Abstract

The author has attempted to assess the value of computer-assisted surgery in arthroplasty of the knee. Basic requisites in TKR include adequate alignment and ligament balance. These requisites have become easier to meet as ancillary instrumentations have improved over time. Numerical tools are now available; they are sometimes presented as an essential technical step. The author reviews the various available options, with their advantages and disadvantages. Satisfactory alignment in the three planes classically relies on anatomic landmarks, the reliability of which is limited, and on ligament tension. Targeting systems, intra- or extramedullary, all have a margin of error. Computer-assisted surgery aims at increasing the precision of implant positioning and achieving optimal ligament balance. Among the systems currently available, a distinction must be made between active and passive systems. The former correspond to the "surgical robots", which are capable of performing the various parts of the operation following adequate preparation, at least regarding the bone cuts. Passive systems remain under control from the surgeon and assist him in positioning the cutting jigs. Among localization systems, a distinction must be made between optical and magnetic systems. Certain systems require preoperative imaging--usually CT scan--in order to first reconstruct a 3-D model of the knee. This step is time-consuming, but this will likely improve in the future. Image matching requires the use of a software, with specific landmarks defined preoperatively by the surgeon. Such systems may be used in cases with major deformities; their main drawback is the need for preoperative imaging. Other systems do not require preoperative imaging: a few points are identified by kinematic analysis of the hip, knee and ankle; they are used for 2-D or 3-D reconstruction. Computer-assisted systems may improve the precision in defining anatomic landmarks and achieving accurate location and orientation of the bone cuts, but this must still be validated. Computer-assisted surgery allows for intraoperative control of the orientation of bone cuts, the mechanical axes, the ligament balance and the range of motion. It remains to be demonstrated that using these systems will result into a clinically relevant benefit for the patient, as several of the prostheses widely implanted so far have demonstrated high survivorship rates at 15 or 20 years, so that the demonstration of a possible superiority would need considerable investment. Using these techniques does not solve all problems anyway. The patellar cut remains empirical, and these techniques also do not give any information about femoropatellar joint function. The present systems make use of the classical instrumentation, but it may be anticipated that instrumentations especially intended for computer-assisted surgery will be developed in the future. Regarding imaging, some possibilities have not yet been explored, such as plain radiography or fluoroscopy. Computer-assisted surgery should be of particular interest in revision arthroplasties, but it poses additional problem. The indications are relatively scarce however, and no major advances should be anticipated in the near future. On the whole, this is an emerging technical field. Development is slow, none of the available products is fully mature and caution is warranted in the face of enthusiastic reports and presentations. The value of computer-assisted surgery in TKR has not been demonstrated and its real usefulness still needs to be evaluated. It will, anyway, by no means exempt the surgeon from mastering the classical operative technique.