ABSTRACT. Introduction Varying degrees of posterior glenoid bone loss occurs in patients with end stage osteoarthritis and can result in increased glenoid retroversion. The excessive retroversion can affect implant stability, eccentric glenoid loading, and fixation stresses. Ultimately, the goal is to correct retroversion to restore normal biomechanics of the glenohumeral joint. The goal of this study was to identify the optimal augmented glenoid design based on finite element model analysis which will provide key insights into implant loosening mechanisms and stability. Materials and Methods Two different augmented glenoid designs, posterior wedge and posterior step- were created as a computer model by a computer aided design software (CAD). These implants were virtually implanted to correct 20° glenoid retroversion and the different mechanical parameters were calculated including: the glenohumeral contact pressure, the cement stress, the shear stress, and relative micromotions at the bone cement interface. Results: During abduction, high strain was concentrated around the peg and posterior glenoid bone. Strain was noticeably higher in stepped design (1-2%) than the wedged design (0.4-1.2%). Stepped glenoid models sustained 30% and 70% higher stresses than those experienced by the wedged glenoid implant models at two different corrections. Distractions, or implant pulling away from bone, predicted by the stepped designs were found to be at least twice as much as those by the wedged designs. Similarly, in compression values were 1.5-8 magnitudes higher in stepped designs than those of wedged designs. The summative effect of the values for distraction, compression, and translation show that the stepped design experienced more micromotion. However, for the wedged design, the amount of micromotion was not affected by the size of the augment (8° and 16°). Discussion: Our study showed that the wedged design experienced less stress compared to stepped design with abduction loading. Notably, the wedged design experienced less stress as the size of the wedge increased to correct a more retroverted arthritic glenoid. The step design also had the highest amount of micromotion which ultimately points to increased failures rate and decreased performace.

ABSTRACT. Background: Manufacturers of reverse shoulder arthroplasty (RSA) implants have recently designed innovative implants to optimize performance in rotator cuff deficient shoulders. These advancements are not without tradeoff and can have negative biomechanical effects. The objective of this study was to develop an integrated FEA kinematic model to compare the muscle forces and joint reaction force (JRF) of 3 different RSA designs. Methods: A kinematic model of a normal shoulder joint was adapted from the Delft model and integrated with the OpenSim shoulder model. Static optimizations then allowed for calculation of the individual muscle forces, moment arms and JRF relative to net joint moments. Three dimensional computer models of humeral lateralized design (HLD), glenoid lateral design (GLD), and Grammont design (GD) RSA were integrated and parametric studies were performed. Results: Overall there were decreases in deltoid and rotator cuff muscle forces for all 3 RSA designs. These decreases were greatest in the middle deltoid of the HLD model for abduction and flexion and in the rotator cuff muscles under both internal and external rotation. The joint reactive forces in abduction and flexion decreased similarly for all RSA designs compared to the normal shoulder model, with the greatest decrease seen in the HLD model. Conclusions: These findings demonstrate that the design characteristics implicit in these modified RSA prostheses result in kinematic differences most prominently seen in the deltoid muscle and overall joint reactive forces. Further research utilizing this novel integrated model can help guide continued optimization of RSA design and clinical outcomes.

ABSTRACT. Introduction: Reverse Total Shoulder Arthroplasty (rTSA) is an efficient treatment, to relieve from pain and to increase function. However, scapular notching remains a serious issue and post-operative range of motion (ROM) present much variations. No study compared implant positioning, different implant combinations, different implant sizes on different types of patients representative to undergo for rTSA, on glenohumeral ROM in every degree of freedom.

Material and Methods: From a CT-scan database classified by a senior surgeon, CT-exams were analysed by a custom software Glenosys® (Imascap®, Brest, France). Different glenoid implants types and positioning were combined to different humerus implant types. Range of motion was automatically computed. Patients with an impingement in initialization position were excluded from the statistical analysis. To validate those measures, a validation bench was printed in 3D to analyze different configurations.

