➢ Computer-assisted surgery for total knee arthroplasty can be performed with use of computer-assisted navigation, handheld navigation, partially or fully robot-assisted technology, and patient-specific instrumentation.
➢ Computer-assisted navigation leads to improved component alignment and a reduction in the likelihood of mechanical axis outliers after total knee arthroplasty, but it is not known whether these differences have any long-term benefit on clinical or functional outcomes.
➢ Handheld navigation is a form of computer-assisted navigation that includes accelerometer and sensor-based technology. While largely unproven in the clinical literature, it offers many potential advantages over traditional navigation.
➢ Robot-assisted surgery has not been extensively studied in the context of unicompartmental and total knee arthroplasty, and, although initial reports have been promising in terms of accuracy and precision, this method is associated with substantial cost and a steep learning curve.
➢ Patient-specific instrumentation was designed to overcome many of the intraoperative challenges associated with navigation or robotic surgery, but early reports have demonstrated only minor improvements in surgical accuracy, and no change in outcomes, compared with conventional total knee arthroplasty.
Total knee arthroplasty (TKA) is one of the most clinically successful and cost-effective procedures in medicine, with excellent long-term outcomes1-3. The number of TKA procedures performed every year is steadily increasing because of the aging population and the desire for enhanced mobility in later years4,5. The increased long-term survival following TKA is therefore becoming more important to both the patient and the health-care system as a whole.
Proper component alignment, rotational position, and soft-tissue balancing are all critical factors in determining implant survival after TKA6,7. Establishing a neutral mechanical axis minimizes eccentric stresses on load-bearing surfaces and also may decrease the shear stresses on the bone-prosthesis interface8-10. Mechanical axis malalignment (defined as a hip-knee-ankle angle outside the acceptable range of 180° ± 3°) is estimated to occur after as many as 30% of TKA procedures performed with conventional instrumentation and has been shown to increase the likelihood of implant failure and bone collapse11-15.
Computer-assisted surgery (CAS) was developed in an attempt to increase the precision of component implantation and alignment during TKA. There are several terms that are used interchangeably in association with CAS, including computer-assisted surgery, computer-guided surgery, computer navigation, image-guided surgery, robotic surgery, and stereotactic surgery16. These terms may cause confusion as they have different meanings in various fields of medicine and differ among professionals, regulators, and industry personnel. It has been suggested that all of these technologies can be broken down into 4 general categories for CAS: (1) computer-assisted navigation, which uses infrared or electromagnetic technology for both preoperative planning and real-time intraoperative feedback during a surgical procedure; (2) fully-automated active robotic systems, which can perform autonomous movements under the indirect supervision of the user; (3) semi-active robot-assisted technologies, which require direct surgeon control to perform a portion of the procedure; and (4) surgery simulation systems, which are used for teaching purposes and are outside the scope of the present review16.
The present review will focus on computer-assisted navigation systems, fully-automated active robotic systems, and semi-active robot-assisted surgery as they apply to TKA. While unicondylar knee arthroplasty (UKA) will be mentioned, the present review will focus primarily on TKA. Handheld navigation and patient-specific instruments are both different forms of computer-assisted navigation and will be considered separately. With the development of CAS, it is perhaps more important than ever for the surgeon to have a firm understanding of the basic principles of arthroplasty—specifically, the ability to perform preoperative templating on radiographs with mechanical alignment devices and to use osseous landmarks to assess the size, alignment, and position of the components. The goal of CAS is not to replace these techniques, but to enhance their precision, accuracy, and reliability.
The most widely studied CAS technique for TKA has been computer-assisted navigation. Computer-assisted navigation initially was developed in the late 1990s along with advances in 3-dimensional (3D) sensor technology17. Since that time, there have been substantial advancements in both the hardware and the software associated with this technology18. The purpose of computer-assisted navigation is to provide precise implantation by means of digital mapping based on standard anatomical landmarks and kinematic analysis. Digital mapping is performed by first constructing a 3D model of the involved knee intraoperatively. The computer can then correlate this model with the surgical instruments to allow for precise cutting and implantation. It is important to note that femoral and tibial component rotation is based on anatomical landmarks at the knee such as the epicondylar axis and the Whiteside line and therefore does not account for extra-articular factors such as femoral neck version or ankle alignment. The systems that are currently available for navigation during knee arthroplasty can be generally categorized as either (1) image-guided systems or (2) image-free systems, as discussed below.
