Abstract
This paper describes the features of KIN-nav, a navigation system specifically dedicated to intra-operative evaluation of knee laxity, and assesses the reliability of the system during surgery. The acquisition protocol for its intra-operative use, the original user interface, and the computational methods for elaboration of kinematic data are described in detail. Moreover, an extensive and specific validation of the system was performed in order to evaluate its intra-operative performance and usability. KIN-nav's reliability and accuracy were analyzed in a series of 79 patients undergoing ACL reconstruction.
The intra-surgeon repeatability computed for ACL-deficient and reconstructed knees at different flexion angles was less than 0.6° for varus-valgus (VV) rotation, less than 1 mm for AP translation, and less than 1.6° for IE rotation. The inter-surgeon repeatability is less than 2° for VV rotation, 5° for internal-external rotation, and less than 3 mm for AP translation.
The proposed method was fast (requiring an additional 10 minutes of surgical time on average), required only a short learning period (5 cases), was minimally invasive, and was robust from the numerical perspective.
Our system clearly shows that the use of navigation systems for kinematic evaluation provides useful and complete information on the knee state and test performance, and is simple and reliable to use. The good repeatability in manual kinematic tests is an improvement on the present ability to discriminate knee kinematics intra-operatively, and thus offers the possibility of better discrimination between knee pathologies and the prospect of new surgical applications.
Introduction
Navigation systems are becoming increasingly popular in orthopaedic surgery, and are being used to address a growing number of applications, including different types of surgical intervention on various joints and new types of computation.
Knee surgery was one of the first applications of navigation systems and remains an important target area for this technology. Many commercial navigation systems assist the surgeon in performing geometrical evaluations, such as alignment and prosthesis positioning in total knee reconstruction, determination of the proper direction for screw insertion on external fixators, or selection of cutting planes in osteotomies Citation[1–5]. Kinematic evaluations are also performed in some navigation systems, e.g., to help select the most isometric placement of the graft, or to predict ligament elongation and impingement during ACL reconstruction Citation[6–8]. At present, however, the ability of navigation systems to track bones and joint motion is seldom used, and quantitative elaboration of motion data is rare, even though functional assessment of the surgical result is a critical step in most interventions. In the present surgical routine, evaluation of knee motion and stability is performed during and after all types of reconstruction, and can influence the surgeon's decision regarding further treatment or surgical actions Citation[9].
Recently, a new interest in navigated intra-operative kinematic evaluations seems to have developed, and some initial clinical applications have already been reported Citation[10–12]. However, the use of navigation systems in knee kinematic evaluation remains neglected. Our laboratory has developed a navigation system, called KIN-nav, which is designed with an original user interface for intra-operative real-time assessment of knee laxity. The development of KIN-nav follows the new philosophy of navigation systems designed specifically for kinematic evaluation.
This paper describes the features of KIN-nav, which is specifically dedicated to intra-operative evaluation of knee laxity, and an evaluation of its reliability during actual surgery. The acquisition procedure is summarized, along with the computational methods for elaboration of kinematic data, which represent an optimization of the protocol proposed in reference Citation[11], taking into account suggestions arising from an ongoing clinical trial in order to make the system user-friendly, fast and accurate.
The reliability and accuracy of KIN-nav were investigated first by conducting an analysis of sensitivity to experimentally introduced errors using simulation and numerical analysis, and then by conducting an experimental analysis of the repeatability of the measurements on 79 patients undergoing ACL reconstruction.
Material and methods
KIN-nav integrates an optoelectronic localizer (Polaris, Northern Digital, Inc., Waterloo, Ontario, Canada), which is able to track the position and orientation of different reference arrays, with a custom software developed in accordance with the clinical application requirements. The system is used according to a specific protocol for intra-operative knee kinematics assessment, as described below.
System description
The navigated protocol consists of two phases. The first phase is the anatomical reference acquisition phase: At the beginning of surgery, two reference arrays are mounted on the femur and tibia via the normal surgical incisions for arthroscopic access and graft harvesting and fixation, using 3-mm screws to reduce morbidity (). This particular configuration does not interfere with the surgical instruments or maneuvers and is suitable for routine surgical use. During the intervention, prior to kinematic testing, the surgeon identifies standard anatomical landmarks on the limb and performs percutaneous acquisition of the hip center, epicondyles, medial and lateral tibial plateaus and malleoli.