Results: 25 patients were included; 50 configurations were realized per patient. The validation bench on 5 configurations retrieved an error of 1,5° ± 0,88°. The impingement rate and ROM were improved using lateralized glenoid implant types, inferior positioning glenoid implant types, 42mm glenospheres, decreased Neck Shaft Angles for humerus implants and humerus inset.

Conclusion: Impingement in resting arm at side position and ROM can be maximized with an adequate implant choice. A surgical planning software could assist the surgeon to choose the best configuration for each patient to maximize the post-operative outcome (scapular notching and global range of motion).

ABSTRACT. PHYSICAL SIMULATION OF THE INTEROSSEOUS LIGAMENTS DURING FOREARM ROTATION F. Péan, F. Carrillo, P. Fürnstahl, O. Goksel Computer-assisted Applications in Medicine (CAiM), ETH Zurich, Switzerland, fabien.pean@vision.ee.ethz.ch

The Interosseous Membrane (IOM) of the forearm is made up of ligaments, which are involved in load balancing of the radioulnar joint and the shaft. Motion models of the forearm are necessary for planning orthopedic surgeries, such as osteotomies, which aim at solving limit of the range of motion or instabilities. However, existing models focus on a pure kinematic approach, omitting the physical properties of the ligaments, thus limiting the range of application by missing dynamical effects. We developed a model that takes into account the mechanical properties of the IOM. We simulated the pro-supination by creating an elastic coupling to the desired motion around the standard axis of rotation. We tested our model on a healthy subject, using CT-reconstructed bone models and literature data for the ligaments. Multiple parameters, including forces of ligaments and positions of landmarks, are output for analysis. The length of the ligaments over pro-supination was in agreement with the literature. Their rest lengths must be recorded in future anatomical studies. The IOM helps in maintaining the contact with cartilage, except in late pronation. Scarring of the central band increases the force generated along the axis of rotation toward the wrist, while scarring of the proximal part does the opposite in pronation. In contrast to kinematic model, the proposed model is helpful to study the effect of physical properties such as ligament forces or scarring on the forearm motion. Future work will be to apply our model to pathological cases, and to compare to clinical observations.

ABSTRACT. The glenohumeral joint is an important joint with large mobility of the human upper extremity. In shoulder arthroplasty patients often has an unsatisfactory outcome. In order to understand the biomechanical complexity of the shoulder, a novel computer controlled experimental shoulder simulator with an innovative muscle control were constructed. The main component of the simulator includes the active pneumatic muscles to replicate the deltoid and the rotator-cuff function and two springs as passive muscle. The aim of this study is to evaluate the impact of a variation of shoulder joint geometries on shoulder biomechanics in the basis of motion analysis. The radius of the glenoid cavity varied from 28-33mm with 2.5mm increment while the radius the humeral head are varied from 20.1-25.1 with 2.5mm increment. The “teach-in” function of the simulator allows an operator to assign the movement to the simulator where the lengths of the pneumatic muscles are recorded. Then the simulator repeats the assigned movement according to the recorded muscles length. The daily living activities includes abduction/adduction, internal/external rotation with adducted arm, and circumduction. The results show promising repeatability of the simulator with minor deviation. However, damage on the surface of the humeral head has been found which should be further studied for both shoulder behavior investigation and the shoulder simulator optimization. Therefore, this study is a decent initial study toward the verification of the simulator and lead to a better understanding of shoulder biomechanical behavior to cope with the clinical problems in the future.

