Hip resurfacing is once more being used to treat primary and secondary osteoarthritis with considerable success and few failures according to the reference centers.4,8,10 It is considered particularly suitable in the younger patient because it is believed a more conservative operation.4,9 There is often a deformity underlying the premature failure of a young person's hip. This means the operation needed will not be standard and may not be easily performed using a standard technique. The surgeon will have to cope with morphologic distortion and compensate for this in ways that are not familiar to a THA surgeon accustomed to having a range of prostheses at hand with a range of neck lengths and offsets to allow him or her to correct any deformity with a different prosthesis rather than by altering a single device's position.
For many THA surgeons, the technique of hip resurfacing and the limited options available to change the neck length, offset, and anteversion will be novel. One unpublished study (Mont MA, Delanois RE, Plate JF, Seyler TM. Femoral neck fractures following metal-on-metal total hip resurfacing. Presented at the Annual Meeting of the American Academy of Orthopaedic Surgeons, February 14-18, 2007, San Diego, CA) described a substantial learning curve and associated high failure rate (22%), although technical errors are not cited as the principal reasons for such otherwise inexplicably high failure rates. The instrumentation provided by manufacturers varies, but all are based around the premise that a guidewire is passed along an axis composed of the center of the femoral head and the center of the neck at an appropriate angle. However, the center of the femoral head is only rarely at the end of the femoral neck axis in joints that fail early, more commonly being translated posteriorly and inferiorly to an extent so considerable variation in alignment results. Components are often put into varus, resulting in higher failure rates.2 Improvement in techniques to enhance implant placement accuracy have been suggested using either fluoroscopic guidance11 or fluoroscopic-based surgical navigation.6 In previous work, we demonstrated the advantages of precise preoperative planning as part of computer-assisted arthroplasty.3,7 We have also shown the dose of computed tomography scans acquired for planning arthroplasty can now be reduced to the level of a single long-leg radiograph.5 Because improvements in computer processing have meant a simpler and faster planning process available on any laptop, we believed the impact of detailed planning alone might be sufficient to improve accuracy to an acceptably high level and thereby obviate the need for navigation systems.
We therefore sought to ascertain the individual influence of three-dimensional planning and navigation when compared with conventional planning and instrumentation in the adoption of a novel technology such as hip resurfacing.
MATERIALS AND METHODS
We recruited 20 medical students studying for their Bachelor of Science in surgical technology for this randomized trial. They were asked to pass a guidewire through the femoral head and neck in an optimal orientation for hip resurfacing using three different orientation tools to aid them. Their accuracy would be measured against the optimal position, and the differences in accuracy of the different systems would be calculated using a t-test. The students had already learned about arthroplasty in general and had been taught the principles of hip resurfacing as well as the place of planning and navigation in THA. They were randomized into three groups. Every student performed the three tasks in rotation. Group 1 (eight students) performed the tasks in the order described using conventional instruments followed by a plan and then followed by a navigation system, a stepwise development in training. Group 2 (eight students) started with the detailed plan, then registered and navigated, and finally used conventional instruments; these students might show navigation and planning improve conventional surgery. Group 3 (five students) started by registering and navigating, then used conventional instrumentation, and finally used the detailed plan without the navigation system; these students might show, after the use of navigation and conventional instruments, the use of a detailed plan was sufficient to provide a good result.
We used the Wayfinder system (Acrobot Co Ltd, London, UK) developed in collaboration with the authors with the tasks of navigating for joint replacement specifically in mind. The planner has been developed specifically for hip resurfacing. Based on a low-dose computed tomography scan, it allows the surgeon to place the components of the hip resurfacing procedure in what he or she considers the optimal position for each case. By using a mixture of two-dimensional and three-dimensional visualization, it is possible to make very small adjustments that ensure an optimal fit of both components, avoiding oversizing, notching, or angular malpositioning of either component (Fig 1). In our unit, each case is planned preoperatively, showing the patient exactly what size devices will be used and precisely where they will be positioned.
The Wayfinder navigation system uses two digitizing arms that are coregistered and controlled by a foot switch and a touch-screen monitor. One arm is fixed to the bone with two screws. It tracks the bone continuously while the other arm tracks whichever instrument is attached to it. To make the task simpler and quicker in this experiment, each bone was immobilized in a vice; thus, only one arm was needed because no bone motion took place so no updating of position was needed.