Figure 1. Percutaneous acquisition of lateral epicondyles during the registration phase. Reference arrays are mounted on the femur and tibia through the skin incision made for ACL reconstruction.
The second phase is the kinematic evaluation of joint laxity through standard clinical maneuvers performed intra-operatively, before and after the surgical procedure. For ACL reconstruction cases, the surgeons performed varus-valgus (VV) stress tests at 0° (full leg extension) and 30° of flexion, internal-external (IE) rotation tests at 30° and 90° of flexion, and antero-posterior (AP) stress tests at 30° and 90° of flexion, all with maximum force.
Data are elaborated in real time to compute the knee state and knee laxity. The femoral anatomical reference system was defined by setting the Z-axis as the femoral mechanical axis (with the hip center identified through a least-square optimization algorithm, as described in references Citation[13] and Citation[14]), the X-axis as the transepicondylar line normalized with respect to the Z-axis, the Y-axis as the cross product between the Z- and X-axes, and the origin as the midpoint between the epicondyles. Similarly, the tibial anatomical reference system was defined by setting the Z-axis as the tibial mechanical axis, the X-axis as the line joining the tibial plateau extremities normalized with respect to the Z-axis, the Y-axis as the cross product between the Z- and X-axes, and the origin as the midpoint between the epicondyles. The 6 degrees of freedom of the knee joint were computed from the relative motion of the tibial frame with respect to the femoral one using Euler decomposition in the sequence X-Y-Z to compute the instantaneous rotations and the Grood and Suntay method to compute instantaneous displacements Citation[15], Citation[16]. Laxity value is computed as the difference between maximal and minimal rotational values achieved during the IE and VV tests, and as the difference between maximal and minimal displacement during the AP test.
The user interface of KIN-nav guides the surgeon through the acquisition steps and shows the results of the computations via a single window interface operated by buttons.
The left side of the interface contains the field of view of the tracker (at top), a section with the command buttons (in the middle) and a section in which the tools of the localizer can be activated (at the bottom) (see ).
Figure 2. Software interface during (a) the anatomical registration phase and (b) antero-posterior laxity tests.
The main, central part of the interface display contains the interaction buttons, the results of the algorithm, such as limb position and kinematic tests, and the 3D visualization of the limb. At present, the central panel contains only two states, corresponding to two different phases during navigated surgery: the first state is for anatomical reference acquisitions and the second is for kinematic acquisitions.
In the first state (), the software suggests the sequence of reference points to be acquired and must be run before switching to the kinematic test state. In this state, the 3D view is centered in the field of view of the localizer and shows three concentric circles in order to help the operator center the cameras.
When the anatomical acquisition phase is terminated, the software switches automatically to the kinematic acquisition state (). In this phase, the software interface shows, at upper right, the main axes of the limb as a schematic representation of the acquired anatomy in two OpenGL frames, corresponding to the frontal and lateral limb views. In these frames, the mechanical axes of the segments are shown, and it is possible to have qualitative visual feedback of the previous anatomical acquisitions. In the lower part of the screen, the numerical values of the angles between the fixed and mobile segments are shown in real time. These values help the surgeon to place the limb in the correct initial position, called the neutral position, before performing the kinematic tests. Between the OpenGL frames and the numerical display there is a black box that displays the computed laxity during the performance of the kinematic tests. This value is important and must be clearly visible during the intervention; it therefore occupies a suitable portion of the screen.
In this second state, the buttons for the acquisition of kinematic tests are activated. Near the test buttons, in the central panel, the results of each test are visualized. This sequence of results is implemented to enable the surgeon to obtain a complete evaluation of the joint kinematics and to compare the pre- and post-operative tests, in order to evaluate the final outcome of the intervention.
System validation
The development of a navigation system requires a complex validation process. In the following, we describe the assessment of each requirement.
