ABSTRACT. Introduction: Balancing at surgery is important for clinical outcome in terms of pain relief, flexion range, and function. The methodology usually involves making bone cuts to achieve correct leg alignment, and then obtaining equal gaps in extension and flexion using spacer blocks or tensor devices. In this study, we describe a method for quantifying balancing throughout the flexion range and show the effect of different surgical corrections from an unbalanced to a balanced state. In this way, we quantified how accurately balancing could be achieved within the practical time frame of a surgical procedure. Methods: Data was obtained from 80 primary procedures using a PCL-retaining device. Initial bone cuts were made using navigation. Instrumented tibial trials were used to measure the contact forces and locations on the lateral and medial sides. Video/audio recordings were made of all aspects of the surgeries. The initial balancing was recorded during the Heel Push Test, namely the lateral and medial contact forces for the flexion range. The data was expressed as medial/total force ratio (total=medial + lateral), with 0.5 being equal lateral and medial forces. Surgical corrections to correct the specific imbalance pattern, determined from previous research, were carried out. The Heel Push Test was repeated after each correction and at final balancing. Results: The initial balancing before correction showed that although the average ratio was 0.52+/-0.27 from 0-90 degrees, the data was scattered between 0.0 (lateral force only) and 1.0 (medial force only). The most common surgical corrections used to achieve balancing were: soft-tissue releases (49), changes in tibial insert thickness (27), bone adjustments (15), tibial rotational adjustments (7). In 84% of the cases, 0-2 corrections were needed to obtain balancing (Range: 0-5). 80% of the cases in early flexion (0-30 degrees) were balanced within 15% of the balanced state (79 % for 30-60 deg of flexion, 77% for 60-90 deg of flexion). The mean ratio for all flexion angles was 0.52 with standard deviation 0f 0.16. Discussion: By following a set of logical steps, accurate balancing was achieved in the majority of cases with only 1-2 Surgical Corrections being necessary. Corrections of both bone cuts and soft tissue were applicable. The most important range for balancing was early in flexion. There was no target value of the total forces because the ligament stiffnesses varied substantially between patients. Further, while a 0.5 ratio was aimed for, we expect that the ideal value will be in the region of 0.6, higher medial forces than lateral, in line with functional forces. Studies are now underway to determine the effect of balancing on the functional results.

ABSTRACT. Introduction: The role of soft tissue balancing in optimizing functional outcome and patient satisfaction after total knee arthroplasty surgery is gaining interest. Consistent soft tissue balancing has been aided by novel technologies that can quantify loads across the joint at the time of surgery. Based on free body diagram theory, compressive load equilibrium should be correlated with ligamentous equilibrium between the medial and lateral collateral ligaments. The authors propose to use the Collateral Ligaments Strain Ratio (CLSR) as a functional tool to quantify and track the effectuated surgical change in laxity of the medial and lateral collateral ligaments and correlate this ratio to validated functional scores and patient reported outcomes. The relationship with intra-operative balancing of compartmental loads can then be scrutinized. Methods: The study is a prospective clinical study with three cohorts of 50 patients each: (1) a surgical prospective study group with ligamentous testing pre-operatively, at 4 weeks, 3 months and 6 months post-operatively; (2) a matched control group of non-operated high function patients; (3) a matched control group of high function knee arthroplasty recipients. Standard statistical analysis method is applied. The testing of the CLSR is performed using a validated Smart Knee Brace developed by the authors and previously reported. The output variables consist of the maximum angular change of the knee in the coronal plane at 10 degrees of flexion produced by a controlled torque application in the varus and valgus (VV) directions. This creates measureable strain on the lateral and medial collateral ligaments, which is reported as a ratio (CLSR). The New Knee Society Score is used to track outcomes. The intra-operative balance is achieved by means of an instrumented tibial tray (OrthoSensor,Inc). The applied torque was standardized to 10Nm with a handheld wireless dynamometer. Results: Pre-operative scatter graphs demonstrate a wide distribution of absolute VV values, reflecting the spectrum of pathological states. The relative distribution of strain after surgery trends towards consolidation. The median CLSR is 0.55 with a SD of 0.20. This asymmetrical value indicates a shift toward a tighter medial side. The overall displacement values range from 0 -4 degrees. The angular changes under standard torque appear to correlate with the linear displacement values previously reported by Bellemans et al.

ABSTRACT. Objectives: To compare computer-assisted total knee arthroplasty with the conventional technique in operative time. Materials and Methods: 30 patients with different degrees and forms of knee osteoarthritis were divided into 2 groups. Group 1 (15 patients) had TKA using patient specific instrumentations (PSI). Group 2 (15 patients) had TKA using the conventional technique. Operative time was measured for each patient of each group. Results: In comparison to conventional group, patients in the PSI group had shorter operative time by 24.3 minutes which is statistically significance. Conclusion: PSI technique has advantage over conventional instrumentation as it reduces operative time.

ABSTRACT. Total knee arthroplasty (TKA) is an effective technique to treat end-stage knee osteoarthritis, targeting the restore a physiological knee kinematics. However, studies have shown abnormal knee kinematics after TKA which may lead to suboptimal clinical outcomes. Posterior slope of the tibial component may significantly impact the knee kinematics. There is currently no consensus about the most appropriate slope. The goal of the present study was to analyze the impact of different prosthetic slopes on the kinematics of a PCL-preserving TKA, with the hypothesis that posterior slopes can alter the knee kinematics.

A PCL-retaining TKA (Optetrak CR, Exactech, Gainesville, FL) was performed by a board-certified orthopedic surgeon on one fresh frozen cadaver that had a non arthritic knee with an intact PCL. Intact knee kinematic was assessed using a computer-assisted orthopedic surgery (CAOS) system (ExactechGPS®, Blue-Ortho, Grenoble, FR) Then, TKA components were implanted using the guidance of the CAOS system. The implanted tibial baseplate was specially designed to allow modifying the posterior slope without repeatedly removing/assembling the tibial insert with varying posterior slopes, avoiding potential damages to the soft-tissue envelope. Knee kinematic was evaluated by performing a passive range of motion 3 separate times at each of the 4 posterior slopes: 10°, 7°, 4° and 1°, and recorded by the navigation system. Femorotibial rotation, antero-posterior (AP) translation and hip-knee-ankle (HKA) angle were plotted with regard to the knee flexion angle.

Tibial slopes of 1° and 4° significantly altered the normal rotational kinematics. Tibial slopes of 7° and 10° led to a kinematics close to the original native knee. All tibial slopes significantly altered the changes in HKA before 90° of knee flexion, without significant difference between the different slopes tested. The magnitude of change was small. There was no significant change in the AP kinematics between native knee and all tested tibial slopes.

Changing the tibial slope significantly impacted the TKA kinematics. However, in the implant studied, only the rotational kinematics were significantly impacted by the change in tibial slope. Tibial slopes of 7° and 10° led rotational kinematics that were closest to that of a normal knee. Alterations in knee kinematics related to changing tibial slope may be related to a change in the PCL strain. However, these results must be confirmed by other tests involving more specimens.

ABSTRACT. One main perceived drawback for the adoption of computer assisted orthopedic surgery (CAOS) during total knee arthroplasty (TKA) relates to the increased surgical time compared to the use of standard mechanical instrumentation. This study compared the time efficiency between a next generation CAOS system (ExactechGPS®, Blue-Ortho, Grenoble, FR) and conventional mechanical instrumentation, and assessed the impact of surgeon experience level on the efficiency. Surgical time was retrospectively reviewed on 63 primary TKAs performed by a board-certified orthopedic surgeon (PP) using the Optetrak Logic® PS knee system (Exactech, Gainesville, FL), grouped as 1) Group I (control): 21 TKAs using conventional mechanical instruments; 2) Group II: 21 KAs performed using the CAOS system with an early experience level (first 21 cases); and 3) Group III: 21 TKAs using the CAOS system with an advanced experience level (beyond 30 cases). Patient condition (age, BMI, gender, etc.), surgical technique, and post-operative guidelines were similar across the three groups. No cases were lost and no patient had any intra-operative complications. Surgical time was compared across the three groups (with significance defined as p<0.05). Compared to the TKAs using conventional mechanical instrumentation, the average surgical time for the navigated TKAs performed with an early experience was 7 minutes longer. However, with an advanced experience level, the average surgical time was 2 minutes less than the time required using conventional mechanical instrument. Further, navigated TKAs with an advanced experience level exhibited the least variability among the three groups. None of the time differences were significant (p>0.20). No significant difference in TKA surgical time was found between the evaluated CAOS system (both within or pass the learning curve) and the conventional instrumentation. Nevertheless, once the learning curve was reached, the system decreased the time variability compared to conventional mechanical instrumentation. The comparable efficiency reported in this study to the conventional mechanical instrumentation may be attributed to the unique features of the ExactechGPS system, such as indication for use inside the sterile field, blood occlusion-resistant tracker design, customizable operative technique tailored to the surgeon’s preference, and compact and reduced number of instruments.

ABSTRACT. One of four normal people had mechanical alignment of 3 degrees varus and more that so-called “constitutional varus”. Parallel joint line to the floor found in both neutral and varus alignment. Therefore, joint line orientation may play an important role in clinical outcomes after TKA. For reconstituting joint line parallel to the floor advocated by 30 varus tibial cut that was introduced by Hungerford et al. The aims of this study attempt to compare between difference radiographic parameter in term of clinical outcomes. The prospective study conducted on 94 primary varus osteoarthritis knees undergone CAS TKA using either classical method (51 knees) or anatomical method (43 knees). Clinical outcomes including WOMAC scores, Oxford knee scores and ROM were evaluated preoperatively and 6 months postoperatively. Full leg standing hip-knee-ankle were measured mechanical axis, tibial cut angle and tibial joint line angle at 6 months after surgery. The results revealed that postoperative neutral alignment (mechanical axis 0± 3°), 4-5°varus and ≥6°varus showed no significant difference in term of WOMAC scores, Oxford scores and ROM. Including comparison between classical tibial cut and anatomical tibal cut, postoperative joint line parallel to the floor and oblique joint line had no significant in clinical outcomes. Nevertheless, anatomical tibial cut and joint line parallel to the floor had significant WOMAC scores improvement than the others. In conclusion, the joint line parallel to the floor may be one of key successes after TKA more than postoperative limb alignment.

ABSTRACT. Computer assisted surgery can improve accuracy and precision in many surgical techniques. Moreover, imaging based computer assisted procedures allow for planning and visual control during all steps. As an alternative to computed tomography (CT) and magnetic resonance imaging (MRI), ultrasound imaging (US) is investigated. It offers high resolution images in real-time while being non-invasive, cost-efficient and broadly available. However, due to a low signal to noise ratio (SNR), speckle, low contrast, acoustic shadowing and a small field of view, new methods must be developed to extract and collate the relevant information. We developed a new protocol and algorithm to segment partial bone surfaces from ultrasound volume images. The applicability of our approach is demonstrated on distal femur surface reconstructions. The bone surfaces are extracted in a multistage process employing level set methods. The basic idea behind this is that due to acoustic shadowing the area below a bone surface has low intensity values and thus, expanding a curve from inside yields the surface. The actual segmentation is done using a distance regularised level set evolution (DRLSE) algorithm incorporating an edge and a ridge indicator function, respectively. The latter is used in a refinement step which is based on the fact that US bone responses exhibit symmetric intensity profiles. The approach was validated on a solid foam bone phantom and a human knee. Ground truths were obtained from CT and MRI data, respectively. Bone surfaces could be reconstructed with an average distance error of 0.30 mm ± 0.37 mm in case of the femur phantom and between 0.25 mm ± 0.17 mm and 0.42 mm ± 0.53mm in case of the human knee.

