Accuracy of Navigation: A Comparative Study of Infrared Optical and Electromagnetic Navigation

Abstract

We evaluated the accuracy of navigation systems for measuring the mechanical axis in patients undergoing total knee arthroplasty and in the synthetic bone model. Infrared optical and electromagnetic navigation systems were compared. Both systems were found to be accurate and reproducible in an experimental environment. However, the accuracies of both systems were affected by erroneous registration, and the optical system was found to be more reproducible. In clinical situations, the mean difference was 1.23°, and difference greater than 3° occurred in 15% of clinical trials. These discordances may have been due to ambiguous anatomic landmarks causing registration errors and the possibility of electromagnetic signal interference in the operating room.

The accuracies of lower extremity alignment and implant position significantly influence long-term results of total knee arthroplasty (TKA). Recent advances in computer technology have improved navigation systems and reduced lower extremity alignment and implant positioning outliers compared with conventional alignment tools. ^1-6 Computer-assisted navigation includes optical navigation and electromagnetic navigation systems and the ultrasound-guided navigation system introduced recently. Experimentally, all navigation systems are known to have errors of less than 1 mm or 1°.^2,7-11

Of these different systems, the infra-red optical navigation system has been well popularized, and clinical results accumulated over 10 years confirm its expected accuracies.^2,12-15 Although the accuracy of electromagnetic navigation remains controversial with regard to metallic interference, recently developed equipment is known to be more accurate and much less affected by intraoperative metallic instrumentation.^16,17

Moreover, few comparative studies have been conducted on the accuracies of optical navigation and electromagnetic navigation systems. This study was undertaken to evaluate the accuracy of infrared optical and electromagnetic navigation systems under clinical and experimental conditions.

Materials and Methods

In Vivo Experiment

We compared the preoperative lower extremity mechanical axis of 20 cases of TKA using the OrthoPilot optical navigation system (B. Braun Aesculap, Tuttlingen, Germany) and the AxiEM electromagnetic navigation system- (Medtronic Navigation, Coal Creek, Colorado). Preoperatively, mechanical axes using weight-bearing anteroposterior full leg radiography was taken.

Figure 1: Synthetic bone models. The hip, knee, and ankle joints are made of titanium, which has no effect on electromagnetic field. The knee joint is constrained to not allow varus or valgus motion.

For OrthoPilot navigation, transmitters were fixed to the distal femur and proximal tibia. Kinematic registration was done for the hip, knee, and ankle joint centers for a range of motion study. Anatomic landmarks such as the center of distal femur, proximal tibia, and ankle and both malleoli of the ankle were registered using probes, and mechanical axes were measured. For AxiEM navigation, trackers were fixed to femur and tibia. Only the hip center was registered using the kinematic method, and the other anatomic landmarks were pointed and registered using a probe. Anatomic landmarks that can affect the mechanical axis include the center of distal femur and proximal tibia and both malleoli of the ankle. The order of application of both navigations was randomized, and single surgeon performed these clinical trials.

Comparison of Accuracies Using a Lower Extremity Synthetic Bone Model

To check the mechanical axis differences in an experimental model between the two navigation systems, we used a lower extremity bone model (Sawbones; Pacific Laboratories, Vashon, Washington), which extended from the pelvis to the foot. Mobile hip, knee, and ankle joints were made of titanium that was not affected by electromagnetic field. Varus or valgus motion was not allowed in the knee joint (Figure 1). Mechanical axes of synthetic bone were evaluated using the OrthoPilot the optical system and the AxiEM electromagnetic system. The registration process of both navigation systems was the same as a previously described process in an in vivo study. Four orthopedic surgeons participated in this experiment and applied both navigations 10 times independently. Two of the surgeons had performed more than 100 TKAs with navigation, whereas the other two had no such experience.

To obtain a true mechanical axis of the synthetic bone, the Orthodoc system (Robodoc preoperative total knee arthroplasty planning software; Curexo Technology Corporation, Sacramento, California) was used after obtaining helical computed tomography images (1.0-mm section thicknesses). We created 3D reconstruction images of saw bone and defined the center of the femur head, distal femur, proximal tibia, and ankle (Figure 2). The true mechanical axis was then obtained by computer after connecting the centers of the hip joint, distal femur, proximal tibia, and ankle. Two orthopedic surgeons checked it 5 times each.

Figure 2: Orthodoc system. A, Femoral head center; B, center of distal femur; C, center of proximal tibia; D, ankle center; E, measurement of mechanical axis.

