The rapid development and ultimate future application of computer-assisted orthopaedic surgery can be compared with the development of intraoperative fluoroscopy and arthroscopy in past decades.19 Computer-assisted orthopaedic surgery was first clinically introduced in 1998, but there are few studies validating the safety of the different systems.12 Software modules for fracture treatment are used less than elective joint replacement modules, therefore, their safety and efficacy have not been proven. Registration of digital maps to be identified by the computer is not easily feasible in situations with different bone fragments such as in comminuted fractures of long bones.7 Computer-navigated instrumentation and implantation of screws and pins in some relatively stable fractures (eg, pelvis, spine) are relatively successful.1,21 Intraoperative fluoroscopy is the most valuable method to observe underlying bone, implant, and surgical tool positions.17 Fluoroscopy has led to the development of minimally invasive orthopaedic techniques to stabilize several fractures with cannulated screws. However, the limited field of view, two-dimensional (2-D) property of fluoroscopic images, and restricted working area hinder intraoperative control of surgical parameters.2 Surgical navigation systems based on fluoroscopy might provide missing information essential for these procedures. More intraoperative data acquisition results in more accurate fracture reduction and internal fixation, which will further improve the quality of fracture surgery.17,19
Computer-assisted orthopaedic surgery is believed to have the potential to improve the precision of trauma surgery and limit radiation and operative time.1,7,8,20 Several manufacturers produce hardware and software products for use in orthopaedic surgery. Validation of these systems in sawbones before clinical application is mandatory.
We developed a measuring method on sawbones to validate any computer navigation system designed for orthopaedic trauma. This measuring method is an adaptation of reported methods validating computer-assisted surgical techniques.11,13,18 We applied this technique to a sawbones model, which simulated the percutaneous insertion of cannulated hip screws for stabilization of impacted or undisplaced femoral neck fractures. Percutaneous insertion of cannulated screws is an accepted treatment for this common fracture, which is particularly advantageous because minimal invasive treatment of hip fractures has been reported to reduce operative trauma in elderly patients.6,8,16 Additional studies to sustain the validation of this model may be necessary, but we are confident in testing new computer-assisted orthopaedic navigation systems with our method using a sawbone model.
Does the virtual image of my screws correspond with the actual position? This is a question considered by all surgeons who work with computer navigation. The primary aim of our study was to compare computer-navigated screw fixation with conventional screw fixation using fluoroscopy in a sawbone model. The primary outcomes were the number of drilling attempts, the screw position in the femoral head, and fluoroscopy time. The secondary outcome was the ability of computer navigation to improve surgical efficiency by decreasing the operative time required for screw placement.
MATERIALS AND METHODS
We assumed the success of treating femoral neck fractures with cannulated screws highly depends on correct positioning of the screws through the lateral cortex into the femoral head.4,10 For biomechanical properties, the most favorable position of the cannulated screws is as wide and as parallel to each other as possible to achieve maximal rotational stability and to facilitate future, active dynamization.8,9,15 The tip of the screws should be as close to the cortex of the femoral head as possible without damaging it to ensure optimal anchoring.3
We used intact proximal femoral sawbones (Synthes®, Davos, Switzerland) to simulate a virtually stable intracapsular femoral neck fracture. Soft tissue coverage was omitted to simplify and facilitate orientation in the axial and anteroposterior (AP) views in both procedures. All sawbones were mounted on a bench clamp in the same manner (Fig 1). Femoral neck anteversion was rotated horizontally to facilitate and simplify the operation. In a standardized procedure, three cannulated hip screws with 7-mm thread width and 16-mm thread length and screw lengths of 95, 95, and 100 mm, respectively, (Synthes®, Zeist, The Netherlands) were inserted in each of 20 sawbones under guidance of a 2-mm Kirschner wire with a threaded tip.
Previously marked sawbones were randomly selected for screw fixation either with fluoroscopic computed navigation or with a conventional fluoroscopic technique.
The conventional and computer-navigated techniques were performed without preoperatively planning the exact screw position on radiographs or computer images. The conventional operative technique using fluoroscopy was performed following the manufacturers' instructions for insertion of cannulated hip screws, except that no parallel drill guide was used. Using a power drill, a 2-mm guide wire was gradually advanced with free-hand technique under optimal image intensification until the subchondral bone of the femoral head was reached. After measuring length, a 4.5-mm cannulated drill bit in a sleeve was used to drill the appropriate canal. Tapping was performed along the entire drill canal and the correct screw was inserted over the guide wire. We performed this procedure for the other two screws in a similar manner.
Computer-assisted insertion of cannulated screws was done using image intensification (C-arm) to acquire optimal AP and axial views of the femoral neck and head. These images were downloaded to the navigator (Vectorvision Fluoro Brainlab®, Heimstetten, Germany). A slightly different technique was used to insert the guide wire to minimize distortion of the navigator. A 4.5-mm cannulated drill bit was calibrated and drilled with free-hand technique to the determined depth guided by the two images on the navigator screen. (Fig 2). A 2-mm guide wire was placed through the drill bit after which the drill bit was withdrawn. After measuring length and tapping, the correct screw was inserted over the guide wire. This procedure was repeated for the other two screws.