Registration using the Wayfinder relies on the surgeon acquiring approximately 30 points in total from three regions of interest: the head and the greater and lesser trochanter (Fig 2). The algorithm developed from that used at the knee7 and is based around an iterative closest point matching system. Typically after the acquisition of this minimal data set, the root mean square error will be on the order of 0.5 mm. The surgeon then performs a “reality test,” drawing the registration probe off the side of the head-neck junction, dropping into the fovea if it is present, and running around the back edge of the greater trochanter. These tasks rapidly provide the surgeon with the reassurance he or she needs that the registration is accurate enough. Further points can be acquired or the entire process can be repeated if any question arises in the surgeon's mind that the coregistration is not adequate. The time taken to acquire the points for registration and accept the accuracy level is typically 2 minutes.
During navigated hip resurfacing arthroplasty, the surgeon attaches a tracker to the drill guide, which he uses to guide a drill-tipped guidewire antegrade through the femoral head and neck instead of the conventional surgical alignment guides. He is guided by the images on screen with both visual and numeric cues (Fig 3). Visually he sees a disc enlarging and changing color as the optimal position is reached and then passed. The angular and translation distances from the planned position are also visible in numeric form on the screen.
The guidewire determines the position of the femoral head with 4° of freedom. A head top cutting guide can also be navigated. The surgeon can also use the navigation system to validate the position achieved; after implantation of each component, which invariably involves banging, the position of the implant against the planned position is checked against the planned position on screen, in particular with regard to full seating of the implant on the head (Fig 4).
Scans were taken of the four different types of dry bone femora using a protocol we have developed from our work on the knee.5 This low-dose algorithm is developed to minimize the dose to the ovaries and other pelvic viscera by scanning with coarse cuts through the top of the iliac crest and then only taking detailed cuts through the acetabulum and femoral heads as far as the lesser trochanters, thereby avoiding irradiating the testes. The estimated dose is now 2.5 mSev for the hip, less than half the annual radiation dose recommended for medical investigation.
The computerized tomography slices were then semiautomatically segmented by one of the authors (VK) using software provided by the navigation company. Each bone therefore had a unique three-dimensional computer model and three unique plans: (1) a “conventional” position based on the center of the neck of the femur and a valgus angle that was clinically optimized using the alignment jigs provided by the implant manufacturers (Corin, Cirencester, UK); (2) a “planned” position offset from the “conventional” position by several millimeters. We explained to the students this was necessary to measure their ability to achieve a plan that might not look right but was what had been decided was biomechanically optimal; and (3) a “navigated” position also offset by several millimeters from the “conventional” and “planned” positions. Students were thus able to use the same bone three times, knowing the correct position of the guidewire for each task did not start in the identical place on the femoral head nor necessarily follow the same trajectory. Three surgical aids were then used to help the student while inserting the femoral guidewire. Each generated a unique guidewire insertion hole and trajectory. After each task, the entry hole was labeled 1, 2, or 3 to signify which method had been used. The students could see the entry holes for successive tasks and had been informed that each task required a different entry site for optimal positioning for that specific task.
For conventional instrumentation, we used alignment jigs supplied by the manufacturer (Corin). Every hip resurfacing system uses a set of jigs to help the surgeon center the post of the hip resurfacing implant in the center of the neck (Fig 5). These jigs are based around a set of circular tongs. The surgeon places these around the neck of the femur. An outrigger mounted on the tongs ensures the guidewire passes through the center of the tongs and thus down the middle of the neck of the femur. Varus and valgus angular alignment and anteversion are obtained using two alignment rods offset from the femoral neck.
For the second task, the student was provided with a unique plan made for each individual bone by the authors using the planning software provided. This plan was illustrated with a series of pictures showing the exact site of entry from the top, the side and end on, together with anteroposterior and lateral radiograph-type views so the student could visualize the position of the guidewire within the bone. Before starting to position the wire, the pictures showed exactly how the wire should look coming out of the femoral head and the angulation of the wire relative to the head and neck that was needed. The conventional instruments were available to assist if requested.
Students were asked to undertake two sequential tasks using a navigation system. They first had to coregister their individual bone to the unique preoperative plan provided. They then had to navigate the guidewire to the desired position as accurately as they believed possible.
Twenty bones were obtained (Sawbones, Portland, OR). They were of four different types: (1) normal anatomy (11 cases); (2) osteoarthritis (five); (3) slipped capital femoral epiphysis (two); and (4) coxa valga (two).