Image-guided systems use either intraoperative (fluoroscopy-assisted) images or preoperative images (computed tomographic [CT] or magnetic resonance imaging [MRI] scans) to construct a 3D model of the knee. The intraoperative position of the knee is recorded with use of markers that are drilled into the femur and tibia. The markers (also known as trackers) are also placed on the surgical instruments. The navigation computer receives information from these markers through special cameras and can then guide the placement of cutting jigs.
Image-free systems do not require a preoperative CT or MRI scan. When using these systems, the surgeon obtains anatomical registration points intraoperatively, which are then used to create a 3D knee model based on a database of knee CT scans. The knee position and cutting jig are then positioned according to the image-guided system. The major advantage of image-free systems is that they do not require the additional radiation, cost, and planning associated with preoperative CT scans.
Numerous studies and meta-analyses have shown that computer-assisted navigation improves component alignment and restores the mechanical axis during TKA15,19-22. Thienpont et al. aggregated the results of 10 meta-analyses and were able to compare the results for 28,763 patients who had undergone either conventional or computer-assisted navigational arthroplasty19. The authors found that there was a significantly lower likelihood of mechanical axis outliers (defined as >2°) when computer-assisted navigation was compared with conventional arthroplasty (p < 0.05). Of note, the authors did not mention how conventional TKA was performed, whether an intramedullary or extramedullary guide was used, or how ligament balancing was achieved. Given the evidence that computer-assisted navigation can improve the accuracy of establishing a neutral mechanical axis, it has been suggested that this technology is most useful for patients with extra-articular deformities (such as previous fracture malunions) that preclude the use of intramedullary guides and for obese patients (in whom establishing alignment can be challenging)8.
While the majority of studies have indicated that computer-assisted navigation improves alignment, controversy remains as to whether this improved alignment leads to any additional clinical benefit. Some studies have demonstrated no differences between computer-assisted navigation and conventional arthroplasty in terms of functional or clinical outcomes at up to 10 years after TKA23-27. Spencer et al., in a randomized controlled trial of 71 patients with a minimum of 2 years of follow-up, found no significant difference in functional outcome scores between patients managed with computer-assisted navigation and those managed with conventional arthroplasty23. However, in a meta-analysis of 21 studies involving 1,713 knees that were treated with either computer-assisted (n = 869) or conventional (n = 844) arthroplasty, Rebal et al. found that, in addition to improving alignment, computer-assisted navigation was associated with significantly greater increases in Knee Society Scores at both 3 and 12 months postoperatively (p < 0.03)20.
On the basis of the majority of available studies in the literature28-30, the institutional costs of computer-assisted navigation for TKA are not justified with regard to improved clinical outcomes. However, the cost-effectiveness of computer-assisted navigation may depend on the hospital volume of TKA procedures. Using a Markov decision model, Slover et al. found that hospitals that perform a high volume of TKA procedures (>150 TKAs/yr) likely would experience a cost benefit in association with the use of computer-assisted navigation as only modest reductions in the revision rate would make the technology cost-effective31. As this technology continues to evolve, its associated cost will likely decrease and it may eventually become a cost-saving technology that also improves patient outcomes.
Although TKA performed with navigation is associated with improved radiographic alignment, the use of computer-assisted navigation is associated with a number of potential drawbacks. Many navigation systems rely on the intraoperative registration of various anatomical landmarks, which, if marked incorrectly, may lead to component malposition. The most common registration errors occur at the distal femoral epicondyles; such errors may lead to malrotation of the femoral component32. Component malposition also may occur when markers that are drilled into the femur or tibia move during the course of surgery, particularly in patients with osteoporosis. Last, fractures have been reported at the drilling sites33,34. Given these drawbacks, it is important that the surgeon not rely on this technology alone to achieve proper component position and alignment. Fortunately, studies have shown that, after a relatively short learning curve, inexperienced surgeons can attain the same level of accuracy in component positioning and alignment as experienced surgeons35,36.