ABSTRACT. Component alignment cannot fully explain TKA performance with regards to patient reported outcomes and pain. Patient specific variations in musculoskeletal anatomy are one explanation for this. Computational simulations allow for the impact of component alignment and variable patient specific musculoskeletal anatomy on dynamics to be studied across populations. This study aims to determine if simulated dynamics correlate with Patient Reported Outcomes. Landmarking of key anatomical points and 3D registration of implants was performed on 96 segmented post-operative CT scans of TKAs. A cadaver rig validated platform for generating patient specific rigid body musculoskeletal models was used to assess the resultant motions. Resultant dynamics were segmented and tested for differentiation with and correlation to a 6 month postoperative Knee injury and Osteoarthritis Outcome Score (KOOS). Significant negative correlations were found between the postoperative KOOS symptoms score and the rollback occurring in midflexion (p<0.001), quadriceps force in mid flexion (p=0.025) and patella tilt throughout flexion (p=0.009, p=0.005, p=0.010 at 10°, 45° and 90° of flexion). A significant positive correlation was found between lateral shit of the patella through flexion and the symptoms score. (p=0.012) Combining a varus/valgus angular change from extension to full flexion between 0° and 4° (long leg axis) and measured rollback of no more than 6mm without roll forward forms a ‘kinematic safe zone’ of outcomes in which the postoperative KOOS score is 11.5 points higher (p=0.013). The study showed statistically significant correlations between kinematic factors in a simulation of postoperative TKR and post-operative KOOS scores. The presence of a ‘kinematic safe zone’ in the data suggests a patient specific optimisation target for any given individual patient and the opportunity to preoperatively determine a patient specific alignment target.

ABSTRACT. In total knee arthroplasty (TKA) the implant design is one key factor for a proper functional restoration of the diseased knee. Therefore, detailed knowledge on the shape (morphology) is essential to guide the design process. In literature, the morphology has been extensively studied revealing differences, e.g. between ethnicity and gender. However, it is still unclear in which way gender-specific morphological differences are sexual dimorphism or explained by differences in size.

The aim of this study was to investigate the morphology of the distal femur under gender-specific aspects for a large group of patients. Statistical analysis was used to reveal significant differences and subsequent correlation analysis to normalize the morphology.

A dataset of n=363 segmented distal femoral bone surface reconstructions (229 female, 134 male) were randomly collected from a database of patients which underwent TKA. In total, 34 morphological features (distances, angles), quantifying the distal femoral geometry, were determined full automatically. Subsequently, graphs and descriptive statistics were used to check normality and gender-specific differences were analyzed by calculating the 95% confidence intervals for women and men separately. Finally, significant differences were normalized by dividing each feature by appropriate distance measurements and confidence intervals were recalculated.

Looking at the confidence 95% intervals, 6 of 34 features did not show any significant differences between genders. Remarkably, this primarily involves angular (relative) features whereas distance (absolute) measurements were mostly gender dependent. Then, we normalized all distance measurements and radii according to their direction of measurement: Features defined in medial/lateral (ML) direction were divided by the overall ML width and those following the anterior/posterior direction were normalized based on the overall AP length. The results demonstrated that gender-specific differences mostly disappear by using an adequate normalization term.

In conclusion, implant sizes (femoral components) should not be linearly scaled according to one dimension. Instead, ML and AP directions should be regarded separately (non-isotropic scaling). Taking this into consideration, gender-specific differences might be neglected.

ABSTRACT. Total knee arthroplasty (TKA) is widely accepted as a successful surgical intervention to treat osteoarthritis and other degenerative diseases of the knee. However, present statistics on limited survivorship and patient-satisfaction emphasize the need for an optimal endoprosthetic care. Although, the implant design is directly associated with the clinical outcome comprehensive knowledge on the complex relationship between implant design (morphology) and function is still lacking.

The goal of this study was to experimentally analyze the relationship between the trochlear groove design of the femoral component (iTotal CR, ConforMIS, Inc., Bedford, MA, USA) and kinematics in an in vitro test setup based on rapid prototyping of polymer-based replica knee implants.

The orientation of the trochlear groove was modified in five different variations in a self-developed computational framework. On the basis of the reference design, one was medially tilted (-2°) and four were laterally tilted (+2°, +4°, +6°, +8°). For manufacturing, we used rapid prototyping to produce synthetic replicates made of Acrylnitril-Butadien-Styrol (ABS) and subsequent post-processing with acetone vapor. The morpho-functional analysis of the replicates was performed in our experimental knee simulator. Tibiofemoral and patellofemoral kinematics were recorded with an optical tracking system during a semi-active flexion/extension (~10° to 90°) motion.

Looking at the results, the patellofemoral kinematics, especially the medial/lateral translation and internal/external rotation were mainly affected. During low flexion, the patella had a more laterally position relative to the femur with increasing lateral trochlear orientation. The internal/external rotation initially differentiated and converged with flexion. Regarding the tibiofemoral kinematics, only the tibial internal/external rotation showed notable differences between the modified replica implants.

