Rubino F. Robot-assisted remote surgery: technological advances, potential complications, and solutions. Surg Technol Int. 2004;12:23–26. MEDLINE 11. 11Marescaux J, Smith MK, Fölscher D, et al. Telerobotic laparoscopic cholecystectomy: initial clinical experience with 25 patients. Ann Surg. 2001;234:1–7. MEDLINE |
CrossRef
12. 12Okudera H, Kobayashi S, Kyoshima K, et al. Three-dimensional Hi-Vision system for microneurosurgical documentation based on wide-vision telepresence system using one camera and one monitor. Neurol Med Chir (Tokyo). 1993;33:719–721. MEDLINE |
CrossRef
13. 13Tsirbas A, Mango C, Dutson E. Robotic ocular surgery. Br J Ophthalmol. 2007;91:18–21. MEDLINE |
CrossRef
14. 14Ang WT, Khosla PK, Riviere CN. Design of all-accelerometer inertial measurement unit for tremor sensing in hand-held microsurgical instrument. Proc IEEE Int Conf Robot Autom. 2003;2:1781–1784. 15. 15Mitchell K, Koo J, Iordachita I, et al. Development and application of a new steady-hand manipulator for retinal surgery. Proc IEEE Int Conf Robot Autom. 2007;April:623–629. 16. 16Charles S, Das H, Ohm T, et al. Dexterity-enhanced telerobotic microsurgery. Proc IEEE Int Conf Adv Robot. 1997;July:5–10. 17. 17Riviere CN, Rader RS, Khosla PK. Characteristics of hand motion of eye surgeons. Conf Proc IEEE Eng Med Biol Soc. 1997;4:1690–1693. 18. 18Riviere CN, Jensen PS. A study of instrument motion in retinal microsurgery. Conf Proc IEEE Eng Med Biol Soc. 2000;1:59–60. 19. 19Singh SP, Riviere CN. Physiological tremor amplitude during retinal microsurgery. Proc IEEE Northeast Bioeng Conf. 2002;April:171–172. 1 Department of Ophthalmology, School of Medicine, The University of Tokyo, Tokyo, Japan 2 Department of Engineering Synthesis, School of Engineering, The University of Tokyo, Tokyo, Japan 3 Department of Neurosurgery, School of Medicine, The University of Tokyo, Tokyo, Japan 4 NHK Engineering Service, Inc., Tokyo, Japan  Correspondence: Yasuhiro Tamaki, Department of Ophthalmology, School of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan
Manuscript no. 2008-1267. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Supported by 3.2 million Japanese Yen (2007–2008) from Grant-in-Aid (N0.19659443) of the Ministry of Education, Culture, Sports, Science and Technology of Japan, Chiyoda-ku, Tokyo, Japan. PII: S0161-6420(09)00231-0 doi:10.1016/j.ophtha.2009.03.001 © 2009 American Academy of Ophthalmology. Published by Elsevier Inc. All rights reserved. | |
|
FULL TEXT00231-0/journalimage?img=PIIS0161642009002310.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0161-6420&ishighres=false&allhighres=false&free=yes','journalimage');)
00231-0/journalimage?img=PIIS0161642009002310.gr2.lrg.gif&fig=fig2&kwhquery=&issn=0161-6420&ishighres=false&allhighres=false&free=yes','journalimage');)
00231-0/journalimage?img=PIIS0161642009002310.gr3.lrg.gif&fig=fig3&kwhquery=&issn=0161-6420&ishighres=false&allhighres=false&free=yes','journalimage');)
00231-0/journalimage?img=PIIS0161642009002310.gr4.lrg.gif&fig=fig4&kwhquery=&issn=0161-6420&ishighres=false&allhighres=false&free=yes','journalimage');)
00231-0/journalimage?img=PIIS0161642009002310.gr5.lrg.gif&fig=fig5&kwhquery=&issn=0161-6420&ishighres=false&allhighres=false&free=yes','journalimage');)
00231-0/journalimage?img=PIIS0161642009002310.gr6.lrg.gif&fig=fig6&kwhquery=&issn=0161-6420&ishighres=false&allhighres=false&free=yes','journalimage');)