After the tasks were completed, all bones were reregistered and the Acrobot navigation system was used to measure the trajectory and offset of each hole from the expected position for that task. Because each task had a position that was unique, each pin track provided an observation that could be compared with the expected position for that individual task. The navigation system provided the offset and angular deviation from the task in millimeters and degrees.
We measured 75 hip resurfacings performed by arthroplasty surgeons in a government-run treatment center using conventional instrumentation. Their angular results had a range of 44° and a standard deviation of over 9°. The accuracy of the Wayfinder in 10 dry bone cases by the authors suggested a range of 5° and a standard deviation of 2°. A power calculation suggested the navigation system would demonstrate considerable improvement over conventional instrumentation with a sample size of eight cases in each group. We increased this to 20 cases because the subjects of this trial were medical students. Statistical Package for the Social Sciences (SPSS Inc, Chicago, IL) was used for the t-tests to compare the accuracy of the three groups.
The range of error in varus-valgus angulation found when using the conventional method to insert a guidewire was 23° (standard deviation [SD], 6°); using the computed tomography plan method, it was 22° (SD, 7°) (Fig 6); and using the navigation method, it was 7° (SD, 2°). Navigation was more accurate (p < 0.002) than either conventional or planned surgery. The mean error in angulation was less than 1° for both the planned and navigated group, whereas the mean error for conventional group was greater (p < 0.008) with 5° more valgus than planned.
The students in Group 3, who used the navigation system to register and then guide their guidewire without any exposure to conventional instrumentation, were as accurate as both other groups and were more trained before they started navigation (Fig 7). Experience in the use of conventional instruments did not influence the ability to achieve a plan or navigate. We detected no effect in the sequence of training. Students produced similar accuracy even in their first attempt on difficult anatomy when provided with navigation technology.
Abnormal anatomy did not affect the accuracy (Fig 8).
The accurate positioning of the femoral guidewire for hip resurfacing is an essential and novel task that is demanding even for well-trained surgeons. In this small study, we set out to assess the impact of two different surgical aids on the training of novices in this particular task. In addition to the conventional instrumentation, we introduced a specific three-dimensional plan and a surgical navigation system. We further anticipated there would be an optimal sequence of training for the delivery of accurate pin placement. We therefore designed a rotation of tasks to investigate this effect. Although the conventional task was straightforward, the other two tasks of guidewire placement, with the three-dimensional plan as an aid or with a navigation system as an aid, were less simple; the subjects were not asked to plan the position of the guidewire, but simply to carry out the detailed plan.
We chose medical students who were entirely naïve in the field of hip surgery but technically inclined by their choice of Bachelor of Science courses as an experimental group, because the techniques of hip resurfacing have become quite widespread in the United Kingdom, so obtaining a group of surgical residents who were entirely naïve would have been hard. The subjects' lack of experience in the use of conventional instruments is a major limitation of the trial and the differences would therefore not likely reflect those of experienced surgeons. We used different bone types, some of which were very deformed, but four students allocated to the more deformed group failed to appear for their sessions to perform their tasks, so this phase of the study was underpowered. We used the same bone for each student to perform the three tasks for reasons of economy. The students were told the site of entry into the bone and the trajectory of the wire in each task were different, but it was not possible for them to be blind to the position of the previous holes. It is difficult to compare these data directly with clinical experience; we set very specific targets and measured accuracy against those targets. However, the range and standard deviation of varus-valgus malalignment measured from the radiographs of cases performed by a group of surgeons using conventional instruments is similar to students using conventional instruments on a dry bone in a laboratory and thee times worse than the same students using surgical navigation for the first time. Poor outcome is considerably more likely with varus neck alignment1 of only 133°. Over 30% of the cases performed by surgeons in our clinical control are in this category compared with only 5% (one case) of those performed by students with navigation.
The range of inaccuracy of 22° shown by all groups when using conventional instruments matches the range of angular variation we have seen in ordinary surgeons resurfacing using conventional instrumentation, who have a similar scatter in the angular inaccuracies of femoral component position when assessed radiographically (Fig 9). Of course, comparing radiographic data against a mean of 135° of inclination for the femoral component is simple on two fronts: radiographs are notoriously inaccurate and the surgeons were not necessarily trying to achieve those angles. However, the scatter probably represents an error that is not so different from the students on their first adventure into the field of operative orthopaedics. We have shown clearly the student's accuracy using appropriate technology is substantially superior to an average surgeon on an average day using conventional instruments.
