Kinematic Alignment With Image-Based Robotic Instrumentation

Motivation to use computer and robotic assistance for kinematically aligned total knee arthroplasty

There is growing evidence that computer assistance in general results in higher precision in TKA component position and reduces outliers, compared with manual instrumentation. ^, However, it has not been shown that this improved precision results in better patient outcomes in the context of MA. KA has the potential to improve patient outcomes compared with MA. Several studies have shown that complications and outcome are related to the precision of component position. ^, Higher failure rates have been reported when the sagittal profile (femoral flexion or tibia slope) is changed ^, and inferior outcomes have been seen when desired KA alignment was not achieved. Additionally, changing joint line obliquity from the native results in increased joint loading and possible higher wear.

This apparent need for accurate three-dimensional implant positioning of both components in KA TKA raises interest in computer-assisted technologies. However, current available technologies have their limitations. Some important parameters, such the reconstruction of the anterior aspect of the knee or the sagittal profile, are not solved satisfactorily in manual or navigation-based techniques. Purely image-based technologies, such as patient-specific instruments (PSI), are limited in that the soft tissues are not included in the planning and wear is often difficult to predict precisely before surgery. This makes intraoperative adaptations still necessary. These limitations make further evolution in technology for KA TKA reasonable.

Introduction to robotic-assisted technologies

Currently, five robotic systems for TKA are available, which can be categorized according to four basic principles. First, there are systems that actively position the cutting jig, which is assembled on a robotic arm to the desired cutting plane in the knee (Rosa®, Zimmer Biomet, Warsaw, IN, USA; OMNIBotics®, Corin Group, Cirencester, United Kingdom [UK]). The cut itself is conducted traditionally by the surgeon through the jig. Second, one system is based on a navigated burr that is used to mill the implant bed or to accurately position traditional cutting jigs (NAVIO, Smith & Nephew, Watford, UK). The burr has a retraction mechanism that prevents removal of any bone outside the desired resection plane or drill hole position. A third system employs a saw directly assembled onto a robotic arm with haptic control (MAKO, Stryker, Warsaw, IN, USA). ^, The robotic arm controls the desired resection plane and sets boundaries in which the saw can be activated by the surgeon. Fourth is an automatic robot conducting the bone resections on its own, based on a surgeon’s planning (TSolution One® Surgical System, Think Surgical, Fremont, CA, USA). So, considerable differences can be recognized already between these robotic systems in the planning programs and methods of executing the bone cuts.

In terms of the underlying navigation and planning, most systems rely on bone models selected from an atlas of knees. During surgery, several landmarks are captured on the surface of the femur and tibia by the surgeon, to which a best-fit bone model is selected to represent the patient’s knee. One system has the option to use preop imaging (X-ray based) to preselect a bone model from the database, best matching the dimension of the patient’s knee (Rosa®, Zimmer Biomet, Warsaw, IN, USA). This also enables preplanning of the prosthesis position following KA principles before surgery. Another system has the option to create a surface map of the bone during surgery, in addition to the atlas model (NAVIO, Smith & Nephews, Watford, UK). From that surface map, fake computed tomography (CT) is created on the planning screen that can be used to kinematically align the prosthesis with the femur. The two other systems are image-based (MAKO, Stryker, Warsaw, IN, USA; TSolution One® Surgical System, Think Surgical, Fremont, CA, USA). They rely on a CT scan of the hip, knee, and ankle, conducted before surgery. The bone morphology is segmented to an individual knee model. Based on this, a KA preplan can be created before surgery and adapted during surgery with respect to soft tissue, as described later.

All the systems presented are equipped with virtual planning software. This means that the prosthesis planning is initially based on the principles of measured resection. Intraoperatively, the flexion and extension spaces are recorded, and the position of the components can be adapted to achieve a balanced joint without releases. These adaptations are virtual and made before the final bone cuts are conducted. The effect on the gaps is predicted by the software. Differences between the systems are in terms of the knee positions that are included in the planning; some analyze the whole range of motion, others just the extension and flexion space at 90 degrees.

Benefits and limitations of current robotic systems for kinematic alignment

Generally speaking, all said technologies can be used to achieve KA TKA. In particular, the hybrid approach of measured resection of the femur and gap-balanced planning of the tibia cut in extension to establish a tight symmetrical extension space fits perfectly with the KA concept. With virtual planning and the precision of the technology, errors and recuts can potentially be minimized. Also, the final alignment can be controlled, helping the decision-making process during surgery. Besides this, each system has some advantages and disadvantages.

The advantages of imageless robotic-assisted technologies over manual instrumentation are pretty much the same as for conventional navigation (e.g., higher precision, visualization of alignment, objective balancing, virtual planning of component position before cuts). So is the intraoperative workflow analog to navigated KA, as described in Chapter 7. However, with the use of any standard knee model there is still a lack of information on the anterior aspect of the knee, the aspect that usually shows the greatest variability in knees. Although KA in general leads to better restoration of patella kinematics compared with MA TKA, the individual relationship between trochlea orientation and posterior condylar line remains unknown and thus the individual effect of the replacement surgery on patella tracking. Also, the individual anterior offset representing the lever arm of the quadriceps is unknown. With the addition of individual surface mapping, as described for the second technology, this limitation is addressed. With a better visualization of the patients’ individual knee geometry, more parameters are available for correct component sizing, flexion, and mediolateral (M-L) position. However, the model creation and planning are done during surgery, which takes time. Also, the more posterior aspect of the knee is difficult to reach with the probe. Thus, surface-based technology has its limitations, especially in visualizing the tibial slope.

Compared with surface-based technology, there are several advantages to image-based technologies, especially in the context of KA TKA. Analog to the mentioned surface model, the individual anterior knee aspect is visible and accessible for appropriate planning. In addition, the tibia slope can be visualized in CT-based planning and precisely adjusted to the native situation (see Basic Bernese kinematic alignment planning workflow). Second, new algorithms can be created to set the tibia rotation with reference to the transverse kinematic femoral axes, as described below. This is difficult to control in imageless navigation, where only the surgeon-defined landmarks are available. Third, the osteophytes affecting soft tissue balance are clearly visible. With image-based navigation, osteophytes can be identified during surgery and removed with navigation control (see section Bernese surgical workflow robotic-assisted kinematically aligned total knee replacement). This helps to ensure a reliable soft tissue analysis with the system, especially in difficult cases. Last, the image-based model enables the measurement of individual cartilage thickness during surgery, potentially resulting in a better reconstruction of the individual joint line. In model or surface-based methods, the cartilage is routinely assumed to be 2 mm on average. ^,

From the surgical point of view, the robotic devices themselves are of less interest in KA than the underlying navigation component. One difference between robotic systems could theoretically be the precision of each system. However, no comparative data are available. Another aspect would be possible active soft tissue protection by robotic systems. This would be applicable for the retracting burr or the saw with haptic boundaries, but not for jig-based systems used with traditional saws. However, there is little information yet on positive effects on patients’ outcomes.

In summary, the authors see several advantages to image-based planning and surgery over imageless robotics or manual techniques, making these technologies an interesting option for achieving KA. Potential pitfalls owing to individual knee morphotypes, massive osteophytes, or bone defects are more easily visible, helping to make the surgery reproducible and safe. Key is to correctly coalign the implant to the prearthritic surface of the knee. The more information available, the better this works. Additional to that, the balance of the knee is objectified. The robotic device assists the transfer of the plan to the knee, with the potential benefit for systems with resection boundaries to protect the soft tissue envelope. In the following, our KA planning and surgical workflow for one specific image-based robotic arm–assisted surgical system is described in detail.

Basic kinematic alignment planning workflow