The position of the screws was evaluated with spiral computed tomography (CT). The accuracy of the positioning of the three cannulated screws was determined by recording parallelism, wide triangular positioning, subchondral position, and perforation of the femoral head. We drew a line running through the core of each screw to assess parallelism. The angle between the lines was documented, and points were awarded for the angle of placement in two directions according to the following procedure: 4 points for 0° to 3° deviation, 2 points for 3° to 6° deviation, and 0 points for 6° or greater deviation. All three screws were compared with each other. A lateral CT slice of the femoral neck was used to assess wide triangular position and the center point then was determined (Fig 3). To achieve a triangular position, all screws had to be placed perpendicular to that center point. Points were awarded according to the following procedure: 1 point for the screw touching the perpendicular line, 0 points if the screw did not touch the line. Two points were given if the distance from the screw remained within 2 to 5 mm from the cortex, 1 point for 5 mm or greater, 0 points for 0 to 2 mm, and -2 points for damaging the lateral cortex of the femoral neck. The subchondral placement of the tip of the screw to the cortex of the femoral head was measured and documented. Points were based on the following procedure: 4 points for 4 to 8 mm distance, 2 points for 0 to 4 mm distance, and 0 points for greater than 8 mm distance. After implanting the screws we inspected the bone for any visible penetrations of the femoral head. No visible penetration was given 5 points and penetration was given 0 points. One investigator, blinded to navigation use, performed all of the measurements.
The total score per sawbone was calculated by adding up the total scores. This final total score varied from 0 (not accurate) to a maximum of 50 points (very accurate). We documented the total duration for operative time of each bone. The starting time of each operation was the moment of fixation of the bone to the operation table after setting up the systems. The end time was the end of the placement of the final screw, including the final radiograph. The time needed to acquire the necessary images for both procedures was included in this registration. In both procedures, maximum effort was taken to gain optimal axial and AP views with the image intensifier. No separate time intervals were registered. We also documented the total number of drilling attempts. Every correction of angle, including a backward movement of the drill, was counted as a drilling attempt. The total of radiation time visible on the C-arm was documented, with CT-evaluations not included.
All data were entered in a computer database and analyzed by an independent statistician using SPSS software (SPSS Inc, Chicago, IL). Because we compared nonGaussian distributions of sets of variables in an ordinal scale we used the Mann-Whitney U test. Statistical significance was defined for p < 0.05.
The total score to assess screw position per sawbone for both operating techniques was equivalent (Fig 4). The mean score for the computer-assisted surgical technique was 34.8 (standard deviation [SD], 6.5). The mean score for the conventional technique was 36.6 (SD, 3.35). The mean operation times for the computer-assisted and conventional techniques were 23.7 minutes (SD 6.5) and 22.4 minutes (SD 5.6), respectively (Fig 5). The mean radiation time for the computer-assisted technique (8.4 seconds; SD 4.8) was less than (p = 0.0003) the time for the conventional technique (28.2 seconds; SD 8.2) (Fig 6). The average number of drilling attempts for the computer-assisted technique (3.8; SD 0.6) was less than (p = 0.0001) for the conventional technique (13.3; SD 4.4) (Fig 7).
Safety and accuracy are the most important issues when working in virtual reality. Most advantages of computer-assisted orthopaedic surgery will be lost when every virtual move needs verification. This study with cannulated hip screws for stable femoral neck fractures provides a suitable model to evaluate the safety and accuracy of computer-navigated systems. Measurements in this model are well defined and simple and they provide a good representation of the three-dimensional aspect. This makes it suitable to compare different fluoroscopy-based navigation systems.
Manufacturers of navigation systems promise more accuracy compared with conventional methods. This claim was not supported by our study or another previously reported series.14
Our findings and another study show computer navigation systems may reduce the number of drilling attempts required to place the cannulated screws.5 Multiple attempts to insert the drill in the correct position cause perforations of the lateral cortex and damage to the cancellous bone in the femoral neck and head. Weakening of bone structures, which already are in poor condition from osteoporosis, may cause early failure of fixation. Reducing drilling attempts may be regarded as an important advantage of computer-assisted insertion of cannulated hip screws.
The radiation time obviously was less in the computer-assisted group, probably because of the limited number of images required. Operating time and radiation time decreased after the first three sawbones, suggesting a learning curve as could be expected for this sort of newly developed procedure.
This computer navigation system showed some limitations. It did not display the previously positioned screws on the screen to guide the next one. The manufacturers have included this feature in their new software modules. Additionally, there were some shortcomings in the hardware. The surgeon has to be constantly aware of the position of the infrared cameras. Blocking the field of view of these cameras will cause loss of tool reference. Furthermore, the computer does not compensate the virtual image for unintended tool deformation.