The clinical phase of our study involved 79 consecutive patients (15 female and 64 male; mean age: 32 years [range: 17–49]) who underwent arthroscopic ACL reconstruction at our institute after being diagnosed with anterior knee joint instability. Patients were enrolled between March 2005 and March 2006, and were treated by the expert surgical staff of our department with well-known surgical techniques and equipment. Thirty-nine patients underwent a single-bundle reconstruction and 40 underwent a double-bundle reconstruction using the hamstring tendon technique Citation[17], Citation[18]. The study design was approved by the Institutional Review Board of the Institute, and all patients provided informed consent.
Descriptive statistics of all tests were obtained and normal distribution of data was verified with the Shapiro--Wilk test. To assess intra-operator repeatability, an expert surgeon, blinded to the laxity results, repeated the kinematic tests 3 times in 30 cases. Intra-operator repeatability was evaluated by computing the average standard deviation of the tests with 1-way ANOVA.
To assess inter-operator repeatability, two expert surgeons and one non-expert surgeon, blinded to the laxity results, repeated the kinematic tests during 30 cases, both with the knees in their pathological condition and after reconstruction. Differences between surgeons were evaluated using pairwise 1-way ANOVA with the LSD error protection method.
For statistical analysis, the significance level was set at 95%.
The intrinsic sensitivity of the two kinematic algorithms to the definition of the anatomical reference systems was analyzed by introducing random errors of varying amplitude and direction in the identification of anatomical landmarks. Starting from the literatu Citation[19], Citation[20], we specifically introduced errors of 40 mm in hip joint center identification and 10 mm in the identification of femoral epicondyles, and medial and lateral points of tibial plateaus and malleoli.
Furthermore, the surgical performance of the novel procedure was evaluated by monitoring the additional time required for the surgery and observing the clinical outcome.
A report on software failures and intra-operative abnormalities was compiled after each use and an interview on clinical usability and efficacy was conducted with five surgeons, in order to obtain both objective and subjective data on the advantages and pitfalls of this methodology and its potential clinical benefits.
Results
Results of laxity tests on patients were found to be normally distributed (Shapiro--Wilk coefficient > 0.91) and were comparable with published in vivo data in the literature Citation[21–25] (see ).
Table I. Average results of laxity test on patients.
The intra-surgeon repeatability computed for ACL-deficient and reconstructed knees at different flexion angles is reported in . To summarize the data obtained for each test, intra-surgeon repeatability was less than 0.6° for VV rotation; less than 1 mm for AP translation; and less than 1.6° for IE rotation.
Table II. Intra-surgeon repeatability of laxity tests. Standard deviation (St dev) of the repeated tests by the same surgeon in ACL-deficient (pre) and reconstructed (post) knees. These values were computed with ANOVA and the significance (p) is reported.
The inter-surgeon repeatability computed for ACL-deficient and reconstructed knees at different flexion angles is reported in . Inter-operator repeatability between all surgeons, expert and non-expert, was less than 2° for VV rotation, 5° for IE rotation, and less than 3 mm for AP translation. The ANOVA showed that there was no statistical difference between the surgeons’ values, except for post-operative VV rotation at 0° of flexion.
Table III. Inter-surgeon repeatability of laxity tests. Median results of laxity tests for two expert surgeons (1 and 2) and one non-expert surgeon (3) in ACL-deficient (Pre) and reconstructed (Post) knees. Differences between surgeons were evaluated with pairwise 1-way ANOVA; the significance (p) of each test is reported.
The effect on the computed laxities of introducing random error of up to 10 mm for anatomical landmarks on the knee and up to 40 mm for the hip center is reported in . The simulation showed that the initial error is not amplified by this elaboration, in either rotational or translational values.
Table IV. Results of error propagation in kinematic tests due to errors in reference point acquisition of 40 mm for the hip center and 10 mm for other anatomical landmarks.
The proposed method proved to be fast (requiring an additional 10 minutes of surgical time on average) and required only a short learning time (5 cases). It is barely invasive from the perspective of the patients and surgical staff, as it requires no additional incisions and does not modify the standard surgical technique, except for the short setup phase of the navigated arrays (which requires an average of 3 minutes). No serious complications were reported post-operatively; in 9 cases, some swelling appeared in the area of the pins during the first week after surgery.