We presented a workflow for an experimental morpho-functional analysis of the knee and demonstrated its feasibility on the example of the trochlear groove orientation which might be used in the future for comprehensive implant design parameter optimization, especially in terms of image based computer assisted patient-specific implants.

ABSTRACT. Despite of the high success of TKA, 20% of recipients remain dissatisfied with their surgery. There is an increasing discordance in the literature on what is an optimal goal for component alignment. Furthermore, the unique patient specific anatomical characteristics will also play a role. The dynamic characteristics of a TKR is a product of the complex interaction between a patient’s individual anatomical characteristics and the specific alignment of the components in that patient knee joint. These interactions can be better understood with computational models. Our objective was to characterise ligament characteristics by measuring knee joint laxity with functional radiograph and with the aid of a computational model and an optimisation study to estimate the subject specific free length of the ligaments. Pre-operative CT and functional radiographs, varus and valgus stressed X-rays assessing the collateral ligaments, were captured for 10 patients. CT scan was segmented and 3D-2D pose estimation was performed against the radiographs. Patient specific tibio-femoral joint computational model was created. The model was virtually positioned to the functional radiograph positions to simulate the boundary conditions when the knee is stressed. The model was simulated to achieve static equilibrium. Optimisation was done on ligament free length and a scaling coefficient, flexion factor, to consider the ligaments wrapping behaviour. Our findings show the generic values for reference strain differ significantly from reference strains calculated from the optimised ligament parameters, up to 35% as percentage strain. There was also a wide variation in the reference strain values between subjects and ligaments, with a range of 37% strain between subjects. Additionally, the knee laxity recorded clinically shows a large variation between patients and it appears to be divorced from coronal alignment measured in CT. This suggest the ligaments characteristics vary widely between subjects and non-functional imaging is insufficient to determine its characteristics. These large variations necessitates a subject-specific approach when creating knee computational models and functional radiographs may be a viable method to characterise patient specific ligaments.

ABSTRACT. Improving the accuracy of measuring 6 degree of freedom tibiofemoral kinematics is a crucial step in gait analysis, but skin-marker estimated kinematics are subject to soft tissue artefacts. Fluoroscopic systems have been reported to achieve high accurate kinematics, but their induced irradiation, limited field of view, and high cost hampers routine usage on large patient cohorts. The aim of this study is to assess the feasibility of measuring tibiofemoral kinematics using multi-channel A-mode ultrasound system in cadaver experiment and to assess its achievable accuracy. A full cadaver was placed with its back on a surgery table while its legs were overhanging the edge of the table. Upper body was fixated and right leg was moved by means of pulling a rope. Two bone pins with optical markers were mounted to the femur and tibia separately to measure the ground truth of motion. Six custom holders containing 30 A-mode ultrasound transducers and 18 optical markers were mounted to six anatomical regions. By measuring the bone to ultrasound transducer distance and using the spatial information of the optical markers on the holders, 30 bone surface points were determined. The corresponding bones (femur and tibia) were registered to these acquired points after which the tibiofemoral kinematics were determined. This study presents a multi-channel A-mode ultrasound system and the first results have shown its feasibility of reconstructing tibiofemoral kinematics in cadaver experiment. Although the reconstructed tibiofemoral kinematics is less accurate than a fluoroscopic system, it outperforms a skin-mounted markers system. Thus, this A-mode Ultrasound approach could provide a non-invasive and non-radiative method for measuring tibiofemoral kinematics, which may be used in clinic gait analysis or even computer-aided orthopaedic surgery.

ABSTRACT. In vitro evaluation of knee kinematics remains an essential part during pre-clinical testing of new implants and surgical procedures. To assess the kinematics, markers are rigidly attached to the bone segments and tracked using infrared cameras. Subsequently, the position of the markers relative to the bone is determined using computed tomography (CT). Although the accuracy of the aforementioned, CT-based method is not doubted, no real-time information is provided. Therefore, this paper presents a real-time method that uses a registration phase in combination with a pre-operative CT scan to determine the location of the bone relative to the markers. During this registration phase, the bone surface location is identified touching surface points with a tracked pen. The kinematic parameters obtained using this real-time method is compared to the golden standard, CT-based, method. Under optimal conditions, rotational and translational differences around 1mm and 1degree are obtained. This is in the range of the inter- and intra- observer variability in determining the landmarks used for these kinematic calculations. It is therefore concluded that the accuracy of the real-time method allows effectively evaluating the knee kinematics in six degrees of freedom.

