THEORY

In conventional hip replacement the operating surgeon relies on visualized and palpated bony landmarks on the pelvis and lower limb to gain feedback when positioning components. Mechanical alignment guides used for acetabular and femoral component insertion have set angles that act as mechanical guides when components are implanted. These devices in general rely on accurate positioning of the patient and may introduce error if the patient moves on the table after initial positioning.

In computerized hip navigation the computer is provided with information from trackers fixed to the pelvis and femur, as well as real-time information on the position of instruments or implants that similarly have attached trackers. The computer screen is the interface with the surgeon, providing accurate, real-time information on position of components as they are implanted.

Hip navigation systems can be divided into two groups: image-guided navigation and imageless navigation. Image-based systems require either a preoperative CT scan or intraoperative fluoroscopic images.

In the first group, a CT scan is taken preoperatively. With planning software the optimum position of the cup can be determined by overlaying virtual templates on the axial, sagittal, and coronal pelvic images. All measurements are related to the anterior pelvic plane, a plane incorporating both anterior superior iliac spines (ASISs) and the center of the pubic symphysis, which can be defined on the three-dimensional reconstruction of the CT images. However, a significant variation between the frontal plane of the pelvis and the gravitational axis of the body has been well documented. Some current systems employ algorithms that enable the surgeon to estimate the gravitational or functional plane as well as the pelvic frontal plane. It is still then left to the surgeon to decide the appropriate position for acetabular component placement.

Another key component is the optical tracking camera, which can track optical targets equipped with light-emitting diodes, so-called “trackers.” One tracker is rigidly fixed in the pelvis. Key points on the pelvis are registered using a pointer attached to a tracker that is optically tracked. The surface geometry can be registered and matched to the preoperative CT scan. This pelvic geometry is related to the fixed pelvic tracker. The acetabular reamer and cup holder can also be attached to an instrument tracker, allowing the computer to visualize and guide reaming, broaching, and actual component insertion.

Most surgeons, however, have found it hard to justify the extra radiation exposure and planning time and the costs of scanning every patient as required by CT-based systems. Fluoroscopic-assisted hip navigation was introduced as an alternative.

When fluoroscopic-assisted hip navigation is performed, patient setup is no different from that for a standard hip replacement, although attention has to be paid to access for the C-arm of the image intensifier. A fixed tracker is attached to the pelvis and the femur. The ASISs are defined using a pointer with an integral tracker, with the geometric centers of both pubic tubercles identified by image intensification such that the anterior pelvic plane is defined. Through image intensification, femoral reference points are defined, including the center of the femoral head, the medullary axis, the knee center, and the posterior aspect of the medial and lateral femoral condyles. A femoral reference plane is defined by the medullary canal axis and a tangential line linking the posterior medial and lateral femoral condyles. The acetabular reamers, cup introducer, broaches, and stem introducer can be attached to trackers, their position can be referenced to the defined anterior pelvic plane and femoral plane, and overlaid in real time onto the stored fluoroscopic images.

Imageless hip navigation systems require no preoperative or intraoperative scanning or fluoroscopy apart from the plain radiographs that would be performed by all surgeons before any hip arthroplasty procedure. At the time of surgery, a fixed body with a tracker attachment post is attached to the iliac wing and the distal femur on the appropriate side for surgery. As before, these trackers are critical to the procedure, with all subsequent component positioning referenced to these fixed points. They can either be passive, reflective arrays, picked up by the computer’s overhead infrared sensor, or independent, active microprocessors that send signals to an overhead camera ( Fig. 60-1 ).

Computer navigation unit with integral overhead sensor and screen for surgeon feedback regarding component positioning.

The experience of this unit is with active trackers. Users of passive trackers have occasionally had problems with blood on the trackers obscuring the reflection of the infrared camera signal. The sensor or camera is connected by a small gantry to the navigation computer and screen, which is a contained, mobile unit (see Fig. 60-1 ).

The anterior pelvic reference frame is defined by a pointer with an integral tracker, as discussed ( Fig. 60-2 ).

Screenshot showing anterior pelvic plane, which is defined by both anterior superior iliac spines.

The femoral plane, or femoral reference plane, is the plane connecting the piriformis fossa with the midpoint of the popliteal fossa and the tendoachilles, which should incorporate the anatomic axis of the femur and be perpendicular to the coronal plane. This plane is defined by the pointer; the surgeon need not rely on seeing bony landmarks with image intensification. The pointer can also be used to trace the acetabular rim, define the articular surface, and define the true acetabular floor.

As with the reamers, broaches, and implant introduction handles, the pointer is of a fixed length that is automatically incorporated into calculations performed by the software. The positions of the mobile instrument trackers relative to the rigid trackers are also known, and indirectly their position relative to the acetabulum, pelvis, and femur can be calculated by the software.

EVIDENCE

Several studies evaluating CT-based navigation systems have been performed. Jolles, Genoud, and Hoffmeyer recruited 10 surgeons to each perform 15 trial insertions of an uncemented cup into a plastic bone and foam model using standard mechanical guides and the CT-dependent Medivision system (Medivision, Oberdorf, Switzerland). The results were highly significant, with a 60% reduction in maximum error and a 50% reduction in mean error for both cup inclination and version. Widmer and Grutzner evaluated 14 hip replacements performed by two surgeons using a CT-dependent navigation setup (SurgiGATE, Medivision, Oberdorf, Switzerland). In this small study, cup position was optimized (i.e., the range of values for cup inclination and version was significantly narrowed). Leenders and colleagues also showed that CT-based navigation improved accuracy of cup position in a control study of 150 hips, with a significant decrease in the variability of cup position for both inclination and version.