Our research model is a simplification of the real operation and therefore has some limitations. Soft tissue coverage was omitted to simplify and facilitate orientation in the axial and AP views. The acquisition of images was facilitated and radiation time was reduced, although equally in both procedures.
The method we used to assess screw position was not validated because we assumed no intraobserver and interobserver variations. To date, numerous similar measuring methods have been reported in studies validating computer-assisted surgery without any validation, assuming no obvious intraobserver and interobserver variations.18
We believe the navigation system we used is safe and accurate for insertion of cannulated hip screws. Compared with a conventional technique of screw insertion, it has the same accuracy in screw positioning and equal operating time but results in fewer drilling attempts and a substantial reduction in radiation time.
Our simple sawbones model can be used to test sophisticated navigation systems and software or hardware updates. Future studies can focus on validation and comparison with our model. This study provides an inexpensive and useful tool to test surgical navigation systems for safety reasons.
1. Arand M, Schimpf M, Kinzl L, Fleiter T, Ploss D, Gebhard F. Precision in standardized Iso-C-Arm based navigated boring of the proximal femur [in German]. Unfallchirurg
2. Arnold WD, Lyden JP, Minkhoff J. Treatment of intracapsular fractures of the femoral neck. J Bone Joint Surg Am
3. Baumgaertner MR, Curtin SL, Lindskog DM, Keggi JM. The value of the tip-apex distance in predicting failure of fixation of peritro-chanteric fractures of the hip. J Bone Joint Surg Am
4. Bosch U, Schreiber T,Skutek M. Cserhati P, Fekete K, Krettek C. Minimal invasive screw osteosynthesis of the femoral neck fracture in the elderly [in German]. Chirurg
5. Browbank I, Bouazza-Marouf K, Schnabler J. Robotic-assisted internal fixation of hip fractures: a fluoroscopy-based intraoperative registration technique. Proc Inst Mech Eng [H]
6. Hofstetter R, Slomczykowski M,Krettek C. Koppen G, Sati M, Nolte LP. Computer-assisted fluoroscopy-based reduction of femoral fractures and antetorsion correction. Comput Aided Surg
7. Janzing HMJ, Houben BJJ, Brandt SE, Chhoeurn V, Lefever S, Broos P, Reynders P, Vanderschot P. The Gotfried percutaneous plate versus the dynamic hip screw in the treatment of pertrochan-teric hip fractures: minimal invasive treatment reduces operative time and postoperative pain. J Trauma
8. Kofoed H, Alberts A. Femoral neck fractures. Acta Orthop Scand
9. Lindequist S. Cortical screw support in femoral neck fractures: a radiographic analysis of 87 fractures with a new mensuration technique. Acta Orthop Scand
10. Lindequist S, Wredmark T, Eriksson SA, Samnegard E. Screw positions in femoral neck fractures: comparison of two different screw positions in cadavers. Acta Orthop Scand
11. Mayman D, Vasarhelyi EM, Long W, Ellis RE, Rudan J, Pichora DR. Computer-assisted guidewire insertion for hip fracture fixation. J Orthop Trauma
12. Merloz P, Tonetti J, Pittet L, Coulomb M, Lavallee S, Sautot P. Pedicle screw placement using image guided techniques. Clin Orthop Relat Res
13. Ohnsorge JA, Schkommodau E, Wirtz DC, Wildberger JE, Prescher A, Siebert CH. Accuracy of fluoroscopically navigated drilling procedures at the hip [in German]. Z Orthop Ihre Grenzgeb
14. Ozanian TO, Phillips R. Image analysis for computer-assisted internal fixation of hip fractures. Med Image Anal
15. Peter RE, Fritschy D. Fracture of the femoral neck: therapeutic approach [in French]. Helv Chir Acta
16. Rouleau JP, Blasier RB, Tsai E, Goldstein SA. Cannulated hip screws: a study of fixation integrity, cut-out resistance, and high-cycle bending fatigue performance. J Orthop Trauma
17. Rzesacz EH, Weinberg AM, Reilmann H. Covered osteosynthesis through cannulated screws in medial Garden type I and II femoral neck fractures [in German]. Unfallchirurg
18. Schep NW, van Walsum T, De Graaf JS, Broeders IA, van der Werken C. Validation of fluoroscopy-based navigation in the hip region: what you see is what you get? Comput Aided Surg
19. Sikorski JM, Chauhan S. Computer-assisted orthopaedic surgery: do we need CAOS? J Bone Joint Surg Br
20. Slomczykowski MA, Hoffstetter R, Sati M, Krettek C, Nolte LP. Novel computer-assisted fluoroscopy system for intraoperative guidance: feasibility study for distal locking of femoral nails. J Orthop Trauma
© 2007 Lippincott Williams & Wilkins, Inc.
21. Stockle U, Konig B, Dahne M, Raschke M, Haas NP. Computer assisted pelvic and acetabular surgery: clinical experiences and indications. Unfallchirurg