Summary and Future Directions
Computer-assisted navigation has the potential to reduce the complexity of arthroplasty while improving surgeon accuracy and precision. The goals of these technologies are to improve the overall alignment to within 2° to 3° of the mechanical axis and to improve both femoral and tibial component rotation. While several studies have demonstrated that mechanical axis alignment is routinely improved, the benefit of CAS in terms of rotational positioning is still in question37,38. In addition, computer-assisted navigation does not directly address soft-tissue balancing. The use of contact load sensors in conjunction with computer navigation may assist in achieving soft-tissue balance39. The widespread adoption of computer-assisted navigation will depend on its further development and cost. As the procedure becomes faster and less expensive, it may become adopted on a larger scale.
The computer-assisted navigation systems described above require a separate computer display positioned within the operating room and are therefore typically called large-console navigation. With the miniaturization of electronics in the last decade, interest has grown in using smartphone-based technology in orthopaedic surgery. The 2 primary devices that employ this technology are accelerometers and sensors. Accelerometers are small devices that measure the position of an object relative to a given axis and then use this information to calculate the mechanical axes of both the femur and the tibia in order to assist the surgeon in performing the femoral and tibial cuts (Fig. 1). Pressure sensors are small transducers that measure the mechanical force, or contact load, in both the medial and lateral compartments of the knee. Dynamic sensor output through a full knee range of motion allows the surgeon to preview the load within the medial and lateral compartments at all flexion angles. Furthermore, the contact point can be tracked for specific behaviors such as rollback or pivot. The surgeon can then make intraoperative adjustments, including ligamentous releases or osseous resections, to achieve optimal soft-tissue balance (Fig. 2). These technologies do not require preoperative imaging, increased operative time, additional incisions, or a high capital expenditure for the equipment40-43. In addition, handheld devices are typically compatible with multiple implant systems.
Handheld navigation was developed relatively recently, and there have been few high-quality studies investigating these technologies. Nam et al. performed a randomized controlled trial comparing extramedullary guides with accelerometer-based navigation for tibial alignment in 100 patients undergoing TKA40. The authors found that accelerometer-based navigation led to significant improvements in terms of both coronal alignment (95.7% compared with 68.1% within 2° of perpendicular to the tibial mechanical axis, p < 0.001) and posterior slope (95.0% compared with 72.1% within 2° of a 3° posterior slope, p = 0.007). Recently, Goh et al. performed a prospective matched-cohort study to compare accelerometer-based navigation with optical computer-assisted navigation in 76 patients undergoing TKA42. The authors found that there were no significant differences in terms of implant alignment, mechanical axis, or clinical outcomes at 6 months postoperatively. The duration of surgery was noted to be significantly longer in the computer-assisted navigation group (p < 0.05). The added cost of accelerometer-based navigation was approximately $1,000 per operation. The authors concluded that although accelerometer-based navigation may offer a similar level of accuracy without the added time and drawbacks of computer-assisted navigation, additional studies are required before it can be recommended over conventional instrumentation.