ABSTRACT. Accurate 3D pre-operative planning shows significance of improving the precision of Total Hip Arthroplasty (THA) and Total Knee Arthroplasty (TKA). Since CT acquisition leads to high radiation exposure to patients, it is clinically desirable to find an alternative to CT scan for planning THA or TKA such as patient-specific 2D-3D reconstruction from a limited number of 2D calibrated X-ray images acquired with much lower radiation dose e.g. EOS imaging. Feature-based 2D-3D non-rigid registration based on the construction of statistical shape model (SSM) as a priori has been applied to reconstruct the surface models of proximal femur, and also the surface reconstruction of lower extremity for TKA has been validated in a cadaveric study by Zheng et al. On the other hand, intensity-based 2D-3D non-rigid registration can reconstruct the patient-specific intensity volumes like CTs to allow an insight into lower extremity morphology such as intramedullary anatomy, which can provide more comprehensive information in routine clinical practice. In this study, we will present an atlas-based 2D-3D reconstruction method and introduce its application to reconstruct the intensity volumes of lower extremity. Moreover, we take the articulation in the knee joint into consideration so as to avoid the penetration between femur and tibia which is favourable for the pre-operative planning. The results of the experiments demonstrated the efficacy of the proposed method on reconstructing the lower extremity morphology as well as the intramedullary canal anatomy.

ABSTRACT. Active Shape Models (ASM) have been widely used in the literature for the extraction of the tibial and the femoral bones from MRI. These methods use Statistical Shape Models (SSM) to drive the deformation and make the segmentation more robust. One crucial step for building such SSM is the shape correspondence (SC). Several methods have been described in the literature. The goal of this paper is to compare two SC methods, the Iterative Median Closest Point-Gaussian Mixture Model (IMCP-GMM) and the Minimum Description Length (MDL) approaches for the creation of a SSM, and to assess the impact on the accuracy of the femur segmentation in MRI.

28 MRI of the knee have been used. The validation has been performed by using the leave-one-out cross-validation technique. An ASMMDL and an ASMIMCP-GMMM has been built with the SSMs computed respectively with the MDL and IMCP-GMM methods. The computation time for building both SSMs has been also measured.

For 90% of data, the error is inferior to 1.78 mm and 1.85 mm for respectively the ASMIMCP-GMM and the ASMMDL methods. The computation time for building the SSMs is five hours and two days for respectively the IMCP-GMM and the MDL methods.

Both methods seems to give, at least, similar results for the femur segmentation in MRI. But (1) IMCP-GMM can be used for all types of shape, this is not the case for the MDL method which only works for closed shape, and (2) IMCP-GMM is much faster than MDL.

ABSTRACT. Reconstruction of pelvic bone defect after resection for bone tumours is a challenging procedure especially when the hip joint is involved due to the anatomy and the complex biomechanical and structural function of the pelvic ring. This surgery is associated to high complication rate. The additive 3D printing technology allows to produce trabecular titanium custom based implants with an accurate planning of resection using bone cutting jigs. From August 2013 to January 2017, we treated 8 patients for bone pelvic sarcoma with custom-made osteotomy jigs (Nylon) and custom-made trabecular titanium prosthesis produced through rapid prototyping technology based on mirroring of the contralateral hemipelvis. Mean follow up time was 18 months (range 2-30) Wide margins were obtained in all cases, in one a local recurrence developed. Surgical time was 4 hours average (from 180 to 250 mins). No postoperative complications were reported. Rapid prototyping is a promising technique in order to achieve wide surgical margins and restore the anatomy in pelvic bone tumour resection as well as reducing complications