ABSTRACT. CAOS systems have been developed since the mid 90’s for improving cup & stem placement as well as neck adjustment. The use of image-free navigation for adjusting modular-neck implants has been developed and studied by Merloz. Lecerf et al. have reported that navigation can help to optimize implant neck length in order to minimize discrepancies in leg length and offset. In this paper, we propose a novel concept that is intended to combine prediction of optimal neck, using image free navigation data, and instrumentation adapted to minimize reduction/dislocation manoeuvers for testing with trial implants. A novel software and instrumentation has been developed on an existing CAOS system (ExactechGPS®, Blue-Ortho, Grenoble, FR). The HipBall concept designed by Dr Stiehl offers a simple and fast method to optimize the neck length. Tow trackers are placed in the pelvis and femur. The initial conditions are measured and saved. Then, the cup and the stem are implanted using conventional techniques. One specific instrument equipped with a tracker is used to measure the cup center and the neck. Then the CAOS system proposes immediately the optimal neck length to minimize the variation of leg length and offset. The concept has been tested successfully on sawbone by an experienced surgeon. In conclusion, two methods have been proposed, the first one (Hip Ball) is optimizing the neck length, and the second one (Modulhips) is optimizing the neck length, version and inclination in order to suggest the best possible modular neck to be implanted, based on both theoretical measurements and real testing.

ABSTRACT. This study investigated the value of adapting a guidance concept called dynamic physical constraint (DPC) in the context of femoral surface resection for alignment of unicompartmental knee arthroplasty (UKA) prosthesis. DPC facilitates actively moving contact-based constraints and allows smooth motion parallel to the constraints. The concept uses simple mechanism to realize three-dimensional virtual fixture and thus is ideal for a hand-operated bone-mounted robot, which once registered, can operate standalone. The goal of this study was to evaluate whether our DPC robot enables the user to perform surface resection with accuracy similar to that of commercially available UKA robots.

One of the authors (JK) performed resection test using the preliminary robot prototype on 20 half-cylindrical medium-density fibreboard (MDF) samples. The robot prototype and the sample were mounted concentrically to a jig. We equipped the robot with a hand mill and a spherical burr and aligned the burr to the end effector. In this assessment, we deactivated DPC and performed uniform-radius resection. We mated dummy implants to resultant samples and laser-scanned the sets to quantify alignment accuracy.

The robot prototype can achieve alignment accuracy in similar order of commercially available UKA robots. On average, implants deviated from ideal angle by 0.8degree varus, 0.3degree lateral rotation, and 2.3degree flexional rotation. Implants were an average of 1.6mm above the target surface and had average of 0.1mm posterior and 0.1mm lateral shifts. The much higher flexional rotation and the overall upward implant offset are likely caused by fabrication imperfection in our preliminary prototype.

ABSTRACT. Introduction Planning for complex multi-planar osteotomies is difficult on a 2D computer screen. The surgeon has to depend on an accurate frame of reference to translate 2D into 3D, by means of anatomical landmarks. A process that’s potentially prone to inaccuracies. With the introduction of computer assisted surgery (CAS), robotic or patient specific guide based systems the accuracy and precision of the procedures has increased. To optimally plan the surgery we have combined 3D printing into the planning workflow.

Case One case was selected for 3D printing support and CAS planning. High resolution CT-scans were made, following our CAS navigation protocol. CT images were downloaded from the PACS system and imported into 3D-slicer (http://www.slicer.org). The tibia was segmented using threshold selected and then region based growing. Small defects were repaired by means of manual segmentation. The file was then exported as STL file and printed in PLA plastic. A patient tracker was attached to the 3d printed tibia, and registered using pair point and surface matching. Optimal resection planes were identified and the coordinates recorded on the CAS machine. The resections were performed and reconstructed. The print was then CT scanned. This scan was fused to the pre-operative CT scan. The surgical procedure was then performed, using the registered surgical plan. The scan of the print was used to check fragment orientation and height. Osteosynthesis was done using two screws. Registration during the practise session was highly accurate with a mean registration error of 0.5 mm.

Discussion 3D printing is already known as a supporting tool for evaluation of the anatomical situation. As the print can be freely manipulated, planning of resection planes or the marking of points of interest is very easy. Scanning the result allowed us to compare the result achieved during the practice session with the current result during the surgical procedure. With accurate reproduction and easy means of evaluation this technique allows for practice sessions. There is a need for a more bone like material for life like practice. Currently the material is very soft, especially when using the oscillating bone saw. With this, the technique can have use in the scientific evaluation of accuracy and precision of surgical tools and modalities and the development of new surgical techniques (RFA, tools). Ritacco et al and Abrahams et al have argued the need for measurement in the scientific comparison of procedures, toolsets and modalities. With a bone-like printable material or by using molds (3d printed negative), this technique could provide life-like, reproducible, specimens for experimental scientific studies. (PAPER WITHDRAWN)

ABSTRACT. Introduction: The proportion of population over the age of 80 undergoing spinal surgery is growing. The use of robotic computer assisted surgery in these patients is challenging, due to poor bone quality, and at times complex pathology and anatomy. We are present a consecutive series of octogenarian patients who underwent robotic computerized guided spine surgery. Methods: Prospective data in spine referral center was retrospectively collected and analyzed. All patients who were 80 years or older at the time of surgery were identified. These patients were matched to 120 patients under the age of 80 (a 3:1 ratio). Patients’ age, sex and indication for surgery were documented. Time of procedure, accuracy execution, fluoroscopy time, and any instrumentation related complication were documented. Results: Between 2007 and 2013, 192 trajectories were executed in the octogenarian patients and 568 trajectories in the younger patients. The average age was 83.9 years vs 61.2 (P<0.05); BMI was 26.9 in the octogenarians vs 28.9 in the younger patients ; 12 patients were males in the octogenarian vs 50 in the younger patients. The upper instrumented in both groups vertebra was T5 and the lowest was S1. Percutaneous procedures were performed in 25 octogenarian patients and 73 in the younger group of patients. Average robotic usage time was 6 min and 40 seconds in octogenarians, compared to 5 min and 5 seconds in younger patients (P<0.05). Radiation exposure time was 16.3 seconds per screw in the octogenarian's vs 9.3 seconds in the younger patients (P<0.05). 182 (94.5%) executed trajectories were accurate vs 545 (95.9%) accuracy in patients under the age of 80 (NS). No robotic assistance or any hardware related complications occurred. Discussion: Spine surgery in the octogenarians is challenging. The combination of osteoporotic bone and complex spinal pathologies, results in longer robotic procedures and in higher radiation exposure time, compared to younger patients. However, execution accuracy and safety is identical to younger group of patients, allowing to apply optimal instrumentation in these frail patients.

ABSTRACT. The medial patellofemoral ligament (MPFL) is one of the most important medial structures preventing lateral dislocation of the patella (LeGrand 2007). MPFL rupture leads to an altered patellar-tracking and retropatellar pressure which can result in early degeneration and even joint replacement. The goal of this study was to evaluate changes in retropatellar force after surgical reconstruction of the MPFL under simulation of physiological quadriceps muscle loading.

On 9 fresh-frozen cadaveric knees the quadriceps muscle was divided into 5 parts along their anatomic fiber orientation. Muscular loading was achieved by applying weights to each of the fife components in proportion to the cross sectional muscle area. A custom made sensor was introduced between the patella and femur [Pliance, Novel]. The robot-control-unit is linked to a force-torque sensor. The force free knee-flexion-path from 0° to 90° was performed using this hybrid robotic method. At first, the retropatellar pressure distribution was recorded with intact ligaments. After cutting the MPFL double bundle MPFL reconstruction (Schoettle 2009) was performed and the pressure distribution was obtained again. A CT-scan of the cadaver knees was performed to calculate the D surface of the patella. The surface was then divided in 4 quadrants and matched to the corresponding single measuring cells. We followed the hypothesis that MPFL reconstruction can restore native retropatellar forces.

RESULTS: For the total surface and intact MPFL mean force increased during knee flexion from 91.42 N at 5° flexion to a maximum of 159.72 N at 85° which equals an increase of 75%. Mean retropatellar force was 141.1 N for ligament stable knees which was sign. increased to 151.7 N after MPFL reconstruction. The physiological retropatellar force could not be restored through MPFL reconstruction. At the lateral patellar facet of native knees was a sign. higher force than the medial facet. Mean force was 95.2 N at the lateral and 43.7 N at the medial facet. Thus 70% of force was measured at the lateral und 30% at the medial facet. At the medial facet native mean retropatellar force was 43.7 N. Through ligament reconstruction mean force decreased to 41.51 N. Mean force at the lateral facet was 95.22 N for the MPFL intact knees and 99.41 N after ligament reconstruction. In conclusion reconstruction of the MPFL could not restore native conditions and even produced significantly higher retropatellar forces. Therefore our above stated hypothesis was rejected. A possible explanation for this is the applied surgical method which might tend to alter patellar movement and increase retropatellar pressure. Therefore close attention has to be paid to ligament tensioning and anatomical attachment.

ABSTRACT. Objective. To evaluate the learning curve of using the computer assisted navigation minimally invasive spine surgery (CAMISS) and transforaminal lumbar interbody fusion (TLIF) technique for the surgical treatment of lumbar degenerative diseases. Summary of Background Data. Minimally invasive (MI) transforaminal lumbar interbody fusion (TLIF) is technique challenge and long learning curve. We combined computer assisted navigation and minimally invasive spine surgery (CAMISS) to treat lumbar degenerative disease. However, the learning curve of the combine technique is unknown. Methods. Seventy-four consecutive patients with lumbar degenerative disease underwent CAMISS TLIF between March 2011 and January 2013 by a single surgeon. All patients were followed up at least 24 months, and the average follow-up time was 38 months. The learning curve was assessed via a logarithmic curve-fit regression analysis. According to the plateau of asymptote, initially 25 patients was divided as early group and remain cases were the later group. The clinical evaluation data included operative time, anesthesia time, intraoperative blood loss, day for ambulation, post-operative hospital stay, visual analog scale (VAS) leg and back pain scores, the Oswestry disability index (ODI) value, Macnab outcome scale scores, complications, radiological outcomes and rate of conversion to an open. Results. The complexity of cases increased over the series, and the complication rate decreased. There were significant different between two groups on average surgical time and duration of anesthesia (p<0.05). There was no significant difference with intraoperative blood loss, day for ambulation, and post-operative hospital stay in the two groups (p>0.05). The improvement of VAS leg and back pain scores and ODI in each group was similar: there was a significant improvement (p<0.05) at 3 months, 12 months and 24 months after surgery when compared with preoperative scores, but there was no statistical difference (p>0.05) in the VAS leg and back pain scores and ODI between two groups 3 months, 12 months and 24 months after surgery. The complication rate was 12.00% for early group, 6.82% for later group, with no statistical difference (p>0.05) between groups. No significant difference was found between two groups in Macnab outcome scale scores and solid fusion rate. There was no need for conversion to an open procedure for both groups. Conclusions. Our study showed a plateau asymptote of CAMISS TLIF was reached after 25 operations. The latter patients were during shorter operative time and anesthesia duration. CAMISS TLIF is an effective and safety option in the surgical treatment of lumbar degenerative disease.

ABSTRACT. Background Computer assisted navigation technology has been widely applied to spine surgeries. However, ideal accuracy could not be easily obtained and sustained. Many clinical factors were supposed to influence clinical accuracy. Our goal is to determine if and how clinical factors affect the accuracy of CT (computerised tomography) -based active infrared navigation system. Methods Lumbar Sawbone model, of the same size as human lumbar sequence, with dispersive pre-positioned titanium points was used. Clinical accuracy was accurately determined by acquiring primary image coordinates and navigation coordinates. We recorded the clinical accuracy of the navigation system in different clinical status, including with or without shadowless lamps, different distance between camera and model and different posture of surgery bed. Results Clinical accuracy may be influenced when the height or angle of the bed changed. In the visible area of the camera, the nearer the distance between the camera and the model, the better the clinical accuracy. At the distance of 1 meter, the best accuracy was recorded at the vertebra which was used for registration, which was 0.81±0.35 mm. Clinical accuracy of the CT-based active infrared navigation system may worsen due to the exposure to shadowless lamps. Conclusions Clinical accuracy of the CT-based active infrared navigation system may be influenced by different clinical status of shadowless lamps, different distance between camera and model and different posture of surgery bed.