Intentionally Erroneous Identification of Anatomic Landmarks

Using the same bone model, anatomic landmarks were intentionally erroneously identified and changes in the mechanical axis were recorded by both navigation systems (Figure 3). Centers of the distal femur, proximal tibia, ankle, and both medial and lateral malleoli centers were registered 10 mm medially and laterally compared with the original points, and the mechanical axis with original and erroneous data were compared. This process was performed by one operator.

Figure 3: Erroneous identification of anatomic landmarks. A, Distal femur; B, proximal tibia; C, ankle.

Statistical Analysis

The Mann-Whitney test was used to analyze anatomic axis differences, and Pearson’s correlation analysis was used to analyze interobserver and intraobserver variances. SPSS version 12.0 Win (SPSS, Inc, Chicago, Illinois) was used throughout.

Figure 4: Results of mechanical axis evaluation using the OrthoPilot and AxiEM navigation systems. OrthoPilot showed varus values of 0º, 1º, 2° for the entire experimental trial, whereas AxiEM showed varus values of 0º, 1º, 2º, 3°.

Results

Clinical Results

The mechanical axis was varus 9.45º ± 7.9° using weight-bearing anteroposterior radiographs, varus 9.02º ± 5.18° using the OrthoPilot navigation system, and varus 10.25° ± 5.10° using the AxiEM navigation system. AxiEM showed 1.23° more mean varus, but it had no statistical significance (P = .078). A difference greater than 3° occurred in 15% of cases (Table 1).

Experimental Results

The true mechanical axis of the synthetic bone was varus 1.25° by Orthodoc. OrthoPilot displayed varus 1.10° ± 0.64° change, and AxiEM displayed varus 1.78° ± 0.89° change. No mechanical axis differences were observed between the two navigations (P = .124) (Table 2). AxiEM provided greater varus than OrthoPilot. The mean difference was 0.68°.

OrthoPilot showed varus values of 0°, 1°, and 2° for the entire trial, whereas AxiEM showed varus values of 0º, 1º, 2°, and 3° for the entire trial. In 86% of trials, both navigation systems showed varus 1° or 2° of mechanical axis, which demonstrated relatively high accuracy and reproducibility of both navigation systems (Figure 4). No significant interobserver or intraobserver variance was detected. Pearson’s correlation coefficient ranged from 0.611 to 0.791 for interobserver variance and from 0.798 to 0.934 for intraobserver variance (P < .05).

Change of Mechanical Axis After the Erroneous Identification of Anatomic Landmarks

With OrthoPilot, 0.2° valgus or varus changes of the mechanical axis were observed with 10-mm medial or later side erroneous registration of the center of the distal femur (Table 3). However, AxiEM showed a 1.76° valgus or 1.62° varus change in the same study. For the erroneous 10-mm medial or lateral registration from the center of the proximal tibia, OrthoPilot showed 1.24° valgus or 1.43° varus change, whereas AxiEM showed 1.69° valgus or 1.72° varus change.

When a 10-mm medial or lateral side of the medial or lateral malleoli registration occurred, OrthoPilot showed 0.33° varus or 0.33° valgus change, and AxiEM showed 1° varus or 1° valgus change. On the other hand, for incorrect registration of the ankle center, OrthoPilot was significantly affected and had 1.69° varus or 1.62° valgus changes. AxiEM had no ankle center registration process.

OrthoPilot showed less aberration than AxiEM after intentionally erroneously identifying 10-mm medial and lateral side registrations from the center of the distal femur, proximal tibia, or bilateral ankle malleoli.

Discussion

Computer navigation has become an important technology, and in many reports navigation has reduced mechanical axis outliers after TKA.^1-5,12,13 Of the available navigation systems, infrared optical navigation has become widely used, and recently electromagnetic and ultrasound navigated systems were introduced.^7-9,18

Stiehl et al¹⁶ compared the accuracies of optical and electromagnetic navigation systems using a cadaver in a standard operating room, and the investigators reported that precision was satisfactory for both optical and electromagnetic tracking for mechanical axis assessment. However, outliers were identified with electromagnetic tracking, causing concern that accuracy could be affected by electromagnetic forces in the operating room. Until now, few studies have clinically compared the accuracies of these two navigation systems. We wanted to determine whether the new electromagnetic navigation system could measure mechanical axes as precisely as an optical navigation system under standard operating conditions.

In patients undergoing TKA, one operator measured preoperative mechanical axes using the two different navigation systems. The mechanical axis measured using two navigation systems was found to be different, although there was no statistical significance. The mean difference was 1.23°, and electromagnetic navigation showed more varus in the same patients. In 85% of patients, the mechanical axis differences measured using two navigation systems were within 2°, but differences of more than 3° were recorded in 15%.

These differences may be attributable to ambiguous anatomy with soft tissue coverage, which causes registration error, knee joint laxity with arthrotomy, and possible data outliers that result from signal interference by metallic surgical instrument and electrical devices in the operation room.