ABSTRACT. Purpose Bone tumor resection and subsequent reconstruction remains challenging for the surgeon. Obtaining adequate margins is mandatory to decrease the risk of local recurrence. Patient specific instruments (PSI) may greatly improve our ability to achieve the targeted resection. Surgical time might significantly be reduced as well. Reconstruction may gain in rapidity and quality especially when using allografts as PSI can also be designed for allograft cutting. This retrospective study analyzes the mid- to long-term follow up of PSI usage. Methods Between 2011 and 2016, we used PSI to remove bone tumors in 30 patients. Medical files were reviewed for type, size and site of the tumor, metastatic status pre-and post-operatively, resection margins within the bone and the soft tissues, local recurrence, the use of an allograft and a specific graft cutting guide or not for the reconstruction, the fusion of the allograft when applicable, the follow-up time and early/late complications. Results Mean follow-up was 27±20 months. Mean operating time was 6h02±3h44. Mean size of the tumors was 8,4±4,7cm and location was the upper limb in 5 cases, inferior limb in 15 cases and the pelvis in 10 occurrences. Metastatic disease developed postoperatively in 5 patients. Surgical margins in the bone were R0 in all cases but one where a R1 surgery was planned to preserve a nerve root. We did not observe any local recurrence in the bone section. Within soft tissues, margins were classified as R0 in 28 patients and R1 in 2 patients. In 26 cases, an allograft was used to reconstruct the bone defect. In 23 of those patients, the allograft was selected by CT scan and cut using a PSI. In the 3 allografts cut free-handily, only one demonstrated a fusion. Of the 23 cut with a guide, 12 fused completely, 2 demonstrated a partial fusion and 9 were not fused at the last follow-up. At the last follow-up, 2 patients were dead of disease, 5 were alive with metastatic disease and 23 were alive without disease. Conclusions PSI seems to improve the immediate outcome (obtained margins) as well as mid- to long-term follow-up (local recurrence). The reconstruction seems improved too since the bone fusion is relatively better than reported in the literature. With a longer follow-up, these evidences should be stronger to definitely make PSI the best option for bone tumor resection.

ABSTRACT. Computer Assisted Surgery (CAS) and Patient Specific Instrumentation (PSI) have been reported to increase accuracy and predictability of tumor resections. The technically demanding joint-preserving surgery that retains the native joint with the better function may benefit from the new techniques. This cadaver study is to investigate the surgical accuracy of CAS and PSI in joint-preserving surgery of knee joint. CT scans of four cadavers were performed and imported into an engineering software (MIMICS, Materialise) for the 3D surgical planning of simulated, multiplanar joint-preserving resections for distal femur or proximal tibia metaphyseal bone sarcoma. The planned resections were transferred to the navigation system (OrthoMap 3D, Stryker) for navigation planning and used for the design and fabrication of the PSI. Each of the four techniques (freehand, CAS, PSI and CAS + PSI) was used in four joint-preserving resections. Location accuracy (the maximum deviation of distance between the planned and the achieved resections) and bone resection time were measured. The results were compared by using t-test (statistically significant if P< 0.05). Both the CAS+PSI and PSI techniques could reproduce the planned resections with a mean location accuracy of < 2 mm, compared to 3.6 mm for CAS assistance and 9.2 mm for the freehand technique. There was no statistical difference in location accuracy between the CAS+PSI and the PSI techniques (p=0.92) but a significant difference between the CAS technique and the CAS+PSI (p=0.042) or PSI technique (p=0.034) and the freehand technique with the other assisted techniques. The PSI technique took the lowest mean time of 4.78 ±0.97min for bone resections. This was significantly different from the CAS+PSI technique (mean 12.78 min; p < 0.001) and the CAS technique (mean 16.97 min; p = < 0.001). CAS and PSI assisted techniques help reproduce the planned multiplanar resections. The PSI technique could achieve the most accurate bone resections (within 2mm error) with the least time for bone resections. Combining CAS with PSI might not improve surgical accuracy and might increase bone resection time. However, PSI placement on the bone surface depends only on the subjective feeling of surgeons and may not apply if the extraosseous tumor component is large. Combining CAS with PSI could address the limitations.