It has been our experience that the act of planning a task in three dimensions beforehand, effectively performing virtual surgery, adds considerably to the comfort a surgeon feels when performing surgery with conventional instruments, because it shows him or her the optimal entry point and trajectory for a given disease process. We were unable to demonstrated improved accuracy in femur type 3-a gross slipped epiphysis-owing to the small sample size. We had planned for double the number of cases, but as noted four students allocated these trials failed to attend. We were unable to show a reduction in scatter from the presence of a detailed plan, but did demonstrate a considerable difference in the mean value achieved; students using a plan on average achieved that plan within 1°, while using conventional means, on average inserting the guidewire into a more valgus position by 5°. The standard deviation of the two groups was similar, however, as was the range. We did not ask the students to carry out the planning task themselves; in this study, we were specifically interested in the task of guidewire insertion.
Surgical navigation, entailing the registration of a bone to a three-dimensional model on screen and the use of a navigation system to help guide a wire into a position are two tasks that were entirely novel to the subjects. Registration to a computed tomography-based model of that exact bone minimized the error than can be associated with this aspect of surgical navigation, enabling the study to focus more on the guidewire insertion. The subjects were asked to perform the task once only. In that single task, they achieved a range of only 7° and a standard deviation of only 2°. This is three times the level of accuracy of the nonnavigated groups.
We expected students would learn sequentially, and by stepwise training, a skill could be acquired safely. We did not show learning conventionally first improved matters, even in difficult cases. Indeed, we failed to show any influence of the sequence of training on angular accuracy. Instead, we demonstrated motivated and enthusiastic students can achieve an expert level of accuracy. They acquired that level of accuracy on their first attempt when provided with the appropriate level of technology.
In a world in which navigation technology is widely available in cars, on boats, and even on bicycles, its use in the operating room to improve outcomes is increasingly expected by patients. The introduction of hip resurfacing, an exciting new technology with very high expectations, should be paralleled by the appropriate use of computer assistance. These two will combine and make excellent modern surgical practice. Our data suggest surgical navigation may play a major role in reducing the length of the learning curve in hip resurfacing arthroplasty and other tasks in arthroplasty in which a high degree of accuracy is clinically important.
We thank the students from Imperial College who participated in this study.
1. Beaule PE, Dorey FJ, LeDuff M, Gruen T, Amstutz HC. Risk factors affecting outcome of metal-on-metal surface arthroplasty of the hip. Clin Orthop Relat Res
2. Beaule PE, Lee JL, Le Duff MJ, Amstutz HC, Ebramzadeh E. Orientation of the femoral component in surface arthroplasty of the hip. A biomechanical and clinical analysis. J Bone Joint Surg Am
3. Cobb J, Henckel J, Gomes P, Harris S, Jakopec M, Rodriguez F, Barrett A, Davies B. Hands-on robotic unicompartmental knee replacement: a prospective, randomised controlled study of the Acrobot system. J Bone Joint Surg Br
4. Daniel J, Pynsent PB, McMinn DJ. Metal-on-metal resurfacing of the hip in patients under the age of 55 years with osteoarthritis. J Bone Joint Surg Br
5. Henckel J, Richards R, Lozhkin K, Harris S, Baena FM, Barrett AR, Cobb JP. Very low-dose computed tomography for planning and outcome measurement in knee replacement: the Imperial Knee Protocol. J Bone Joint Surg Br
6. Hess T, Gampe T, Kottgen C, Szawlowski B. Intraoperative navigation for hip resurfacing: methods and first results. [in German]. Orthopade
7. Jakopec M, Harris SJ, Rodriguez y Baena F, Gomes P, Cobb J, Davies BL. The first clinical application of a “hands-on” robotic knee surgery system. Comput Aided Surg
8. McMinn D, Treacy R, Lin K, Pynsent P. Metal on metal surface replacement of the hip. Experience of the McMinn prothesis. Clin Orthop Relat Res
9. Treacy RB. To resurface or replace the hip in the under 65-year-old: the case of resurfacing. Ann R Coll Surg Engl
. 2006;88:349-353, discussion 349-353.
10. Vale L, Wyness L, McCormack K, McKenzie L, Brazzelli M, Stearns SC. A systematic review of the effectiveness and cost-effectiveness of metal-on-metal hip resurfacing arthroplasty for treatment of hip disease. Health Technol Assess
© 2007 Lippincott Williams & Wilkins, Inc.
11. Wirth CJ, Gosse F. Improved implantation technique for resurfacing arthroplasty of the hip. Oper Orthop Traumatol