ABSTRACT. INTRODUCTION Restoration of the posterior femoral offset has been described as a significant prognostic factor after TKR. The measurement of the posterior offset prior to the TKR on standard X-ray may be inaccurate because the thickness of the remaining cartilage cannot be assessed. The post-operative evaluation of offset restoration may also be inaccurate, as the rotational positioning of the femur may impact the 2D image of the implant. We hypothesized that the measurement of the restoration of the posterior femoral offset by a navigation system during the implantation is more precise than the conventional post-operative measurement on post-operative standard X-rays.

METHODS 100 cases of navigated TKR have been analyzed. The native offset was measured on pre-operative standard sagittal X-rays with a conventional technique. The thickness of the bone ± cartilage resection performed was measured intra-operatively by the navigation system. The thickness of the femoral implant at the level of the posterior femoral resection was recorded. The final offset was measured on post-operative standard sagittal X-rays. The actual offset change was compared to the measured offset change. Both measures were compared by appropriate tests at a 5% level of significance.

RESULTS The mean pre-operative measured offset was 61.4 ± 5.2 mm. The mean pre-operative actual offset was 57.5 ± 4.5 mm. The mean post-operative measured offset was 62.8 ± 6.6 mm. The mean post-operative actual offset was 56.5 ± 4.0 mm. The mean actual offset change was -0.9 ± 2.6 mm (range, -7 to +5 mm). The mean measured offset change was +1.4 ± 5.3 mm (range, -25 to +13 mm). The mean difference between the paired measurements of the pre-operative actual and measured offsets was 3.9 mm (p<0.001). There was a moderate correlation between the two paired pre-operative measurements (r² = 0.41, p<0.001). The mean difference between the paired measurements of the post-operative actual and measured offsets was 6.3 mm (p<0.001). There was a poor correlation between the two paired pre-operative measurements (r² = 0.27, p<0.001).

DISCUSSION We observed a significant difference between the actual offset and the measured offset. This may be explained by several biases: 1) rotational malposition of the femur; 2) uncertainty of the contours of both condyles; 3) uncertainty about the thickness of the remaining cartilage. According to the present study, the pre-operative X-ray measurement of the femoral offset cannot be used to plan the expected bone resection. Furthermore, the post-operative standard X-ray cannot be used as a quality control to assess the reconstruction of the femoral offset. We suggest using a navigation system as a more accurate tool for pre- and post-operative femoral offset.

ABSTRACT. The revision rate of shoulder arthroplasty remains high in relation to the hip and knee. The glenoid loosening is a major complication and can be driven by poor positioning of the glenoid component of a shoulder prosthesis. The purpose of this study was to determine the efficiency of a new surgical method for placing glenoid component using patient specific templates created using preoperative planning and 3D modeling.

ABSTRACT. The mobility and stability of the human knee joint are influenced mainly by the complex interaction between bony and soft tissue structures. Harmonised systems are crucial for physiological ligament function and overall knee kinematics and should be intended in, e.g. knee arthroplasty. The goal of this study was to analyse the relationship between distal femoral morphology and cruciate ligament attachments by (1) using segmented cruciate ligament attachments based on MRI-datasets (2) developing a robust framework for reliable and automatic identification of the articular geometry for an appropriate comparison. The cruciate ligament attachments were segmented in 6 MRI-datasets (n=6) from healthy male subjects (BH 184±6cm, BW 90±10kg) by a clinical expert. The insertion areas were exported as point clouds and the centres of gravitation served as approximations of the attachments. The femoral epicondyles were selected to determine the transepicondylar axis (TEA) as further reference. The articular geometry was analysed based on a framework implemented in MATLAB. The framework consisted of a pattern recognition algorithm for the identification of the articulating portion of the condyle profiles in the sagittal view. Ellipses were fitted to the contour parts and the 2D dispersion of the posterior foci of these ellipses was calculated. The elliptical axis (EA) was fitted to the 3D positions of posterior foci with minimum dispersion. Finally, the centroids of the cruciate ligament attachments were directly compared to the corresponding TEA and EA by projecting the ligament insertions to a plane, orthogonal to the appropriate reference axis. Looking at the point distributions of the insertions, ACL and PCL were clearly separated for both axes. Concerning the EA, the ACL insertions were located slightly proximal/anterior and for the PCL anterior. The mean distance between ACL and EA was 10.12±3.41 mm and 20.32±1.43 for the PCL, respectively. Despite the high deviations, particularly regarding the PCL, the variance was only 2.06 mm. Regarding the TEA, the distances were smaller but the variance was higher (8.42 mm). The results imply that there is at least a correlation between the EA and PCL attachment. However, the TEA showed a less functional relation. Therefore, the EA could provide additional knowledge for patient-specific therapy based on CT-data where no soft tissue information are available.

ABSTRACT. INTRODUCTION From a mechanical point of view, the wrist represents the most complex joint in the body, providing a large range of motion in the flexion-extension (FE) and radial-ulnar-deviation (RUD) planes. Especially scapho-lunate-ligament (SLL) deficiency remains an unsolved problem in hand surgery. The objective of this pilot study was, to show that the presented simulation model of the wrist joint could be used for dynamic computer based modelling of physiological and pathological motion behaviour of the wrist. MATERIALS AND METHODS The new in the AnyBody-Modelling-System developed wrist model includes detailed information of the wrist: 67 ligaments, five forearm muscles, each wrist bone is allowed to move in six degrees of freedom. Furthermore, a method for patient-specific model adaption was implemented, and a linear scaling approach for considering the cartilage layer of the bones. For three subjects we simulated the physiological RUD, the pathological motion behaviour (SLL-rupture), and of a virtual reconstructed SLL-rupture. The results were directly compared to their corresponding in vivo data. During RUD the relative motion of the two carpal rows are more complex. RESULTS The results visualizes the differences between physiological RUD in comparison to pathological RUD behaviour of the SLL-rupture. In case of SLL-rupture the gap between scaphoid and lunate is extended. Also the orientation of the scaphoid and lunate differs in the pathological case from the physiological. DISCUSSION The wrist model was able to reproduce valid physiologic ranges of motion with small differences to the in vivo data, and also pathological trends, when compared to experimental results. The model replicated the results of the in vivo studies in its entirety, with comparably good results for the physiological case. For the physiological and pathological case the acquired kinematic data of the selected carpal bones was in agreement with previously published data. The kinematical behaviour of the scaphoid and lunate for the RUD were within the same range as reported in the literature.

ABSTRACT. Patient Specific Instruments (PSIs) are becoming an increasingly common method to provide surgeons with assistance in accurately performing procedures; however, to our knowledge, these new instruments have only been applied to traditional, highly invasive surgical approaches. However, PSIs have the potential to decreased surgical invasiveness by reducing the surgeon’s need to clearly visualise anatomical landmarks. Therefore, we designed and evaluated a novel PSI for minimally invasive shoulder arthroplasty. The proposed minimally invasive approach prevents en face access to the articular surfaces and thus the PSI was designed to guide the accurate placement of a trans-humeral bone tunnel which would permit surgical steps to be conducted. To accurately create this tunnel and place a guide pin in the glenoid, the PSI was designed as a two sided guide that incorporates unique anatomical features from both bones, which would lock the two bones in a predefined pose relative to one another. Proper registration of the PSI is aided by the joint’s passive compression force, which is not disrupted due to the soft tissue sparing approach. Once the bones are locked together, a guide pin could be passed through the humeral head – creating a bone tunnel to guide later humeral bone preparation – and into the glenoid to guide reaming and drilling. By designing the guide in this way, it is possible to avoid the need to perform surgical steps with a clear en face view. The PSI was created by loading 3D reconstructed CT models of the humerus and scapula into a CAD package, aligning the desired humeral and scapular guide axes such that the bones’ relative pose is fully defined, and finally constructing the guide itself between and around the articular surfaces, such that sufficient anatomical features are incorporated to provide complete physical registration with the bones. This PSI was subsequently customized, based on a cadaveric specimen and fabricated using a 3D printer. The PSI’s usability and accuracy in achieving the pre-operative plan were then assessed using optical tracking and surface based registration procedure. Results of the evaluation demonstrated that the designed PSI is capable of accurately registering the two bones to within 5mm and 14° of the intended pre-operative plan, while also effectively reducing the invasiveness of the surgical procedure. Therefore, this novel PSI may represent a new avenue to improve the clinical impact of CAOS systems, by achieving good surgical accuracy, but with a greatly reduced invasiveness.

ABSTRACT. Surgical navigation requires an accurate, stable transformation between the tracking system and reference images. This study was the design and evaluation of an additively manufactured calibrator with an integrated verification tool, used to register cone-beam computed tomography (CBCT) image volume to electromagnetic (EM) tracking.

An Aurora EM system was used to track both the calibrator and a surgical probe. Intraoperative CBCT images were acquired with a GE Innova 4100 scanner. The calibrator incorporated 7 tantalum beads, a 6DOF EM sensor, and 7 through-holes for calibrator verification. The calibrator was characterized using the beads and averaged EM reading in 10 poses.

Target Registration Error (TRE) estimation used a device with 14 beads and 18 through-holes. For verification, the probe was placed in each path and the axis and tip location measured relative to the calibrator. This verification task took about 45s. Axial error was the angle between the probed paths and designed axes; translation error was the shortest distance between these lines.

The translation TRE was 3.14±0.96 mm and the angular TRE was 1.7±0.7 degrees, which is consistent with published EM evaluations. The validation axes had an inter-line distance of 0.9±0.78 mm and an axial difference of 1.1±0.7 degrees. The verification errors were smaller than TRE because of the different mathematical formulation. Although the verification calculation was not exactly a tracking error, it provided an alternative quantitative assessment of registration accuracy. This integrated intraoperative registration verification minimizes modifications to the surgical workflow and these results demonstrated highly accurate orientation tracking in a surgical environment.

ABSTRACT. Tracking of the anterior pelvic plane is of interest for medical interventions such as total hip arthroplasty, for which it is used as a reference for the positioning of the acetabular cup. We introduce and evaluate a new portable ultrasound device for the measure of the pelvic tilt in different positions of daily living. This device consists of two ultrasound probes articulated with respect to each other in order to visualize simultaneously the bony landmarks of interest that are one of the anterior superior iliac spine and the pubic symphysis. A series of sensors and the calibration of the ultrasound probes allow the measurement of the relative position of the landmarks with respect to a vertical line. The accuracy of the device has been investigated through a simulation study and showed errors (mean ± standard deviation [minimum; maximum]) as 0.18° ± 0.96° [-3.85°; 4.33°], with 99% of measurements within a ± 2.5° with respect to the actual pelvic tilt. This level of accuracy is similar to what can be found in the literature for the same purposes. Our device gathers advantages such as being portable and user friendly in order to be used during the pre-operative consultation. It is also non invasive and non irradiant. Further investigations will be run to assess this accuracy in vitro and in vivo.

ABSTRACT. C-Arm fluoroscopy is limited by its 2D imaging modality and is incapable of providing accurate 3D quantitative assessment of operative anatomy. In High Tibial Osteotomy (HTO), assessing the distance between the mechanical axis of the leg and the centre of the knee joint is difficult to accomplish due to limited fluoroscopic view size.A previously developed sensor-based tracking system (TC-Arm)adds on to C-arm equipment to provide additional quantitative capabilities. A new image-based tracking module was developed for TC-Arm using a reference panel with an array of fiducial markers. The image analysis software segments the marker positions in each image and identifies image coordinates with respect to the panel. Each image’s parameters are identified by 2D-3D matching of the panel’s 3D model to the marker’s epipolar geometries. Finally, the defined linear transformation matrices are applied for positioning all the fluoroscopic images with respect to the same global reference. A Sawbone model of the leg was used as a phantom and marked with radio-dense fiducial markers at the centres of each joint. An Optotrak optoelectronic tracking system data was used to validate the new module’s functions. First, tracking accuracy was determined by comparing orthogonal-stereo views and the reconstructed positions of the panel’s design. Secondly, TC-Arm’s results were compared to the corresponding digitized references points on the Sawbone model to calculate errors in the varus/valgus angle and mechanical axis deviation. The new addition to the TC-Arm has a reasonable tracking accuracy (<3.6mm, <4°) considering HTO: The system measured the mechanical axis deviation for HTO application with an accuracy of 1.3 mm and 1.4°. Comparing these results with the acceptable tolerance of less than 10 mm for MAD reported in the literature, our demonstrated results are considered to be within an acceptable range. With the new module, the capability for three-dimensional quantitative assessments of operative anatomies of any size can be added to any C-arm equipment in the OR. This can have great potential for many complex orthopaedic trauma, reconstruction, or preservation surgeries including HTO.