The navigated measurements were successfully performed in 75 cases, but were suspended in 4 cases: in 2 cases because of unknown communication problems, and in 2 cases because of software bugs that were later corrected. The risk of failure was found to be very low (< 3%). In the few cases of KIN-nav failure, recovery was always possible after the preliminary trial period. Software crashes could be easily recovered from by resetting the communication; acquisition errors for anatomical landmarks or kinematic maneuvers could be detected immediately by examining the limb alignment or the result of the first sample kinematic test, whereupon the anatomical landmarks were corrected and the measurement repeated.
Discussion
We have described a new navigation system, KIN-nav, which is dedicated to the intra-operative measurement of laxity, and have reported the evaluation of its features and performance by numerical analysis and clinical tests during 79 consecutive ACL reconstructions.
At present, the validation of navigation systems does not follow a set of standard guidelines, and is usually focused on the evaluation of the clinical outcome. In this study, we mainly addressed the intra-operative performance and usability of the system, which we believe to be necessary before a new product enters extensive use.
KIN-nav has clearly shown that the use of navigation systems for kinematic evaluations presents several interesting aspects.
First of all, KIN-nav is easy to use; both senior and less-experienced surgeons used it for laxity evaluations, and all showed a short learning curve. KIN-nav was minimally invasive for the patient, fast to use in clinical routine, did not interfere with the standard surgical steps during arthroscopic ACL reconstruction, and was judged to be user-friendly and helpful for laxity evaluations by all surgeons and the entire surgical staff.
Such a dedicated system has the advantage of being simpler and faster than other navigation systems for knee surgery, and focuses on a new surgical action – kinematic evaluation of the joint – which is, at present, mostly evaluated in a qualitative and subjective manner by surgeons.
KIN-nav was tested clinically in patients undergoing ACL reconstruction, and the measured laxity values were in agreement with those in the literature and with surgeons’ expectations Citation[20], Citation[23–26].
An important and original finding of this study is the estimation of KIN-nav's repeatability and sensitivity for surgical applications. The repeatability of AP laxity was 1 mm when used by an expert surgeon and 3 mm when used by different surgeons, including a non-expert surgeon; the repeatability of IE laxity was 1.6° for a single examiner and 5° for different examiners; while the repeatability of VV laxity was 0.6° for a single examiner and 2° for different examiners. These results are very encouraging when compared with those for other instruments used for manual kinematic tests. In fact, KIN-nav showed a repeatability for AP laxity that is double with respect to that for KT or Rolimeter arthrometers Citation[27–31]. Moreover, KIN-nav also provided a simple quantitative evaluation of rotational laxity, which is more difficult to estimate intra-operatively, and the uncertainty in this type of evaluation was found to be better than the repeatability of in vivo instrumented tests reported by other groups Citation[12], Citation[23]. The least repeatable test was the evaluation of IE laxity, even when the test was apparently conducted along the tibial anatomical axis and with a constant flexion angle.
The good repeatability of manual kinematic tests that we obtained may be due to the fact that KIN-nav provides the surgeons with real-time feedback on the knee condition and thus helps to control the knee flexion angle at the initial position for manual kinematic tests.
It should be noted that the numerical analysis of the implemented algorithms for laxity computation showed that the standard algorithms for knee kinematic decomposition can be used reliably for computation of both translation and rotational laxities. It is known that kinematic decomposition is influenced by the acquired reference system, but the decomposition of laxity values does not amplify the uncertainty in the acquisition of anatomical landmarks and is suitable for repeatable measurements Citation[32], Citation[33].
These results suggest that the use of a navigation system for laxity evaluation may improve on the present ability to discriminate knee kinematics intra-operatively, thus offering the possibility of better discrimination between knee pathologies and the prospect of new surgical techniques.
The present implementation of KIN-nav confirmed that navigated laxity measurements are a robust and accurate methodology for intra-operative evaluation, being able to provide a reliable and complete description of the knee state during different steps of surgery and an immediate quantitative assessment of surgical performance. The proposed method is therefore suitable for routine estimation and documentation of the surgical procedure, and can help surgeons to better assess the need for further treatment or optimize the result of a reconstruction, in addition to analyzing different factors that influence the final surgical outcome.