Yau et al¹⁹ investigated the intraobserver errors in obtaining visually selected anatomic landmarks that were used in the registration process and concluded that the maximum error in mechanical axis was 1.32° in the coronal plane and 4.17° in the sagittal plane in a cadaver study. In this study, the mechanical axis was measured only one time by both navigations. Thus, we cannot evaluate the intraobserver variation. However, some variation can be expected even in the same patients because of soft tissue coverage of anatomic landmarks.

The optical navigation system used in this study had been employed clinically for more than 9 years. Although the electromagnetic navigation system is increasingly implemented and has several advantages such as a small tracker that can be fixed on a surgical incision site with minimal trauma, it has no line of sight capability. The system also presents problems with interaction with ferromagnetic instruments or other electrical equipment in operation room.

Experimentally, the accuracies of optical and electromagnetic navigation systems are known to be within 1 mm or 1°.^1,9,17-21 For the accuracy of infrared optical navigation systems, Pitto et al¹ reported that the mean error of the system was within 0.5° in the setting of normal alignment and within 1.0° in the setting of abnormal plane alignment. Many reports have been published on electromagnetic navigation systems.^9,17,18,20 Hummel et al¹⁷ reported relatively accurate results for the Aurora electromagnetic system (Northern Digital Inc, Bakersfield, Calif) and found that the relative positional error was 0.97 mm and that its rotational error was 0.2-0.91°. However, it was also found that significant distortion can occur by interaction with metal (most significantly by 400 series stainless steel). Electromagnetic interference in the operating field had lead to newly developed electromagnetic systems that improved accuracy. Schicho et al²¹ studied the effect of metal instruments on the Aurora electromagnetic navigation system, which caused a mean 1.44-mm distance error when a Langenbeck hook was applied, a mean 0.53-mm distance error when a drill was applied, and a mean 2.37-mm distance error when an ultrasound scan head was applied. In addition, they reported findings for identical experiments using the Treo-EM system, and found a mean 0.21-mm distance error for a Langenbeck hook, 0.23-mm distance error for a drill, and 0.56-mm distance error for an ultrasound scan head.

In this trial, we experienced some discordance between the two systems for measuring the mechanical axis and thus investigated the accuracy of both systems under experimental conditions. To eliminate the interference of metal, the hip, knee, and ankle joints of the synthetic saw bone were made of titanium, which has no effect on an electromagnetic system. Further, the knee joint was constrained to not allow varus or valgus motion for precise measurement of mechanical axis of the synthetic bone model.

As a result, the true mechanical axis of bone model was varus 1.25° for the Orthodoc system, varus 1.1° for the OrthoPilot system, and varus 1.78° for the AxiEM system, which is not significantly different, indicating that both navigation systems are accurate. No intraobserver and interobserver differences were found, which meant both systems had high reproducibility. Further, relatively obvious anatomic structures of the bone model are thought to contribute to high reproducibility of both systems by reducing registration errors compared with some variability in an in vivo study with ambiguous anatomic landmarks. Even though in terms of numerical values, the mean difference between the two navigation system was 0.68°, which does not seem to be a significant difference in measuring the mechanical axis of the bone model. OrthoPilot showed 0°, varus 1°, and varus 2°, whereas AxiEM showed 0°, varus 1°, varus 2°, and varus 3°, indicating that OrthoPilot has better reproducibility with less variability. This study revealed that under the least favorable conditions, the two navigation systems could show a difference of 3° (for example, OrthoPilot 0° and AxiEM varus 3°). Furthermore, experimentally AxiEM yielded higher varus values, which could affect postoperative mechanical axis correction leading to valgus over correction, clinically.

When anatomic registration was incorrect (10-mm registration error study), mechanical axis measurements were affected in both navigation systems. AxiEM was affected more in every step (range, 1.0° to 1.76°). OrthoPilot was less affected (range, 0.2° to 1.69°), but it also had a significant change in mechanical axis measurement especially when proximal tibia and ankle centers were registered inadequately. Surgeons should therefore collect precise anatomic landmarks during the registration process to reduce potential errors in using navigation system.

Conclusion

In this study, infrared optical navigation and electromagnetic navigation systems were found to be accurate and reproducible in an experimental environment. However, the accuracies of both systems were affected by erroneous registration, and the levels of inaccuracy encountered were high for the electromagnetic system. Under clinical conditions, discordances between the two navigation systems were observed and may have been attributable to ambiguous anatomic points that cause registration errors and the possibility of electromagnetic signal interference in the operating environment.