ABSTRACT. High tibial osteotomy (HTO) is a common surgical procedure for treatment of patients with varus malalignment. The success rate of the procedure is strongly dependent on the quality of the correction. Thus, an accurate pre-planning is essential to ensure that the precise amount of alignment is achieved postoperatively. The purpose of this study was to simulate the HTO in a patient with varus deformity in order to explore the interactions between the wedge angle, the mechanical axis, and the knee joint configuration. A finite element model of the knee joint of a patient with varus deformity was developed. The geometry was obtained using the whole limb CT scans the knee MR images. The bones were assumed as rigid bodies, the articular cartilage and the meniscus as elastic solids, and the ligaments as nonlinear springs. A 600N force was applied at the femoral head in the line of the mechanical axis and the resulting knee configuration was studied. The HTO was simulated assuming insertion of wedges with different angles beneath the tibial platue and applying the resulting alteration of the loading axis to the model. The results indicated that the actual change of the mechanical axes was always smaller than what predicted by a geometric pre-planning approach that does not consider the post-operative change of the knee joint configuration. It was suggested that subject-specific models are needed to simulate the HTO in patients before surgery and determine the appropriate wedge angle that locates the mechanical axis in the middle of the knee.

ABSTRACT. Opening-wedge High Tibial Osteotomy (HTO) has been shown to be an effective procedure to treat mild to moderate osteoarthritis of the medial compartment of the knee in active individuals. It has also become a mandatory surgical adjunct to articular cartilage restoration when there is preoperative malalignment. However, its efficacy is directly correlated with the accuracy of the correction, which must be within 3° of the preoperative target. Achieving this goal is a significant challenge with conventional techniques. Therefore, computer-assisted navigation protocols have been developed; however, they do not adequately address the technical difficulties associated with this procedure. We present an integrated solution dedicated to the opening-wedge HTO. Advantages to the technique we propose include: 1) a minimum number of implanted bone trackers, 2) depth control of the saw, 3) improved 3-D accuracy in the location of the lateral tibial hinge, and 4) micrometric adjustment of the degree of correction. The proof of concept has been completed on all six specimens. The following key points have been validated: a)Compatibility with a minimally-invasive (5-6 cm) surgical incision b)The compact navigation station can be placed close to the operative field and manipulated trough a sterile draping device c)Only two trackers are necessary to acquire the required landmarks and to provide 3-D control of the correction. These can be inserted within the surgical wound without any secondary incisions d)The optimized guide accurately controlled the external tibial hinge in all six cases e)The implant cavity could be milled effectively f)The distractor used to complete the desired realignment maintained stability of the distraction until final fixation with the PEEK implant g)The PEEK implant could be fixed to the tibia with excellent stability in a low-profile fashion. The solution presented here has the potential to help surgeons perform a medial opening-wedge HTO more safely and accurately. This will likely result in an increase in the number of HTOs performed for both isolated medial compartment osteoarthritis as well as for lower extremity realignment in association with cartilage restorative procedures.

ABSTRACT. We propose to employ a 'gold standard' imaging technique known as RSA (Roentgen Stereophotogrammetric Analysis) that uses two roughly orthogonal radiographs to measure implant micromotion in joint replacement patients. RSA can quantify motions between an implant and the patient’s bone to submillimeter accuracy, and it has been shown that relative movements above certain limits that emerge within two years post-surgery can reliably predict implant longevity.

The wide adoption of RSA use however, has been stifled by costs. To setup an RSA system typically requires expensive dual X-ray equipment and software, making RSA systems inaccessible to most researchers and clinicians.

In this study, we investigate the feasibility of using mobile C-Arms along with open-source RSA software as the foundation for an RSA system. C-Arms are portable X-ray machines that are already available at most hospitals. By leveraging these low-cost resources, our RSA protocol may aid researchers and clinicians in accelerating efforts aimed at achieving higher success rates in joint replacements.

Presented in this paper is the preliminary validation of our RSA system. Initial testing has yielded an accuracy of 0.033 mm and a precision range of 0.021 - 0.049 mm when measuring fixed distances. These results are comparable to those achieved by existing RSA systems in terms of resolving translations. Future work will include full validation on the accuracy and precision of the system, covering both translations and rotations. Once our marker-based RSA system has been validated, we aim to extend the system to enable markerless (model-based) RSA capabilities.

ABSTRACT. Joint fracture surgery imposes challenging accuracy requirements. Manipulating joint-related bone fragments to a high positional accuracy is a complex problem. Distal-femur-fractures extend into the knee joint and separate the articular surface of the bone into fragments damaging the cartilage surface of the bone, becoming difficult to treat. Robotic assistance can allow more accurate fracture fragment repositioning, combining the required high-positioning accuracy with the minimally invasive approach, reducing in-patient stays and making recovery swifter and more complete. We propose a robot-assisted system for the reduction of intra-articular joint fractures, which will allow the surgeon to reposition a bone fragment attached to the robot with submillimetric accuracy. The system consists of a 6-DOF parallel-robot composed by linear actuators arranged in a hexapod configuration. It is controlled in real-time from a computer workstation employing a host-target structure, i.e. a host-computer, a real-time controller, and a field-programmable-gate-array. Positioning trials were conducted aiming at characterising the system through quantitative measurements of its positioning accuracy. Each trial included moving the robot end-effector to a desired position/target. An optical tracker acquired the actual position of the robot end-effector at each target point by tracking the optical markers placed on the robot, allowing the measurement of positional accuracy as position errors (RMSE). The experiments demonstrated that the system is precise, achieving submillimetric translation errors (0.5mm along x-y axes, 50µm along z), and rotation errors of less than 0.3°. This level of accuracy is highly reliable for fracture surgery, leading to the conclusion that the system meets the requirements.

ABSTRACT. Ultrasound imaging has become commonplace in the biomedical arena and in Non-destructive Testing (NDT). However, despite their common ancestry, the two fields of research have bifurcated, with recent developments finding little crossover. A variant of the Synthetic Aperture Focussing Method (SAFT), known as the Total Focussing Method (TFM) and developed mainly within NDT, allows for the maximum extraction of information using a single phased array. Despite finding success in NDT, it has yet to find a place within biomedical imaging. The superior imaging capabilities associated with TFM could, however, find value in preoperative imaging for robotic knee arthroplasty. Currently, Computed Tomography (CT) is considered to be the gold standard orthopaedics, providing a three dimensional, patient specific representation of the surfaces of the knee. Ultrasonic imaging may offer an alternative to this, with lower cost, no danger of ionising radiation and the capability to provide accuracies comparable to CT. Research in biomedical ultrasound imaging has focussed on depicting soft tissues, meaning the techniques regularly employed in the field are of little use in this application. Therefore, Full Matrix Capture (FMC), TFM and SAFT have been employed in the imaging of the bony surfaces of the knee in an attempt to challenge CT as a method of preoperative imaging. In order to establish a proof of concept for this, an epoxy human distal femur and a bovine distal femur were scanned in a water bath. A 128 element 5MHz one dimensional phased array was employed for ultrasonic data capture. The array was positioned using a Kuka KR5 Arc HW 6-axis robot. The resulting images were combined with positional data to produce a surface point cloud, from which 3D surface mesh models were computed. Comparison with Faro Quantum laser scan ground truth data produced mean errors of 0.82±0.64 mm and 0.88±1.02 mm, for human and bovine distal femurs respectively. Having established a reasonable level of accuracy, future work will seek to incorporate the problems of soft tissue penetration and issues of line of sight within the joint.

ABSTRACT. Studies have shown abnormal knee kinematics after total knee arthroplasty (TKA) that can lead to suboptimal clinical outcomes. Tibial posterior slope (PTS) can significantly impact the kinematics of the reconstructed proximal tibia due to its direct impact on the tension of the posterior cruciate ligament (PCL). Cadaveric tests for assessing the influence of the posterior tibial slope on the kinematics require reproducible measurement methods. This study assessed the reproducibility of a cadaveric test method developed to evaluate the effects of PTS on knee kinematics. A computer-assisted (ExactechGPS®, Blue-Ortho, Grenoble, FR) PCL-retaining TKA (Optetrak CR, Exactech, Gainesville, FL) was performed on one healthy cadaveric knee with presumably intact PCL. Properly sized tibial and femoral components were implanted using bone cement, and assembled with a tibial insert with the thickness of the surgeon’s choice. The implanted tibial baseplate was specially designed to easily adjust the reconstructed tibial plateau to each PTS configuration without damaging the soft-tissue envelope. Knee kinematic was recorded by the guidance system and evaluated by performing a passive range of motion 3 times at each of the 5 PTS sequentially (10°, 7°, 4°, 1°, and back to 10°). The repeatability of the test was investigated by comparing the initial and the last data sets at 10° PTS, and statistical differences (defined as p<0.05) were assessed at ~0° (3°), 30°, 60°, 90° and 120° flexion. Similar knee kinematics was found between the two sets of acquisition at 10° PTS. Notably, the antero-posterior (AP) position of the femur relative to the tibia was identical between the three trials performed at the beginning of the experiment and the three trials performed at the very end of the experiment. No significant differences were found between the two data sets at the sampled flexion angles, except for a clinically negligible difference (<1mm) found in the femur-tibia AP displacement at 30° flexion (p=0.04). The results suggested that the presented test method does not disrupt the soft tissue environment of the knee, therefore provides reproducibly assessment of the knee kinematics with regard to the PTS. The adjustable tibial baseplate offers an improved solution in future investigations on the impact of PTS on knee kinematics, compared to previous proposed and less accurate method, such as adjusting the slope using anterior opening wedge osteotomy with gap filling using bone cement.

ABSTRACT. Computer-assisted orthopaedic surgery (CAOS) improves the accuracy in total knee arthroplasty (TKA). Although accuracy at the system-level has been assessed in a previous study, clinical interpolation/comparison of the results across CAOS systems is challenging. This study quantified and compared the system-level accuracy in the intraoperative measurements of resection alignment between two CAOS systems. Computer-assisted TKAs were performed on 10 leg assemblies (MITA knee insert and trainer leg, Medial Models, Bristol, UK) using System I (5 legs, ExactechGPS®, Blue-Ortho, Grenoble, FR) and System II (5 legs, globally established manufacturer). The surgeries referenced preoperatively defined anatomical landmarks. Post bone cut, the alignment parameters were collected by the CAOS systems (CAOS measured alignment). The pre- and post- operative leg surfaces were scanned, digitized, and registered (Comet L3D, Steinbichler, Plymouth, MI, USA; Geomagic, Lakewood, CO, USA; and Unigraphics NX, Siemens PLM Software, Plano, TX, USA). The alignment parameters were measured virtually, referencing the same pre-defined anatomical landmarks (baseline). The measurement errors between the baseline and CAOS measured alignment were compared between the two CAOS systems with significance defined as p<0.05. Compared to System I, System II had a higher unsigned error magnitude in the measurements of tibial varus/valgus alignment and posterior slope (p≤0.01), and lower magnitude in the measurement of femoral varus/valgus alignment (p=0.03). System II exhibited higher error bias towards tibial varus alignment (up to 2.59°), more posterior slope (up to 1.41°), and more femoral hyper extension (up to 1.6°) than System I (p<0.01). The mean signed and unsigned errors were generally less than 1°, with the exception of System II in the measurement of tibial varus/valgus alignment (signed and unsigned mean errors=1.93°). The results showed that System I generally had lower variability and less bias in the resection alignment measurements than System II. The majority of the significant differences found were clinically irrelevant (<1° in means). Notably, System II was shown to produce on average ~2° measurement errors in tibial varus/valgus alignment biased towards varus. Clinical-level accuracy in alignment outcomes has been shown to be system-dependent, this study further suggested there are differences in system-level accuracy between CAOS systems.