Sensor-assisted surgery is designed to improve soft-tissue balancing, which traditionally has been achieved by surgeon “feel.” Walker et al. evaluated the ability of a surgeon to use sensor-based technology to successfully balance a knee in a cadaver model44. The authors found that sensor-based technology is sensitive to relatively minor adjustments of either 2° in alignment or 2 mm in additional resection and concluded that this technology could be used intraoperatively to achieve knee balance. Gustke et al. evaluated the use of intraoperative sensors in a study of 135 patients undergoing TKA45. The authors found that 18 knees (13%) were unbalanced, defined as having an intercompartmental pressure of >15 lb (6.8 kg). At 1 year postoperatively, the percentage of patients who were satisfied or very satisfied was significantly higher in the balanced group than in the unbalanced group (96.7% compared with 82.1%, p = 0.043). In another study, the same group of authors found that balanced knees had significantly increased Knee Society Scores and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) clinical outcome scores when compared with unbalanced knees at 6 months postoperatively (p < 0.05)43. It is important to note that this clinical study did not include a control group. We are not aware of any randomized controlled trials that have demonstrated improved soft-tissue balancing in association with the use of sensor-based technologies as compared with conventional instrumentation.
Potential Drawbacks and Limitations
Handheld navigation was designed to limit the number of drawbacks encountered in association with computer-assisted navigation or robotic surgery, such as additional incisions, increased operative time, and high capital expense. However, compared with those technologies, handheld navigation is more limited in its ability. Accelerometer-based technology is designed to allow the surgeon to accurately place the cutting guides for the distal femoral and tibial cuts. However, it does not assist with anterior or posterior femoral cuts or implant rotation. Sensor-based technology is used primarily for soft-tissue balancing and is not used to make the initial femoral or tibial resections.
Summary and Future Directions
While handheld navigation has few drawbacks, it is a new technology and remains largely unproven in the clinical literature. It has been reported that this technology may cost up to an additional $1,000 per case42. With increasing pressure to decrease the episodic cost of care for total joint arthroplasty, studies will have to show that this new technology provides a clinically significant benefit in comparison with conventional instrumentation for TKA before it is adopted for widespread use.
Although computer-assisted navigation may improve component alignment, errors can still occur during the process of making the bone cuts, leading to component malposition. This limitation has led to the development of robotic assistance during TKA in an effort to facilitate the preparation of bone surfaces. The term robot refers to any mechanical device that is accurately controlled by a computer with use of intelligent software. Robot-assisted surgery requires the patient to undergo a preoperative CT scan for planning. The optimal coronal, sagittal, and rotational alignment can be determined, and the surgeon can then virtually implant the components to determine the appropriate size. The robotic system then allows for 3D milling or cutting on the basis of both the preoperative planning and, in the case of robot-assisted systems, tactile feedback from the surgeon.
The majority of the literature on the use of robot-assisted surgery in knee arthroplasty has involved UKA, a procedure in which the technology has been shown to improve component positioning and alignment46-48. UKA is particularly well suited for robotic surgery as the procedure is technically challenging, is minimally invasive, and requires resection of the diseased compartment with preservation of all other compartments (Fig. 3). Fewer studies have evaluated the use of robot-assisted surgery for TKA. Most of those studies have shown that robot-assisted TKA helps to restore mechanical alignment, especially femoral component rotation, to a greater degree than conventional surgery does49-52. Park and Lee randomized patients to either conventional TKA (n = 30) or robot-assisted TKA (n = 32) and found that, after a mean duration of follow-up of 4 years, there were significant differences between the groups in terms of the coronal femoral component angle (mean, 97.7° compared with 95.6°; p < 0.01), sagittal femoral angle (mean, 0.2° compared with 4.2°; p < 0.01), and sagittal tibial angle (mean, 85.5° compared with 89.7°; p < 0.01), all in favor of robot-assisted surgery50. The authors found no differences in terms of the Knee Society Score53, tibiofemoral angle, or coronal tibial component angle. In a more recent randomized controlled trial, Liow et al. randomized 60 patients into 2 treatment groups: robot-assisted TKA (n = 31) and conventional TKA (n = 29)51. In the robot-assisted group, there were no mechanical axis outliers (defined as >3° from neutral) or notching, compared with rates of 19.4% (p = 0.049) and 10.3% (p = 0.238), respectively, in the conventional TKA group. There were no differences between the groups in terms of clinical outcome scores after a minimum duration of follow-up of 6 months.