The main limitation of this methodology is that it is very expensive and requires the rigid fixation of rigid arrays to the bone, which may produce effusion and post-operative complications. These limitations may be improved by ongoing and future technical advances in the design of navigation systems and associated tools.
Further clinical studies will be performed with KIN-nav, especially using its high sensitivity (3 mm for AP laxity; 5° for IE laxity; 2° for VV laxity) to investigate differences between surgical techniques or individual kinematic features.
Acknowledgments
The authors would like to thank all the nurses and staff at “Casa di Cura Madre Fortunata Toniolo” (Bologna) for their collaboration in the experimental acquisitions; Dr. F. Iacono, Dr. M. Lo Presti, Dr. A. Russo, Dr. M.P. Neri, and Dr E. Kon for their support of the clinical research and KIN-nav validation; and the personnel of the Biomechanics Lab, particularly G. Bernagozzi, M. Bonfiglioli and C. Carcasio.
References
- Cossey AJ. Spriggins. The use of computer-assisted surgical navigation in preventing malalignment in knee arthroplasty. Proceedings of the Fourth Annual Meeting of the International Society for Computer Assisted Orthopaedic Surgery (CAOS-International 2004), F Langlotz, BL. Davies, SD Stulberg, Chicago, IL June 2004; 2–3
- DiGioia A, Jaramaz J, Picard F, Nolte L-P. Computer and robotic assisted knee and hip surgery. Oxford University Press, New York 2004
- Hankenmeier S, Hufner T, Wang G, Kendoff D, Zeichen J, Zheng G, Krettek C. Navigated open-wedge high tibial osteotomy: Advantages and disadvantages compared to the conventional technique in a cadaver study. Knee Surg Sports Traumatol Arthrosc 2006; 14(10)917–921
- Wang G, Zheng G, Keppler P, Gebhard F, Staubli A, Mueller U, Schmucki D, Fleutsch S, Nolte LP. Implementation, accuracy evaluation, and preliminary clinical trial of a CT-free navigation system for high tibial opening wedge osteotomy. Comput Aided Surg 2005; 10(2)73–85
- Hankenmeister S, Hufner T, Wang G, Kendoff D, Zheng G, Richter M, Gosling T, Nolte L, Krettek C. Navigated intraoperative analysis of limb alignment. Arch Orthop Trauma Surg 2005; 125(8)531–535
- Sati M, Stäubli H-U, Bourquin Y, Kunz M, Nolte L-P. Real-time computerized in situ guidance system for ACL graft placement. Comput Aided Surg 2002; 7: 25–40
- Sati M, Deguise JA, Droving G. Computer assisted knee surgery: Diagnosis and planning of knee surgery. Comput Aided Surg 1997; 2(2)106–123
- Picard F, DiGioia AM, Moody J, Martinek V, Fu FH, Rytel M, Nikou C, LaBarca RS, Jaramaz B. Accuracy in tunnel placement for ACL reconstruction. Comparison of traditional arthroscopic and computer-assisted navigation techniques. Comput Aided Surg 2001; 6(2)279–289
- Müller W. The Knee: Form, Function, and Ligament Reconstruction. Springer, Berlin 1983
- Colombet P, Allard M, Granchi C, Plaskos C, Lavallée S. Anterior cruciate ligament reconstruction using knee joint laxity measurements and a bendable ligament model. Proceedings of the Fifth Annual Meeting of the International Society for Computer Assisted Orthopaedic Surgery (CAOS-International 2005), F Langlotz, BL. Davies, D Schlenzka, HelsinkiFinland June 2005; 69–71
- Zaffagnini S, Martelli S, Bignozzi S, Imakiire N, Marcacci M. Navigated intraoperative clinical test for kinematic assessment of ACL graft behavior. Proceedings of the Fifth Annual Meeting of the International Society for Computer Assisted Orthopaedic Surgery (CAOS-International 2005), F Langlotz, BL. Davies, D Schlenzka, HelsinkiFinland June 2005