ABSTRACT. A typical total knee arthroplasty (TKA) procedure using a computer-assisted orthopaedic surgery (CAOS) system involves inputting target resection parameters at the beginning of the surgery and measuring the achieved resection after bone cuts. This study assessed system-dependent differences between two contemporary CAOS systems by comparing the alignment deviation between the measurement of the achieved resection and the targeted parameters. Computer-assisted TKA resections were performed on ten leg assemblies (MITA knee insert and trainer leg, Medial Models, Bristol, UK) by a board-certified orthopaedic surgeon (BH) using System I (5 legs, ExactechGPS®, Blue-Ortho, Grenoble, FR) and System II (5 legs, globally established manufacturer). The surgeon was experienced (>30 surgeries) with both systems. The deviations of the intraoperative alignment measurements on the achieved distal femoral and proximal tibial resection from the target were calculated and compared between the two systems with significance defined as p<0.05. System II had significantly higher deviation towards varus (on average 2.2°) than System I (on average 0.83° valgus) for the tibia (p<0.01). System I tended to measure slightly more in flexion (~1°) than System II (~0.5° extension) (p=0.03). System I demonstrated lower variability of the signed deviation than System II in tibial varus/valgus alignment, femoral flexion/extension, and femoral varus/valgus alignment. No significant differences were found in between systems in the unsigned errors. This study demonstrated that there exists system-dependent variability associated with intraoperative measurement of the achieved resection during TKA. Assuming a consistent surgical variability exhibited by the same surgeon with equal experience on both systems, the data showed that both systems had measurement within the perceived acceptable range (within 3°). However, some systems (System II) may have higher variability than others (System I), and exhibit clinically meaningful bias (tibial varus/valgus) while achieving or quantifying the resections. The variability may be caused by the cumulated effect of the differences between the two CAOS systems. As clinical alignment accuracy has been found to be system-dependent, and data in the surgical report has been used as key inputs in relevant studies, the results here emphasizes the importance of taking into account the specific CAOS system in both clinical application and CAOS research.

ABSTRACT. During computer-assisted total knee arthroplasty (TKA), depending on surgeon preferences or patient factors, resection planning (guided adjustment of cutting blocks) is performed with different knee flexion, abduction/adduction (ABD/ADD) and internal/external (I/E) rotation angles, potentially leading to measurement errors in the planned resections due to a modified tracker/localizer spatial relationship. This study developed a method to assess the variation in the measurement of the planned resection due to leg manipulation during TKA. Computer-assisted TKA (ExactechGPS®, Blue-Ortho, Grenoble, FR) was performed on a whole leg assembly (MITA knee insert and trainer leg, Medial Models, Bristol, UK) by a board-certified orthopaedic surgeon (BH) at his preferred leg flexion, ABD/ADD, and I/E rotation angles. A cutting block was adjusted and fixed to the tibia, targeting the resection parameters (neutral tibial and femoral alignments with 3° tibial posterior slope). An instrumented resection checker was then attached to the cutting block to measure the planned resection at the same leg position (baseline). Next, the surgeon moved the leg to 9 sampled positions, representing typical leg manipulation during TKA. The planned resection was tracked by the CAOS system at each leg position. Tibial resection parameters at each sampled position were compared to the baseline. Regression was performed to identify the variables (flexion, ABD/ADD, I/E rotation) that significantly contribute to the measured variation (p<0.05). Clinically negligible variations were found across the 9 positions, with mean errors ≤0.1mm in resection depths and ≤0.2° in alignment. For the system studied leg flexion strongly correlated with the measurement errors in resection depths (p≤0.01) and posterior slope (p<0.01), with a trend of measuring less in resection depths and posterior slope with an increased leg flexion. However, the magnitude of the errors was clinically negligible (between ±0.5 mm or ° at a confidence level of 95%). The measurement accuracy of planned resection due to change of leg position has been overlooked in the evaluation of CAOS systems, even though it directly impacts the final resection. The results showed that ExactechGPS can provide robust measurements of the planned resection parameters during TKA, independent of the ABD/ADD and I/E rotation of the knee. The method can be applied to various CAOS systems to improve the understanding of system-specific clinical variability on the measurement of planned TKA resection.

ABSTRACT. Unicondylar Knee Arthroplasty (UKA) is an alternative surgical option to a Total Knee Arthroplasty (TKA) for knee arthritis suffers with early to mid-stage disease. It has evolved over the last four decades as new instruments and implants have been developed. However, the utilisation of this procedure within the UK has been limited to around 8% of all knee arthroplasty procedures. Robotic assisted UKA has grown in popularity, where 14% of UKA procedures use robotically assisted devices in the USA. The surgical plan for the prosthesis placement is determined using data recorded during an intraoperative stress test to access the ligaments of the knee. Before any cuts are made the position of the femoral and tibial components can be altered to find the optimised component position. The bony surface is then removed using a high speed rotating spherical bur which is robotically controlled to cut the exact plan. The outcome measure for this study was the change in the tibiofemoral angle which was defined as the difference between the planned and achieved angle in the coronal plane reported by the surgical system. 298 patients receiving image free, robotically assisted UKA by 22 surgeons in 16 hospitals across Europe and the USA were accessed. The surgical system calculated the pre-operative tibiofemoral coronal angle, as well as the predicted angle if the surgical plan for the implant placement was achieved. After the cutting phase of the procedure, the tibiofemoral coronal angle was recorded when the implants had been cemented in their final position. The achieved and planned tibiofemoral coronal angles were compared where when the lower limb was in full extension. The achieved tibiofemoral angle was within 1o, 2o and 3o of the surgical plan in 88.9%, 96.6% and 99.3% of the cases respectively. The differences between the tibiofemoral angles planned and achieved are shown in figure 1. The mean difference in the planned and achieved was -0.1o (SD 1.03o). In conclusion, image free robotics provides a new tool for accurately preparing bone in UKA. This study reported excellent levels of accuracy between the achieved and planned tibiofemoral angle.

ABSTRACT. In computer assisted orthopaedic surgery (CAOS), ultrasound (US) bone imaging is emerging as a viable intra-operative imaging modality. US imaging offers numerous advantages over the currently dominant radiation-based systems - specifically no use of ionizing radiation, low cost, portability and real-time response. Although recent state-of-the-art US bone imaging methods demonstrated sub-millimetric accuracy, they still are inhibited by comparatively long runtimes (~50 seconds for an US volume). We therefore assessed whether a downsampling scheme could improve runtime without significantly affecting bone surface localization accuracy. This also allows us to evaluate whether lower frequency US scans which can image deeper structures, albeit with decreased spatial resolution, could potentially provide acceptable accuracy. Our proposed method first downsamples an US volume with an appropriate anti-aliasing filter, and converts thick ridge-like bone boundaries in an US volume into thin ridge-like bone boundaries. Next, we use a bone-ridge detector using appropriately reduced scales of wavelets and appropriately reduced sizes of windows, to detect the bone boundaries. We validated the algorithm on 15 in-vivo clinical pelvic data and found a tenfold reduction in computation time with only modest increases in the surface fitting error (from ~0.45mm to 0.57mm). These results suggest that the proposed method may be useful either in reducing processing time or enabling use of low beam frequencies for deeper tissue penetration. In future, we will do a more extensive accuracy assessment and will reassess the algorithm using a version of the code that we have implemented on a graphical processing unit (GPU) to further reduce computational time.

ABSTRACT. Introduction Unicompartmental Knee Arthroplasty (UKA) has been offered as a tissue sparing alternative to total knee arthroplasty (TKA). Proposed benefits of UKA include more normal kinematics and function. In this study, a novel measurement technique was used to determine proposed benefits between UKA and TKA. Methods UKA or TKA patients were prospectively enrolled in this IRB approved study. Each patient received preoperative education regarding their expected physical therapy (PT) regimen, which was uniform for all patients. 5 functional criteria were tracked: range of motion from 5 to 115 degrees, recovery of flexion and extension strength to 4/5 of preoperative strength, gait with minimal limp and without an assistive device for 250 feet and ability to ascend and descend a flight of stairs with step over gait and good control. The number of PT visits to reach each functional criteria was recorded. Patient reported outcomes measurements (PROM) scores collected pre-operatively, 2 and 6 weeks post-operatively including: Knee Society Score (KSS), Knee Injury and Osteoarthritis Outcome Score (KOOS), Western Ontario and McMaster Osteoarthritis Index (WOMAC), Forgotten Joint Score (FJS) Results 26 patients (27 knees) were enrolled in the study, with 11 females and 15 males. 18 knees received a TKA and 9 received a UKA. Average age was 64.6 ±7.4 for the TKA group and 64.7 ±4.1 for the UKA group. Average body mass index (BMI) was 28.4±5.9 for the TKA group and 30.4±4.3. Age and BMI did not differ between the two groups. The UKA group required less PT visits on average to reach each criteria, a significant difference (p<0.05) was seen in extension of 5 degrees (5 vs 2.4, p=0.0411), flexion of 115 degrees (5 vs 3.8, p=0.0004), gait with minimal limp (6.8 vs 3.8, p=0.0022), stair ascend/descend (10.5 vs 7.1, p=0.0277) and 4/5 extension strength (9.8 vs 6.4, p=0.0319). No Statistical difference in KSS, KOOS, WOMAC, FJS at any follow-up (pre-op, 2 and 6 weeks post-op) Discussion Early results show less PT is required for UKA patients than TKA patients to reach the same functional goals. As PT accounts for a significant portion of the episode of care for knee arthroplasty, this may result in a decreased economic burden for UKA patients. Differences in outcomes of PT testing and Patient Reported Outcomes Measures (PROMs) scores indicates current PROM scores are not sensitive enough to detect changes in early recovery. More patients and longer follow-up are required to determine full clinical and economic benefit of UKA procedure. This study is ongoing and will follow patient out to 1 year post-op.

ABSTRACT. Nowadays, planning of image guided surgeries is often done using computed tomography (CT) and magnetic resonance imaging (MRI). CT offers good image quality but due to its ionizing radiation it is invasive. MRI is affected by distortions and quite expensive. Ultrasound might serve as an alternative imaging modality. It offers high resolution images in real-time while being non-invasive, cost-effective and broadly available. Its weaknesses, however, are a low signal to noise ratio (SNR), speckle, low contrast, acoustic shadowing and a small field of view (FOV). Thus new methods must be developed to extract and collate the relevant information. We developed an image processing chain to reconstruct entire bone surfaces from 3D ultrasound volume data where surfaces can be greater than the FOV. The applicability of our approach is demonstrated on distal femur surface reconstructions. The bone surfaces are reconstructed in a multistage process. The basic idea behind this is that due to acoustic shadowing the area below a bone surface has low intensity values and thus, expanding a curve from inside yields the surface. The actual segmentation is done incorporating distance regularised level set evolutions (DRLSE). Doing so, the refinement is based on the fact that US bone responses exhibit symmetric intensity profiles. Surface patches obtained from the segmentation procedure are roughly aligned manually or using tracking information and then registered via the iterative closest point (ICP) algorithm. Due to acoustic shadowing, the resulting surface is incomplete. The entire surface geometry is reconstructed by fitting an average geometry or a statistical shape model (SSM) computed from 332 training data sets. The approach was validated on a solid foam bone phantom; the ground truth was obtained from CT data. 19 US volume image were acquired. Compared to the ground truth, the segmented bone surface patches could be reconstructed with an average surface registration error (SRE) of 0.30 mm ± 0.37 mm. After registration and model fitting, the SRE of the morphing approach was 1.07 mm (absolute distances) and the average signed distance error was 0.77 mm ± 1.07 mm. The SRE of the SSM approach was 0.74 mm and the average signed distance error was -0.20 mm ± 0.95 mm.