While most studies have shown that robot-assisted surgery is associated with improvements in lower-extremity alignment and component positioning, it is still unproven whether these improvements translate into any clinical benefit for the patient. Song et al. performed a randomized controlled trial of 30 patients undergoing simultaneous bilateral TKA, with 1 knee undergoing robot-assisted TKA and the other undergoing conventional TKA52. While the authors found that the mechanical axes, sagittal femoral component alignment, and mechanical axis outliers were all improved in association with robot-assisted TKA, there were no significant differences in terms of the range of motion, Hospital for Special Surgery (HSS) scores, or WOMAC scores at 1 year postoperatively (p > 0.05)54,55. The authors concluded that they could not show any improvements in terms of radiographic alignment or postoperative outcomes in association with robot-assisted TKA52.
Robot-assisted surgery has some disadvantages, including the additional time, cost, and radiation associated with a preoperative CT scan. Problems encountered in association with the use of markers that are drilled into the tibia and femur, including movement leading to component malposition and fracture, are uncommon but possible32-34. Another cited disadvantage of robot-assisted procedures is the lack of versatility intraoperatively, which can result in the abandonment of the robotic procedure and conversion to a conventional procedure56. As mentioned above, robot-assisted TKA does not assist with soft-tissue balancing, which is critical to the success of TKA. Last, there is a high startup cost (up to $800,000) and a learning curve associated with the use of any new technology, which may be prohibitive for many surgeons and hospital systems51.
Summary and Future Directions
Robot-assisted TKA has been shown to produce consistent and accurate postoperative mechanical alignment. However, it has not been shown to improve clinical outcomes in comparison with conventional TKA in the short-term studies that have been published to date. This method has not been cost-effective since its inception, and it exposes patients to avoidable radiation risks because of the need for a preoperative CT scan. The high cost of this method and regulatory hurdles raised by both the government and insurance companies will likely delay the widespread adoption of this technology.
Patient-Specific Instruments and Implants
Computer-assisted navigation and robot-assisted surgery require intraoperative anatomical registration, which is time-consuming and may lead to pin-site fracture. Patient-specific instruments were designed to reap the benefits of computer-assisted navigation or robot-assisted surgery while simplifying the intraoperative work-flow. A preoperative CT scan or MRI is obtained with use of a manufacturer-specific protocol and is sent to the manufacturer. The manufacturer then creates a virtual plan, which is sent to the surgeon for review. The surgeon is able to modify the plan on the basis of patient-specific factors, such as ligament deficiency, fixed deformity, or flexion contracture. Upon final surgeon approval, the manufacturer fabricates custom cutting guides that are sent directly to the hospital in a sterile instrument pack, usually arriving 3 to 6 weeks later. The purpose of patient-specific instruments is to shift the planning from the intraoperative period (as is required for use of navigation or robotics) to the preoperative period (Fig. 4).
The most frequently cited benefit of using patient-specific instruments rather than navigation or robotics is operating-room efficiency57-60. By avoiding the need to perform intraoperative anatomical registration, and by reducing the number of instrument trays required, both the operative time and the turnover time can be decreased. Patient-specific instruments obviate the need for intramedullary cutting guides, thereby decreasing the risk of systemic and pulmonary emboli61,62. Improved alignment offers the theoretical advantage of decreased rates of polyethylene wear and component failure6,7,9,11.
Given the potential benefits of patient-specific instruments, patient-specific implants were created to replicate the 3D anatomy of the individual patient. In defined cases of implant mismatch, such as posterior condylar offset, an anatomical fit may lead to improved soft-tissue tensioning or knee kinematics. Mont et al. found that there was a decreased number of contaminated trays or instruments when patient-specific implants were used63. Patient-specific implants require further studies to demonstrate the benefit of this technology.