References

1. Pitto RP, Graydon AJ, Bradley L, et al. Accuracy of computer-assisted navigation system for total knee replacement. J Bone Joint Surg Br. 2006;88:601-605.
2. Stulberg SD, Loan P, Sarin V. Computer-assisted navigation in total knee replacement results of an initial experience of thirty five patients. J Bone Joint Surg Am. 2002;84:90-98.
3. Macule-Beneyto F, Hernandez-Vaquero D, Segur-Vilalta JM, et al. Navigation in total knee arthroplasty. A multicenter study. Int Orthop. 2006;30:536-540.
4. Saragaglia D, Picard F, Chaussard C, et al. Computer-assisted knee arthroplasty: comparison with a conventional procedure. Results of 50 cases in a prospective, randomized study. Rev Chir Orthop Reparatrice Appar Mot. 2001;87:18-28.
5. Stoeckl B, Nogler M, Rosiek R, et al. Navigation improves accuracy of rotational alignment in total knee arthroplasty. Clin Orthop Relat Res. 2004;26:180-186.
6. Bolognesi M, Hofmann A. Computer navigation versus standard instrumentation for TKA: a single-surgeon experience. Clin Orthop Relat Res. 2005;440:162-169.
7. Amin DV, Kanade T, DiGioia AM III, Jaramaz B. Ultrasound registration of the bone surface for surgical navigation. Comput Aided Surg. 2003;8:1-16.
8. Chen TK, Abolmaesumi P, Pichora DR, Ellis RE. A system for ultrasound-guided computer-assisted orthopaedic surgery. Comput Aided Surg. 2005;10:281-292.
9. Rosenow JM, Sootsman WK. Application accuracy of an electromagnetic field-based image-guided navigation system. Stereotact Funct Neurosurg. 2007;85:75-81.
10. Huitema RB, Hof AL, Postema K. Ultrasonic motion analysis system-measurement of temporal and spatial gait parameters. J Biomech. 2002;35:837-842.
11. Khadem R, Yeh CC, Sadeghi-Tehrani M, et al. Comparative tracking error analysis of five different optical tracking systems. Comput Aided Surg. 2000;5:98-107.
12. Anderson KC, Buehler KC, Markel DC. Computer assisted navigation in total knee arthroplasty: comparison with conventional methods. J Arthroplasty. 2005;20:132-138.
13. Bathis H, Perlick L, Tingart M, et al. Alignment in total knee arthroplasty. A comparison of computer-assisted surgery with the conventional technique. J Bone Joint Surg Br. 2004;86:682-687.
14. Jenny JY, Boeri C, Picard F, Leitner F. Reproducibility of intraoperative measurement of the mechanical axes of the lower limb during total knee replacement with a non-image-based navigation system. Comput Aided Surg. 2004;9:161-165.
15. Jenny JY, Clemens U, Kohler S, et al. Consistency of implantation of a total knee arthroplasty with a non-image-based navigation system: a case-control study of 235 cases compared with 235 conventionally implanted prostheses. J Arthroplasty. 2005;20:832-839.
16. Stiehl JB, Heck DA. Six sigma analysis of computer-assisted surgery tracking protocols in TKA. Clin Orthop Rel Res. 2007;464: 105-110.
17. Hummel J, Figl M, Birkfelner W, et al. Evaluation of a new electromagnetic tracking system using a standardized assessment protocol. Phys Med Biol. 2006;51:N205-N210.
18. Lionberger R. The attraction of electromagnetic computer-assisted navigation in orthopaedic surgery. In: Stiehl JB, Konermann W, Hacker R, eds. Navigation and MIS in Orthopaedic Surgery. Heidelberg, Germany: Springer Verlag; 2006:44-53.
19. Yau WP, Leung A, Chiu KY, et al. Intraobserver errors in obtaining visually selected anatomic landmarks during registration process in nonimage-based navigation-assisted total knee arthroplasty: a cadaveric experiment. J Arthroplasty. 2005;20:591-601.
20. Poulin F, Amiot LP. Interference during the use of an electromagnetic tracking system under OR conditions. J Biomech. 2002;35(6): 733-737.
21. Schicho K, Figl M, Donat M, et al. Stability of miniature electromagnetic tracking systems. Phy Med Bio. 2005;50:2089-2098.

Authors

Drs Song, Seon, Park, and Yoon are from the Center for Joint Disease, Chonnam National University Hwasun Hospital, Jeonnam, Korea.

Dr. Song discloses a relationship with the B. Braun Aesculap speakers bureau. Drs Seon, Yoon, and Park have no relevant financial relationships to disclose.

Correspondence should be addressed to: Sang Jin Park, MD, Center for Joint Disease, Chonnam National University Hwasun Hospital, 160 Ilsimri, Hwasuneup, Hwasungun, 519-809, Jeonnam, Korea.