ABSTRACT. Key goals of knee arthroplasty include the correction of malalignment and achieving a 'balanced' knee with stability through flexion with applied varus and valgus force. A reliable, non-invasive measure of mechanical alignment with applied force could be a valuable tool in planning and follow-up of TKA. The Physiopilot non-invasive system has been previously validated as reliable for measurement of supine mechanical femorotibial angle (MFTA) in the range of flexion 0°-40°. We aimed to validate the system in measuring MFTA with applied force through the range of flexion in vivo. 23 volunteers with normal knees were recruited, mean age of 33 years. Volunteers were examined by two clinicians, with each conducting two registrations on each knee. The range of laxity in the coronal plane (MFTA) was measured at 15°, 30°, 45°, 60°, 75° and 90° with each examiner manually applying a force to the knee to the endpoint of laxity for both varus and valgus displacement. Intra and inter-observer repeatabilities were evaluated with calculation of Coefficients of Repeatability (CR) and Inter-class Correlation Coefficients (ICC), respectively. Initial MFTA registration agreement for the first eight volunteers was found to be outlying acceptable limits and therefore the final 15 volunteers were analysed, giving 30 normal knees. The system showed adequate intra-observer repeatability in extension with mean CR 2° with varus/valgus force and inter-observer ICC 0.7, however mean CR and ICC were >3° and <0.7 as flexion was introduced. In keeping with previous studies, the non-invasive system was able to give a reliable measurement of MFTA in extension. However, assessment of MFTA with flexion was outwith acceptable limits. We propose that further studies are required to address the errors with this non-invasive system in measuring MFTA with flexion. We believe that the outlying results for our first eight volunteers represent a learning curve in using the non-invasive system.

ABSTRACT. Recently we developed a personalized X-ray reconstruction-based planning and post-operative treatment evaluation system called iLeg for total knee arthroplasty or lower extremity osteotomy. Based on a patented X-ray image calibration cage and a unique 2D-3D reconstruction technique, iLeg can generate accurate patient-specific 3D models of a complete lower extremity from two standing X-rays for true 3D planning and evaluation of surgical interventions at the knee joint. The goal of this study is to validate the accuracy of this newly developed system using digitally reconstructed radiographs (DRRs) generated from CT data of 12 cadavers (24 legs). Our experimental results demonstrated an overall reconstruction accuracy of 1.3±0.2mm.

ABSTRACT. For a proper functional restoration of the knee following knee arthroplasty, a comprehensive understanding of bony and soft tissue structures and their effects on biomechanics of the individual patient is essential. A systematic description of morphological knee joint parameters and a study of their effects could beneficial for an optimal patient-specific implant design. The goal of this study was the development of a full parametric model for a comprehensive analysis of the distal femoral morphology also enabling a systematic parameter variation in the context of a patient specific multiparameter optimization of the knee implant shape. The computational framework was implemented in MATLAB and tested on 20 CT-models which originated from pathological right knees. The femora were segmented semi-automatically and exported in STL-format. First, a 3D surface model was imported, visualized and reference landmarks were defined. Cutting planes were rotated around the transepicondylar axis and ellipses were fitted in the cutting contour using pattern recognition. The portions between the ellipses were approximated by using a piecewise cubic hermite interpolation polynom such that a closed contour was obtained following the characteristics of the real bone model. At this point the user could change the parameters of the ellipses in order to manipulate the approximated contour for e.g. higher-level biomechanical analyses. A 3D surface was generated by using the lofting technique. Finally, the parameter model was exported in STL-format and compared against the original 3D surface model to evaluate the accuracy of the framework The presented framework could be successfully applied for automatic parameterization of all 20 distal femur surface data-sets. The mean global accuracy was 0.09±0.62 mm with optimal program settings which is more accurate than the optimal resolution of the CT based data acquisition. A systematic variation of the femoral morphology could be proofed based on several examples such as the manipulation of the medial/lateral curvature in the frontal plane, contact width of the condyles, J-Curve and trochlear groove orientation. In our opinion, this novel approach might offer the opportunity to study the effect of femoral morphology on knee biomechanics in combination with validated biomechanical simulation models or experimental setups. New insights could directly be used for patient-specific implant design and optimization.

ABSTRACT. The consideration of the individual knee ligament attachments is crucial for the application of patient specific musculoskeletal models in the clinical routine, e.g. in knee arthroplasty. Commonly, the pre-operative planning is based on CT images, where no soft tissue information are available. The goal of this study was to evaluate the accuracy of a full automatic and robust mesh morphing method that estimates locations of cruciate ligament attachments on the basis of training data. The cruciate ligament attachments from 6 (n=6) different healthy male subjects (BH 184±6cm, BW 90±10kg) were identified in MRI-datasets by a clinical expert. The insertion areas were exported as point clouds and the centres of gravitation served as approximations of the attachments. These insertion points were used to annotate mean shapes of femur and tibia. The mean shapes were built up from 332 training data sets each. The surface data were obtained from CT scans by performing an automatic segmentation followed by manual cleaning steps. The mean shapes were computed by selecting a data set randomly and aligning this reference rigidly to each of the remaining data sets. The data were fitted using the non-rigid ICP variant (N-ICP-A). Due to this morphing step, point correspondences were established. By morphing a mean shape to the target geometries, including the cruciate ligament attachments, the distribution of the insertions on the original mean shape was obtained. Subsequently, a statistical mean was computed (annotated mean). The annotated mean shape was again morphed to the target data sets and the deviations of the respective predicted insertion points from the measured insertion points were computed. The training data was successfully morphed to all 6 subjects in an automatic manner with virtually no distance error (10-5 mm). The mean distance between the measured and morphed ligament attachments was highest for the ACL in the femur (4.26±1.48 mm) and lowest for PCL in the tibia (1.63±0.36 mm). The highest deviation was observed for femoral ACL (6.93 mm). In this study, a morphing based approach was presented to predict origins and insertions of the knee ligaments on the basis of CT-data, exemplarily shown for the cruciate ligaments. It has been demonstrated, that the N-ICP-A is applicable to predict the attachments automatic and robust with a high accuracy. This might help to improve patient-specific biomechanical models and their integration in the clinical routine.

ABSTRACT. Utilisation of unicondylar knee arthroplasty (UKA) has been limited due in part to high revision rates. Only 8% of knee arthroplasty surgeries completed in the England and Wales are UKAs. It is reported that the revision rate at 9 years for Total Knee Arthroplasty (TKA) was 3% compared to 12% for UKAs. In the last decade semi active robots have been developed to be used for UKA procedures. These systems allow the surgeon to plan the size and orientation of the tibial and femoral component to match the patient’s specific anatomy and to optimise the balancing the soft tissue of the joint. The robotic assistive devices allow the surgeon to execute their plan accurately removing only ‘planned’ bone from the predefined area. This study investigates the accuracy of an imageless navigation system with robotic control for UKA, reporting the errors between the ‘planned’ limb and component alignment with the post-operative limb and component alignment using weight bearing long leg radiographs. We prospectively collected radiographic data on 92 patients who received medial UKA using an imageless robotic assisted device across 4 centers (4 surgeons). This system is CT free, so relies on accurate registration of intraoperative knee kinematic and anatomic landmarks to determine the mechanical and rotational axis systems of the lower limb. The surface of the condylar is based on a virtual model of the knee created intra operatively by ‘painting’ the surface with the tip of a tracked, calibrated probe. The buring mechanism is robotically controlled to prepare the bone surface and remove the predefined volume of bone. The study shows the 89% of the patients’ post-operative alignment recorded by the system was within 3o of the planned coronal mechanical axis alignment. The RMS error was 1.98o.The RMS errors between the robotic system’s implant plan and the post-operative radiographic implant position was; femoral coronal alignment (FCA) 2.6o, tibial coronal alignment (TCA) 2.9o and tibial slope (TS) 2.9o. In conclusion, the imageless robotic surgical system for UKA accurately prepared the bone surface of the tibia and femur which resulted in low errors when comparing planned and achieved component placement. This resulted in a high level of accuracy in the planned coronal mechanical axis alignment compared to that measured on post-operative radiographs.

ABSTRACT. Despite being demonstrably better than conventional surgical techniques with regards to implant alignment and outlier reduction, computer navigation systems have not faced widespread adoption in surgical operating rooms. We believe that one of the reasons for the low uptake stems from the bulky design of the optical tracker assemblies. These trackers must be rigidly fixed to a patient’s bone and they occupy a significant portion of the surgical workspace, which makes them difficult to use. In this study we introduce the design for a new optical tracker system, and subsequently we evaluate the tracker’s performance. The novel tracker consists of a set of low-profile flexible pins that can be placed into a rigid body and individually deflect without greatly affecting the pose estimation. By relying on a pin’s stiff axial direction while neglecting lateral deviations, we can gain sufficient constraint over the underlying body. We used an unscented Kalman filter based algorithm as a recursive body pose estimator that can account for relative marker displacements. We assessed our tracker’s performance through a series of simulations and experiments inspired by a total knee arthroplasty. We found that the flexible tracker performs comparably to conventional trackers with regards to accuracy and precision, with tracking errors under 0.3mm for typical operating conditions. The tracking error remained below 0.5mm during pin deflections of up to 40mm. Our algorithm ran at computation speeds greater than real-time at 30Hz which means that it would be suitable for use in real-time applications. We conclude that this flexible pin concept provides sufficient accuracy to be used as a replacement for rigid trackers in applications where its lower profile, its reduced invasiveness and its robustness to deflection are desirable characteristics.

ABSTRACT. Background. Accurate insertion of pedicle screws in scoliosis patients is a great challenge for surgeon due to the severe deformity of thoracic and lumbar spine. Meanwhile, malposition of pedicle screw in scoliosis patients could lead to severe complications. Computer-assisted navigation technique may help improving the accuracy of screw placement and reducing complications. Thus, this meta-analysis of the published researches was conducted concentrating on accuracy of pedicle screw placement and postoperative assessment in scoliosis patients using computer-assisted navigation technique.

Methods. PubMed, Cochrane and Web of Science databases search was executed. In vivo comparative studies that assessed accuracy and postoperative evaluation of pedicle screw placement in scoliosis patients with or without navigation techniques were involved and analyzed.

Results. One published randomized controlled trial (RCT) and seven retrospective comparative studies met the inclusion criteria. These studies included 321 patients with 3821 pedicle screws inserted. Accuracy of pedicle screw insertion was significantly increased with using of navigation system, while average surgery time was not significantly different with non-navigated surgery. And Correction rate for scoliosis in navigated surgery was not significantly different with non-navigated surgery.

Conclusions. Navigation technique does indeed improve the accuracy of pedicle screw placement in scoliosis surgery, without prolong the surgery time or decrease the deformity correction effect.

ABSTRACT. Introduction: Optimal orthopaedic implant placement is a major contributing factor to the long term success of all common joint arthroplasty procedures. Devices such as 3D printed bespoke guides and orthopaedic robots are extensively described in the literature and have been shown to enhance prosthesis placement accuracy. These technologies have significant drawbacks such as logistical and temporal inefficiency, high cost, cumbersome nature and difficult theatre integration. A radically new disruptive technology for the rapid intraoperative production of patient specific instrumentation that obviates all disadvantages of current technologies is presented.

Method: An ex-vivo validation and accuracy study was carried out using the example of placing the glenoid component in a shoulder arthroplasty procedure. The technology comprises a re-usable table side machine, bespoke software and a disposable element comprising a region of standard geometry and a body of mouldable material. Anatomical data from 10 human scapulae CT scans was collected and in each case the optimal glenoid guidewire position was digitally planned and recorded. The glenoids were isolated and concurrently 3D printed. In our control group, guide wires were manually inserted into 1 of each pair of unique glenoid models according to a surgeon’s interpretation of the optimal position from the anatomy. The same surgeon used the guidance system and associated method to insert a guide wire into the second glenoid model of the pair. Achieved accuracy compared to the pre-operative bespoke plan was measured in all glenoids in both the conventional group and the guided group.

Results: The technology was successfully able to intraoperatively produce sterile, patient specific guides according to a pre-operative plan in 5 minutes including device set up and planning, at a minimal cost. In the manual insertion group, average accuracy achieved was 6.8º and 1.58mm with respect to the plan compared to the guided group where an average of 0.74mm and 1.72 º was achieved.

ABSTRACT. Although total knee arthroplasty (TKA) is a largely successful procedure to treat end-stage knee osteoarthritis (OA), some studies have shown postoperative abnormal knee kinematics. Computer assisted orthopaedic surgery (CAOS) technology has been used to understand preoperative knee kinematics with an open joint (arthrotomy). However, limited information is available on the impact of arthrotomy on the knee kinematics. This study compared knee kinematics before and after arthrotomy to the native knee using a CAOS system.