As is the case for robot-assisted surgery, there have been few high-quality studies on the radiographic and functional outcomes associated with the use of patient-specific instrumentation. Ng et al., in a retrospective review of 569 TKAs performed with patient-specific instruments and 155 TKAs performed with conventional instruments, found that patient-specific instrumentation led to improved alignment (88% compared with 78%; p < 0.001)58. The major limitation of that study was that it was not randomized and may have been biased in favor of patient-specific instrumentation, given the higher number of patients in the patient-specific instrumentation group. Sassoon et al., in a recent systematic review of 22 studies, found that the majority of studies did not demonstrate an improvement in overall limb alignment in association with the use of patient-specific instrumentation64. There was a mixed consensus between studies with regard to the effect of this technology on reducing mechanical axis outliers. While patient-specific instrumentation did not routinely decrease operating-room time, it did reduce the number of instrument trays required for the operation. The authors found very little information on the effect of patient-specific instrumentation on postoperative pain, range of motion, function, or patient satisfaction. They concluded that patient-specific instrumentation has not decisively resulted in improved overall alignment and therefore does not appear to benefit either operating-room efficiency or the cost-effectiveness of TKA64-66.
The major disadvantage of patient-specific instrumentation is the cost associated with imaging and subsequent fabrication of single-use instruments. In addition, substantial time is required of the patient, surgeon, and office staff during the preoperative planning period. There is a potential for error during the creation of the instrumentation, in which case the surgeon would have to switch to conventional arthroplasty, resulting in substantial costs in terms of both time and frustration. Last, as is the case with any new technology, there is a learning curve associated with the use of patient-specific instrumentation.
The cost-effectiveness of patient-specific instrumentation has been debated in the literature. Proponents of this technology state that operating-room time is decreased as fewer instrument trays must be opened and set up. In addition, the surgeon theoretically should be able to save several time-consuming steps through the use of the prefabricated cutting jigs. In previous studies, the operative time that was saved appeared to be small and not of financial importance59,64. Watters et al. calculated fixed and time-dependent operating-room costs and determined that the use of patient-specific instrumentation for TKA was not associated with cost savings on a per-case basis60. Slover et al. used a Markov decision model to determine the cost-effectiveness of patient-specific instrumentation67. Assuming that the preoperative imaging study cost $1,000 and the fabricated instruments cost $1,500, the authors found that patient-specific instrumentation would have to decrease the 20-year revision rate by 50% relative to conventional instruments in order for it to be cost-effective67. The authors concluded that this reduction is highly unlikely given the known long-term success of TKA.
Summary and Future Directions
Patient-specific instrumentation was developed to enhance surgeon precision and accuracy without requiring substantial intraoperative computer referencing or the use of markers that are drilled into the tibia and femur. Unfortunately, the current literature suggests that such instrumentation is associated with only modest gains, if any, in terms of improving component alignment or saving operating-room time. At the present time, these minimal gains are far outweighed by the time and cost required for this technology.
There has been tremendous advancement in computer-assisted technology for TKA over the last 30 years. Computer-assisted navigation and robot-assisted surgery have been shown to significantly improve component alignment compared with conventional TKA. It is still unknown if this improvement will lead to better patient outcomes or a decreased rate of revision. Handheld navigation does not require additional incisions or substantial initial expense compared with those technologies, but it is still unproven compared with conventional instrumentation. Patient-specific instrumentation has been associated with less-favorable early results, with only minor improvements in component alignment when compared with conventional TKA.
At the present time, most of these technologies are too costly to justify their routine use in place of conventional TKA. Nonetheless, much of the marketing for these technologies tends to emphasize unsubstantiated benefits while disregarding potential drawbacks68. As the health-care environment becomes increasingly competitive, hospitals may embrace these technologies to attract patients. It is the role of the surgeon to explain the benefits and drawbacks of these technologies to patients so that they can make informed decisions regarding surgery. Before these technologies are embraced, future studies must demonstrate improved clinical outcomes combined with affordable costs.
Investigation performed at the Hospital for Joint Diseases, NYU Langone Medical Center, New York, NY
Disclosure: There was no external source of funding for this study. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work and “yes” to indicate that the author had other relationships or activities that could be perceived to influence, or have the potential to influence, what was written in this work.
- Copyright © 2017 by The Journal of Bone and Joint Surgery, Incorporated