Fluoroscopic navigation techniques have also shown promising results. Grutzner and colleagues performed 236 consecutive hip replacements using fluoroscopic images linked to navigated pelvic registration. In the first 50 cases, acetabular position was evaluated using the CT planning software. The mean inclination was 42 degrees (37 to 49 degrees), and the mean version was 21 degrees (10 to 28 degrees). The difference between intraoperatively calculated cup position and postoperatively measured implant position was 1.5 degrees (standard deviation [SD] 1.1 degrees) for inclination and 2.4 degrees (SD 1.3 degrees) for anteversion. Zheng and colleagues also assessed the same fluoroscopic navigation technique in 10 cases. The numbers were small, but in addition to acetabular navigation, the femoral component was also navigated with respect to offset, version, varus-valgus alignment, and leg length. In eight cases, surgeons were satisfied with the femoral component position guided by the navigation. In two cases the navigation had to be abandoned because of technical problems.

However, the results now appearing with imageless computer navigation are most relevant, as there is a general shift toward this most recently developed form of hip navigation. Nogler and colleagues used the Stryker Hip Navigation System (Stryker Leibinger, Freiburg, Germany). Bilateral acetabular cup insertion was carried out in 12 human cadavers through an anterolateral transgluteal approach using a press-fit, uncemented cup. One side was navigated and the other side was a control. The median position for median inclination and median version were similar for both groups, but the 90th percentile range was 15.7 degrees and 18.5 degrees, respectively, for inclination and version in the control group, compared with 4.3 degrees and 7.1 degrees in the navigated group. Kiefer performed 156 cementless total hip replacements with cup placement guided by the OrthoPilot hip navigation system (B. Braun Aesculap, Tuttlingen, Germany). Postoperative cup position was measured on plane x-ray images, with Pradhan’s technique used to calculate cup version. The mean inclination angle given by the computer at cup insertion was 41 degrees (SD 5 degrees) and was 42 degrees (SD 5 degrees) for the inclination measured on x-ray films. Cup version was less accurate, with a mean version at cup insertion of 15.9 degrees (SD 4.5 degrees) and of 10.9 degrees (SD 4.8 degrees) on postoperative radiologic assessment. Wixson and colleagues performed a consecutive series of 82 hips navigated using the Stryker Leibinger Hip Navigation System (Stryker Leibinger, Freiberg, Germany) and compared postoperative cup position with positions from a cohort of 50 standard hip replacements, using digitized x-ray studies. The target range for cup inclination was 40 to 45 degrees and the target range for anteversion was 17 to 23 cases. Only six of the control cases were in the target range, compared with 30 of the navigated cases, a highly significant difference ( P < .001).

Most of the studies presented have focused on acetabular component positioning, and there have been weaknesses in all cases. The limitations of postoperative radiographic assessment with plane x-ray studies, a method used in some of the studies mentioned in the previous paragraphs, were alluded to earlier. Lack of a control group or adequate statistics lowers the value of others. Perhaps most important of all is what implant positions are considered optimal. Many of the papers published make reference to Lewinnek’s guide to optimal cup position. This was a retrospective study on small numbers using plane x-ray images to measure cup position.

We have been prospectively evaluating the outcomes of the Stryker Hip Navigation System (Stryker Leibinger, Freiberg, Germany) as part of a randomized control trial. Postoperative cup and stem position is measured on high-definition postoperative CT scans using the Stryker CT Hip Navigation software for accurate three-dimensional measurement of both acetabular and femoral component positions. We have found this system to have very good accuracy, with low intraobserver and interobserver error. The aim of the study has been to demonstrate the accuracy of hip navigation—that is, how well the position for an implanted component indicated by the hip navigation system relates to the postoperative position. For control hips we asked the operating surgeons to estimate the position they thought they had achieved, and related this to the postoperative three-dimensional CT measurement. We have performed 80 total hip replacements, with navigation used in 39 cases. In the navigated group, two subgroups were studied; in one group, pelvic registration was performed with the patient supine before positioning, and in the other group, registration was performed after the patient was prepared and draped in the lateral position. Both techniques were assessed for accuracy. Forty-two hips were uncemented, and 38 were cemented. There was no difference between the groups with regard to pathology, age, sex, body mass index, or type of implant. The difference between predicted and postoperative position was normally distributed for both the navigated and the control groups. The range of values was significantly narrowed in the navigation group for cup inclination ( P < .05), cup version ( P < .05), and stem version ( P < .05).

As shown in Figure 60-3 , in our hands registration in the supine position was more accurate than registration in the lateral position. Registration in the lateral position was still significantly more accurate than in the control cases. There was one infected hip in the control group and a single dislocation in both groups.

Graph presenting the accuracy of cup anteversion to the nearest 5 degrees. Accuracy is defined as difference between what the surgeon felt was achieved intraoperatively and the postoperative CT measurement in the control cases. In the navigated cases, accuracy is defined as the difference between cup inclination predicted by the computer navigation system intraoperatively and the postoperative CT measurement. All navigated cases were registered in the lateral position. In the patients in whom additional supine registration was carried out, the navigation system would allow the value of anteversion for both types of registration to be displayed.

We also evaluated the accuracy of leg length estimation in 17 navigated cases. In most cases the change in leg length measured to the nearest 5 mm by the surgeon was the same as that predicted by the navigation unit. It is difficult to say just how accurate the predicted change is, but in our experience, having a gross unexpected leg length discrepancy would not be possible with navigation.

The lack of reliance on a CT scan to which the pelvis can be matched by point registration has prompted some to question whether imageless navigation can be as accurate as CT-dependent navigation. Kalteis and colleagues compared 30 hips placed conventionally, 30 hips placed with CT-dependent navigation, and 30 hips placed with imageless navigation. Both forms of navigation significantly improved the accuracy of cup placement compared with conventional freehand arthroplasty. There was no demonstrable difference between the two techniques.