Kinematics of a healthy knee from a fresh frozen cadaver with presumably intact PCL were evaluated using a custom software application in an image-free CAOS system (ExactechGPS, Blue-Ortho, Grenoble, FR). At the beginning of the test, four metal hooks were inserted into the knee away from the joint line (one on each side of the proximal tibia and the distal femur) for the application of 50N compressive load to simulate natural knee joint. Prior to incision, one tracker was attached to each tibia and femur on the diaphysis. Intact knee kinematics were recorded using the CAOS system by performing passive range of motion 3 times. Next, a computer-assisted TKA procedure was initiated with acquisition of the anatomical landmarks. The system calculated the previously recorded kinematics within the coordinate system defined by the landmarks. The test was then repeated with closed arthrotomy, and again with open arthrotomy with patella maintained in the trochlea groove. The average femorotibial AP displacement and rotation, and HKA angle before and after knee arthrotomy were compared over the range of knee flexion. Statistical analysis (ANOVA) was performed on the data at ~0° (5°), 30°, 60°, 90° and 120° flexion.

The intact knee kinematics were found to be similar to the kinematics with closed and open arthrotomy. Differences between the three situations were found, in average, as less than 0.25° (±0.2) in HKA, 0.7mm (±-0.4) in femorotibial AP displacement and 2.3° (±1.4) in femorotibial rotation. Although some statistically significant differences were found, especially in the rotation of the tibia for low and high knee flexion angles, the majority is less than 1°/mm, and therefore clinically irrelevant.

This study suggested that open and closed arthrotomy do not significantly alter the kinematics compared to the native intact knee (low RMS). Maintaining the patella in the trochlea groove with an open athrotomy allows accurate assessment of the intact knee kinematics.

ABSTRACT. For a successful total knee arthroplasty (TKA) and long prosthesis lifespan, correct alignment of the implant components as well as proper soft tissue balancing are of major importance. In order to overcome weaknesses of existing imaging modalities for TKA planning such as radiation exposure and lack of soft tissue visualization (X-ray and CT) and high cost, long acquisition times and geometric distortion (MRI), it is investigated if ultrasound (US) imaging is a suitable alternative. Currently, a reconstruction method of the bony knee morphology based on US imaging is developed at our research institute. For capturing the mechanical axis, being crucial for TKA planning, different approaches could be implemented. This work investigates whether a weight-bearing full leg X-ray registered with the local 3D-US knee dataset can be used for this purpose. Also, the impact of incorrect calibration data (i.e. uncalibrated X-rays) on the accuracy of the estimated mechanical axis is investigated. A 3D-2D projective, feature-based registration algorithm was used to spatially align the 3D US-based model to the 2D X-ray image before transferring the mechanical axis from the X-ray to the model. For validation, a CT-based local model and its projection were used and an initial error in translation and rotation was added. Also, calibration parameters such as the centre ray position and the source-to-image-detector distance were altered. The estimation error of the mechanical axis was less than 1°, the median error lower than 0.1° in the frontal plane. Even if the calibration data is not available, the accuracy remains sufficient for TKA planning. In this study, idealized 2D and 3D image information was used. In the future, this method should be tested using clinical X-ray images and 3D-US data.

ABSTRACT. Important factors affecting quality of life (QOL) after total knee arthroplasty (TKA) include postoperative knee kinematics and geometry, influenced by implant design and placement (Matsuda 2001; Nishikawa 2013; Noble 2005). Although specific design factors and their effect on kinematics or QOL have been investigated previously, the inter-relationships between preop-postop changes in kinematics, geometry and the resulting QOL have not been studied to our knowledge. These are essential to understand the interplay between the different factors, and to determine which factors manufacturers and surgeons should focus on when designing and implanting knee prostheses. In addition, the majority of TKA studies focus on the tibiofemoral (TF) joint, although the patellofemoral (PF) joint is routinely the source of postop complications; the PF joint is difficult to study due to polyethylene radiotransparency and because the femoral component obscures the patella from most directions. The purpose of this pilot study was to correlate changes in knee articular shape, over which the implant designers and surgeons have some control, to changes in kinematics and postop QOL, with a particular focus on the PF joint, to answer the following research questions for a sample population with a given implant design and surgeon: (1) Do changes in knee shape affect knee kinematics? In particular, is patellar tracking affected by groove location? (2) Do changes in knee kinematics affect QOL? (3) Do changes in knee shape (resulting from implant design and placement) affect QOL? (4) Do individuals with worse QOL differ from those with better QOL?

ABSTRACT. Postoperative radiological assessment is used to evaluate the success of knee replacement procedures. Load-bearing long-standing anterior-posterior (AP) x-rays are typically used for this assessment. For knee replacement procedures the five landmarks that are identified are: 1) hip center; 2) femoral knee center; 3) tibial knee center; 4) medial malleolus; and 5) lateral malleolus. These landmarks are used to identify the femoral and tibial reference mechanical axes. However, variations in the x-ray acquisition process and foot rotation can lead to errors in this assessment. In the past, researchers have studied the effect of foot rotation and flexion on estimation of knee alignment. In our study, the use of digitally reconstructed radiographs (DRRs) allows us to vary the x-ray acquisition parameters and observe the effect of these changes to estimations of the mechanical axes. We also measured the inter-user variability in these measurements. Our results show that AP x-rays can be used to accurately estimate the femoral and tibial mechanical axes.

ABSTRACT. INTRODUCTION An appropriate positioning of a total knee replacement (TKR) is a prerequisite for a good functional outcome and a prolonged survival. Navigation systems may facilitate this proper positioning. Patient specific templates have been developed to achieve at least the same accuracy than conventional instruments at a lower cost. We hypothesized that there was no learning curve at our academic department when using patient specific templates for TKR instead of the routinely used navigation system.

MATERIAL The first 20 patients operated on for TKR at our academic department using a patient specific template entered the study. All patients had a pre-operative CT-scan planning with a dedicated software. The patient specific templates were positioned on the bone according to the best fit technique. The position of the templates was controlled at each step of the procedure by the navigation system, and eventually corrected to achieve the expected goal. The discrepancy between the initial and the final positioning was recorded. The paired difference between each set of measurement was analyzed with appropriate statistical tests at a 0.05 level of significance.

RESULTS The mean difference between the initial positioning of the tibia template in the coronal plane was 3° ± 4° (p<0.01); 5/10 cases were off the expected range. The mean difference between the initial positioning of the tibia template in the sagittal plane was 4° ± 3° (p<0.001); 6/10 cases were off the expected range. The mean difference between the initial positioning of the femur template in the coronal plane was 2° ± 2° (p<0.01); 3/10 cases were off the expected range. The mean difference between the initial positioning of the tibia template in the sagittal plane was 3° ± 3° (p<0.01); 4/10 cases were off the expected range. After correction of the position according to the navigation system, all templates were placed in the expected range in all directions. There was a trend to improve the initial positioning of the templates at the end of the study, especially for the femoral template.

DISCUSSION The global accuracy of the system tested was less than the navigated reference. The current system should not be used without extensive intra-operative control of the positioning of the templates. However, when the positioning of the templates was adequate, the orientation of the resection was correct. We conclude that the planning and manufacturing process may be adequate, but the question of the intra-operative positioning should be addressed. As the results about positioning improved during the time of the study, one might argue that one of the relevant question is the learning curve of the surgeon when using such templates.

ABSTRACT. Purpose

This study aims to evaluate the two-year post-operative clinical outcomes of patients undergoing total knee arthroplasty (TKA) with computer-assisted surgery (CAS) using a pinless navigation system (BrainLAB® VectorVision® Knee 2.5 Navigation System) versus standard CAS.

Methods

We analysed prospectively collected data from patients who underwent TKA with CAS from November 2008 to October 2012 over a two year follow-up period by a single senior surgeon. Primary outcome measures include Short-Form 36 (SF-36), Oxford Knee (OKS) as well as Knee Society Scores (KSS).

Results

100 patients were recruited in both arms of the study. Patients in the pinless navigation arm underwent a shorter duration of surgery compared to those in the standard CAS arm with 72±13 min and 83±11 min respectively (p<0.001). There was improvement in OKS for both in pinless navigation arm and standard CAS arm from 34±8 to 18±5 (p<0.001) and 34±9 to 18±5 respectively. This improvement was reflected in the SF-36 as well as KSS. The study did not detect any statistically significant differences in clinical outcomes between the two arms at two-year follow-up.

Conclusion

Pinless navigation in TKA with CAS provides improvement in clinical outcomes that are similar to patients who underwent standard CAS with a shorter duration of surgery.

ABSTRACT. Computer-aided surgical systems commonly use preoperative CT scans when performing pelvic osteotomies for intraoperative navigation. These systems have the potential to improve the safety and accuracy of pelvic osteotomies, however exposing the patient to radiation is a significant drawback. In order to reduce radiation exposure, we propose a new smooth extrapolation method leveraging a partial pelvis CT and a statistical shape model (SSM) of the full pelvis in order to estimate a patient's complete pelvis. A SSM of normal, complete, female pelvis anatomy was created and evaluated from 42 subjects. A leave-one-out test was performed to characterise the inherent generalisation capability of the SSM. An additional leave-one-out test was conducted to measure performance of the smooth extrapolation method and an existing "cut-and-paste" extrapolation method. Unknown anatomy was simulated by keeping the axial slices of the patient's acetabulum intact and varying the amount of the superior iliac crest retained; from 0% to 15% of the total pelvis extent. The smooth technique showed an average improvement over the cut-and-paste method of 1.31 mm and 3.61 mm, in RMS and maximum surface error, respectively. With 5% of the iliac crest retained, the smoothly estimated surface had an RMS surface error of 2.21 mm, an improvement of 1.25 mm when retaining none of the iliac crest. This anatomical estimation method creates the possibility of a patient and surgeon benefiting from the use of a CAS system and simultaneously reducing the patient's radiation exposure.

ABSTRACT. Percutaneous fixation of scaphoid fractures has become popular in recent years, mainly due to its reduced complexity compared to open surgical approaches. Fluoroscopy is currently used as guidance for this percutaneous approach, however, as a projective imaging modality, it provides only a 2D view of the complex 3D anatomy of the wrist during surgery, and exposes both patient and physician to harmful X-ray radiation. To avoid these drawbacks, 3D ultrasound has been suggested to provide imaging for guidance as a widely available, real-time, radiation-free and low-cost modality. However, the blurred, disconnected, weak and noisy bone responses render interpretation of the US data difficult so far. In this work, we present the integration of 3D ultrasound with a statistical wrist model to allow development of an improved ultrasound-based guidance procedure. For enhancement of bone responses in ultrasound, a phase symmetry based approach is used to exploit the symmetry of the ultrasound signal around the expected bone location. We propose an improved estimation of the local phase symmetry by using the local spectrum variation of the ultrasound image. The statistical wrist model is developed through a group-wise registration based framework in order to capture the major modes of shape and pose variations across 30 subjects at different wrist positions. Finally, the statistical wrist model is registered to the enhanced ultrasound bone surfaces using a probabilistic registration approach. Feasibility experiments are performed using two volunteer wrists, and the results are promising and warrant further development and validation to enable ultrasound guided percutaneous scaphoid fracture reduction.

ABSTRACT. We present a novel method to derive the surface distance of an osteosynthesis plate w.r.t. the patient-specific surface of the distal femur based on 2D X-ray images. Our goal is to study from clinical data, how the plate-to-bone distance affects bone healing. The patient-specific 3D shape of the femur is, however, seldom recorded for cases of femoral osteosynthesis since this typically requires Computed Tomography (CT), which comes at high cost and radiation dose. Our method instead utilizes two postoperative X-ray images to derive the femoral shape and thus can be applied on radiographs that are taken in clinical routine for follow-up. First, the implant geometry is used as a calibration object to relate the implant and the individual X-ray images spatially in a virtual X-ray setup. In a second step, the patient-specific femoral shape and pose are reconstructed in the virtual setup by fitting a deformable statistical shape and intensity model (SSIM) to the images. The relative positioning between femur and implant is then assessed in terms of displacement between the reconstructed 3D shape of the femur and the plate. A preliminary evaluation based on 4 cadaver datasets shows that the method derives the plate-to-bone distance with a mean absolute error of less than 1mm and a maximum error of 4.7 mm compared to ground truth from CT. We believe that the approach presented in this paper constitutes a meaningful tool to elucidate the effect of implant positioning on fracture healing.