- Ishibashi Y, Tsuda E, Tazawa K, Sato H, Toh S. Intraoperative evaluation of the anatomical double-bundle anterior cruciate ligament reconstruction with the OrthoPilot Navigation System. Orthopaedics 2005; 8: 1277–1282
- Cannon W, Dilworth MD. Use of arthrometers to assess knee laxity and outcomes. Sports Medicine & Arthroscopy Review 2002; 10(3)191–200
- Ganko L, Engebretsen L, Ozer H. The Rolimeter: a new arthrometer compared with the KT-1000. Knee Surg Sports Traumatol Arthrosc 2000; 8(1)36–39
- Wu. G, Siegler S, Allard P, et al. ISB recommendation on definition of joint coordinate system of various joints for the reporting of human joint motion – Part I: Ankle, hip, and spine. J Biomech 2002; 35(4)543–548
- Grood ES, Suntay WJ. A joint coordinate system for the clinical description of three-dimensional motions: Application to the knee. ASME J Biomech Eng 1983; 105: 136–144
- Marcacci M, Zaffagnini S, Iacono F, Neri MP, Loreti I, Petitto A. Arthroscopic intra- and extra-articular anterior cruciate ligament reconstruction with gracilis and semitendinosus tendons. Knee Surg Sports Traumatol Arthrosc 1998; 6(2)68–75
- Marcacci M, Molgora AP, Zaffagnini S, Vascellari A, Iacono F, Presti ML. Anatomic double-bundle anterior cruciate ligament reconstruction with hamstrings. Arthroscopy 2003; 19(5)540–546
- Nizard R. Computer assisted surgery for total knee arthroplasty. Acta Orthop Belg 2002; 68(3)215–230
- Jenny SY, Boeri C. Low reproducibility of the intra-operative measurements of the transepicondylar axis during total knee replacement. Acta Orthop Scand 2004; 75(1)74–77
- Dennis DA. In vivo determination of normal and anterior cruciate ligament-deficient knee kinematics. J Biomech 2005; 38(2)241–253
- Shino K. Measurement of anterior instability of the knee. J Bone Joint Surg Br 1987; 69(4)608–613
- Shoemaker SC. In vivo rotatory knee stability. J Bone Joint Surg Am 1982; 64(2)208–216
- Schuster AJ, McNicholas MJ, Wachtl SW, McGurty DW, Jakob RP. A new mechanical testing device for measuring anteroposterior knee laxity. Am J Sports Med 2004; 32: 1731–1735
- Markolf LK, Kochan A, Amstutz HC. Measurement of knee stiffness and laxity in patients with documented absence of the ACL. J Bone Joint Surg Am 1984; 66(2)242–252
- Pedowitz RA, Popejoy DJ. Initial laxity does not correlate with subsequent laxity after ACL reconstruction in humans: A prospective evaluation. Am J Orthop 2004; 33(11)560–564
- Pollet V, Barrat D, Meirhaeghe E, Vaes P, Handelberg F. The role of the Rolimeter in quantifying knee instability compared to the functional outcome of ACL-reconstructed versus conservatively-treated knees. Knee Surg Sports Traumatol Arthrosc 2005; 13(1)12–18
- Forster I, Warren-Smith C, Tew M. Is the KT-1000 knee ligament arthrometer reliable?. J Bone Joint Surg Br 1989; 71: 843–847
- Edixhoven P, Huiskes R, de Graaf R, van Rens TJ, Slooff TJ. Accuracy and reproducibility of instrumented knee-drawer tests. J Orthop Res 1987; 5(3)378–387
- Daniel D, Malcolm L, Losse G. Instrumented measurement of anterior laxity of the knee. J Bone Joint Surg Am 1985; 67: 720–726
- Boyer P, Djian P, Christel P, Paoletti X, Degeorges R. Reliability of the KT-1000 arthrometer (Medmetric) for measuring anterior knee laxity: Comparison with Telos in 147 knees. Rev Chir Orthop Reparatrice Appar Mot 2004; 90(8)757–764
- Piazza SJ, Cavanagh PR. Measurement of screw-home motion of the knee is sensitive to errors in axis alignment. J Biomech 2000; 33: 1029–1034
- Pennock SY, Clark KJ. An anatomy-based coordinate system for the description of the kinematic displacements in human knee. J Biomech 1990; 21(12)1209–1218