ABSTRACT. A challenging problem in ultrasound based orthopaedic surgery is the identification and interpretation of bone surfaces. Recently we have proposed a new fully automatic ultrasound bone surface enhancement filter in the context of spine interventions. The method is based on the use of a Gradient Energy Tensor filter to construct a new feature enhancement metric, which we call the Local Phase Tensor. The goal of this study is to provide further improvements to the proposed filtering method by incorporating a-priori knowledge about the physics of ultrasound imaging and salient grouping of enhanced bone features. Typical ultrasound scan of the spine, there is a large soft tissue interface present close to the transducer surface with high intensity values similar to those of the bone anatomy response. Typical ultrasound image segmentation or enhancement methods will be affected by this thick soft tissue response. In order to weaken this soft tissue interface we calculate a new transmission map where features deeper in the ultrasound image have higher transmission values and shallow features have lower transmission values. The calculation of this new US transmission/attenuation map allows the proposed image enhancement method to mask out erroneous regions, such as the soft tissue interface, and improve the accuracy and robustness of the spine surface enhancement. The masked US images were used as an input to the LPT image enhancement method. In order to provide a more compact spine surface representation and further reduce the typical US imaging artifacts and soft tissue interfaces we calculate saliency Local Phase Tensor features. The saliency images are computed using Difference of Gaussian filters. Qualitative results, obtained from in vivo clinical scans, show a strong correspondence between enhanced features and the actual bone surfaces present in the ultrasound scans. Future work will include the extension of the proposed method to 3D and validation of the method in the context of intra-operative ultrasound image registration in CAOS applications.

ABSTRACT. Ultrasound (US) imaging is recommended for early detection of Developmental Dysplasia of the Hip (DDH) to guide decisions about possible surgical treatment. However, a number of studies have raised concerns over the efficacy of US in early diagnosis. The main limitation of US-based diagnosis is sub-standard reliability of the primary dysplasia metric measurements: namely, the alpha and beta angles. In this study, we have proposed a novel and automatic method to extract dysplasia metrics from 2D US, which we hope will improve the overall reliability of US-based DDH measurements by removing error due to subjective measurements. We hypothesize that improvements in reliability of dysplasia metric measurements will reduce the chances of missed early-diagnosis, and therefore reduce the need for later complex surgical treatments.

We evaluated performance of the algorithm on 4 infants diagnosed with US scans for DDH. The typical runtime of our algorithm is less than 1 second for an US image. We found a 6° bias between manual and automatic measurements, with automatic measurements tending to be lower in value; the standard deviation in the discrepancy values was also relatively high at 7°. This suggests that there is considerable variability in the angle estimation process, which is typically done manually, which supports our contention that further work needs to be done to establish an accurate and repeatable measurement technique. Further, we found agreements in the Graf-classification types in six out of seven sessions. For the one patient where there was a discrepancy in classification, later US sessions suggest the manual technique possibly missed the opportunity for early detection, in contrast to the automatic method which classified the patient as having evidence of dysplasia. Thus, such an automatic method may improve the reliability of current US-based DDH diagnosis techniques. The primary limitation of this study is that we have done strictly an intra-image discrepancy analysis and have not compared the results with what could be considered a 'gold standard' reference. In future work, we plan to assess these indices on 3D images of the hip and assess the accuracy of proposed 2D and 3D-based automatic index calculation techniques against a 3D reference model.

ABSTRACT. INTRODUCTION Hip and knee joint replacement is nowadays one of the most common surgeries in Germany. The frequency of peri- and post-operative complications varies depending on the study. Since 2001, every hospital in Germany is required to report any peri- and post-operative complication to an external institute for quality control. The purpose of this study was to evaluate the published data of these institutes and to differentiate between the rate of peri- and early postoperative complications of conventional and computer navigated surgical procedures. The hypothesis of the study was that there is no increase in the rate of peri- and early post-operative complications as a result of the navigated surgical procedure. MATERIALS AND METHODS A retrospective analysis of the data on primary total knee and hip replacements between 2004 and 2012 were conducted. The share of navigated procedures, additional operating time due to navigation, the peri- and early post-operative surgical and general rates of complications and the comparison of patient population (age, sex and ASA-classification) were subject of the analysis. RESULTS Overall, the number of implanted knee endoprostheses rose from 110.000 in 2004 to 133.000 in 2012, including a doubling in the share of the navigated knee endoprostheses from 6.8% to 11.2%. Additional operative time for the implantation of knee prostheses decreased from initially 20 min. to 11.3 min.. The rate of patients with at least one surgical intra and post-operative surgical complication decreased nearly 50.0% both, conventional and computer-navigated total knee arthroplasty. Focusing solely on the rate of general post-operative complications, a decrease from 5.9 % (conventional) or 4.7% (navigated) to 2.9%, respectively, was observed.

In the area of hip prostheses, the total number increased from 138.000 (2004) to 152.500 (2012), with a relatively stable share of navigated prostheses from 1.63% to 1.4%. Additional operative time in case of the navigation decreased from 14.8 min. to 5.3 min.. The rate of patients with at least one surgical intra or post-operative complication showed a decrease from 5.0% to 3.0% for both. The rate of post-operative general complications with navigated prostheses showed a higher drop in percentage points from 4.5% to 1.85% than with non-navigated prostheses, which decreased from 5.0% to 3.0%. DISCUSSION Despite high expertise and having used navigation for years, still very few endoprostheses are navigated. However, through a regular application of the navigation, a significant reduction of the operational time was accomplished in knee and hip replacement. The analysis did not show an increase in either peri- or early post-operative appearance of complications.

ABSTRACT. Introduction: Inappropriate soft tissue tension around an artificial hip is regarded as one cause of dislocation or abductor muscle weakness. It has been considered that restoration of leg offset is important to optimize soft tissue tension in THA, while it is unclear what factors determine soft tissue tension around artificial hip joints. The purpose of the present study was to assess how postoperative leg offset influence the soft tissue tension around artificial hip joints. Materials and Methods: The subjects were 89 consecutive patients who underwent mini-incision THA using a navigation system through antero-lateral or postero-lateral approach. Soft tissue tension was measured by applying traction amounting to 40% of body weight with the joint positioned at 0º, 15º, 30º, and 45º of flexion. The distance of separation between the head and the cup was measured using the navigation system. Results: The distance of cup/head separation differed significantly for different angles of flexion, with the greatest distance at 15º of flexion which was 11±5 (SD) mm. Stepwise multiple regression analysis showed that postoperative leg offset discrepancy, antero-lateral approach, preoperative abduction ROM were correlated with the distance of cup/head separation at 15º of flexion. Postoperative leg offset discrepancy were also correlated negatively with the distance of cup/head separation at 0º and 30º of flexion. Conclusion: Postoperative leg offset discrepancy influenced significantly the soft tissue tension around THA at a wider range of flexion.

ABSTRACT. Perthes disease is a childhood disorder often resulting in femoral head deformity. Categorical/dichotomous outcomes of deformity are typical clinically, however quantitative, continuous measures, such as Sphericity Deviation Score (SDS), are critical for studying interventions. SDS uses radiographs in two planes to quantify femoral head deformity. Limitations of SDS may include non-orthogonal planes and lost details due to projections. We applied this method in 3D, with specific objectives to: 1. Develop SDS-like sphericity measures from 3D data 2. Obtain 2D and 3D sphericity for normal and Perthes hips 3. Compare slice-based (3D) and projection-based (2D) sphericity CT images of 16 normal (8 subjects) and 5 Perthes hips (4 subjects) were segmented to create 3D hip models. Ethics board approval was obtained for this study. SDS consists of roundness error (RE) in two planes and ellipsoid deformation (ED) between planes. We implemented a modified SDS which was applied to (a) orthogonal projections simulating radiographs (sagittal/coronal; 2D-mSDS), and (b) largest radii slices (sagittal/coronal; 3D-mSDS). Mean 2D-mSDS was higher for Perthes (27.2 (SD 11.4)) than normals (11.9 (SD 4.1)). Mean 3D-mSDS showed similar trends, but was higher than 2D (Perthes 33.6 (SD 5.3), normals 17.0 (SD 3.1)). Unlike 2D-mSDS, 3D-mSDS showed no overlap between groups. For Perthes hips, 2D-mSDS was consistent with SDS. For normal hips, 2D-mSDS was higher than expected (similar to Stulberg II). Projection-based (2D) measures may produce lower mSDS due to spatial averaging. Slice-based (3D) measures may better distinguish between normal and Perthes shapes, which may better differentiate effectiveness of treatments.

ABSTRACT. Component placement and the individual’s functional posture play key roles in mechanical complications and hip dysfunction after total hip arthroplasty (THA). The challenge is how to measure these. X-rays lack accuracy and CT scans increase radiation dose. A newer imaging modality, EOSTM, acquires low-dose, simultaneous, perpendicular anteroposterior and lateral views while providing a global view of the patient in a functional standing or sitting position, leading to a 3D reconstruction for parameter calculation. The purpose of the present study was to develop an approach using the EOS system to compare patients with good versus poor results after THA and to report our preliminary experiences using this technique. A total of 35 patients were studied: 17 with good results after THA (G-THA), 18 with poor results (P-THA). The patients were operated on or referred for follow-up to a single expert surgeon, between 2001 and 2011, with a minimum follow-up of at least two years. Acetabular cup orientation differed significantly between groups. Acetabular version relative to the coronal plane was lower in P-THA (32º±12º) compared to G-THA (40º±9º) (p=0.02). There was a strong trend towards acetabular cup inclination relative to the APP being higher in P-THA (45º±9º, compared to 39º±7º; p=0.07). Proportions of P-THA vs. G-THA patients with cup orientation values higher or lower than 1 SD from the overall mean differed significantly and substantially between groups. All revision cases had a least four values outside 1 SD, including acetabular cup orientation, sagittal pelvic tilt, sacral slope, femoral offset and neck-shaft angle. This is the first study to our knowledge to provide acetabular, pelvic and femoral parameters for these two groups and the first to provide evidence that a collection of high/low parameters may together contribute to a poor result. The results show the importance of acetabular component placement, in both inclination and version and the importance of looking at individuals, not just groups, to identify potential causes for pain and functional issues. With the EOS system, a large cohort of individuals can be studied in the functional position relatively quickly and at low dose. This could lead to patient-specific guidelines for THA planning and execution.

ABSTRACT. The anterior pelvic plane (APP) is used as a reference in various pelvic surgeries in orthopaedics. Current methods for identifying the APP are limited in accuracy and efficiency. A quick and accurate method for registering the pelvis orientation can be very useful. Previously, we have introduced a Tracked C-arm (TC-arm) system for use with any C-arm fluoroscopy for producing spatially calibrated imaging views. This system has been tried for estimating the APP. Early results, however, has shown limited repeatability in identifying the anterior superior iliac spine (ASIS) landmarks. This study improves the previous algorithms for a robust registration of the APP. A Sawbone pelvis was used, and its APP was marked by radio-dense ball-bearings. In the new addition, the TC-arm allowed segmenting the ASIS in an interactive user-interface by taking guidance from a reference line tangential to the ipsilateral pubic tubercle for marking the most anterior point on the iliac-crest. The imaging and analysis was repeated 10 times. The results were compared to reconstruction of the fiducial markers placed on the true APP. Accuracy of 1.4° and 4.4° were found for registering the pelvic tilt and rotation, correspondingly. The overall accuracy and precision of registration of the APP were 4.7° and 0.82°, correspondingly. The new method showed 7.5 times improvement in repeatability of measuring the pelvic tilt (SD<0.4°) compared to the previous fluoroscopic methods. This technique addresses an important challenge in estimation of the pelvic bone which is crucial for reliable device placement and producing standard radiographic views in surgery.