ABSTRACT. Navigation-assisted surgery has been reported to enhance resection accuracy in bone sarcoma surgery. Patient-specific instruments (PSIs) have been proposed as a simpler alternative with fewer setup facilities. We investigated the use of 3D surgical planning and PSI in realizing computer planning of complex resections in bone sarcoma patients with regards to surgical accuracy, problems, and early clinical results. We retrospectively studied twelve patients with bone sarcoma treated surgically by PSIs with 3D planning. The procedure was planned using engineering software. The resection accuracy was accessed by comparing CT images of tumor specimens with the planned in seven patients. Mean age was 30.9 (9 – 64). Mean follow-up was 3.1 year (0.5 – 5.3). 31 planes of bone resections were successfully performed using the technique and were considered accurate. The mean time required for placing PSIs was 5.7 minutes (1 – 10) and performing bone osteotomies with the assistance of PSIs was 4.7 minutes (2 – 7). The mean maximum deviation error was 1.7mm (0.5 – 4.4). One PSI was broken during bone resection, and one patient needed re-resection using the same PSI. One pelvic patient died of local recurrence and lung metastases six months postoperatively. One patient developed a soft tissue local recurrence and lung metastasis at 20 months after surgery. The mean MSTS functional score was 27.9 (21 – 30). There were no complications related to 3D planning and PSIs. In selected patients, 3D surgical planning and PSIs replicate complex bone resections and reconstructions in bone sarcoma surgery. Comparative studies with conventional or navigation-assisted resections are required.

ABSTRACT. In recent years, there has been a growing interest to incorporate ultrasound into computer assisted orthopaedic surgery procedures in order to provide non-ionizing intra-operative imaging alternative to traditional fluoroscopy. However, identification of bone boundaries still continues to be a challenging process due low signal to noise ratio and imaging artifacts. The quality of the collected images also depends on the orientation of the ultrasound transducer with respect to the imaged bone surface. Shadow region is an important feature indicating the presence of a bone surface in the collected ultrasound data. In this work, we propose a framework for the enhancement of shadow regions from extended field of view spine ultrasound data. First bone surfaces are enhanced using a combination of local phase based image features. These enhanced features are used as an input to a L1 norm based regularization method which emphasized uncertainty in the shadow regions. Validation on phantom and in vivo experiments achieve a mean dice coefficient value of 0.93 and 0.9 respectively.

ABSTRACT. Robotic surgical systems reduce the cognitive workload of the surgeon by assisting in guidance and operational tasks. As a result, higher precision and a decreased surgery time are achieved, while human errors are minimised. However, most of robotic systems are expensive, bulky and limited to specific applications. In this paper a novel semi-automatic robotic system is evaluated that offers the high accuracies of robotic surgery while remaining small, universally applicable and easy to use. The system is composed of a universally applicable handheld device, called Smart Screwdriver (SSD) and an application specific kinematic chain serving as a tool guide. The guide mechanism is equipped with motion screws. By inserting the SSD into a screw head, the screw is identified automatically and the required number of revolutions is executed to achieve the desired pose of the tool guide. The usability of the system was evaluated according to IEC 60601 1 6 using pedicle screw implementation as an example. The achieved positioning accuracies of the drill sleeve were comparable to those of fully automatic robotic systems with -0.54 ± 0.93 mm (max: -2.08 mm) in medial/lateral-direction and 0.17 ± 0.51 mm (max: 1.39 mm) in cranial/caudal-direction in the pedicle isthmus. Additionally, the system is cost-effective, safe, easy to integrate in the surgical workflow and universally applicable to applications in which a static position in one or more DOF is to be adjusted.

ABSTRACT. Pedicle screw fixation is a technically demanding procedure with potential difficulties and reoperation rates are currently on the order of 11%. The most common intraoperative practice for position assessment of pedicle screws is biplanar fluoroscopic imaging that is limited to two-dimensions and is associated to low accuracies. We have previously introduced a full-dimensional position assessment framework based on registering intraoperative X-rays to preoperative volumetric images with sufficient accuracies. However, the framework requires a semi-manual process of pedicle screw segmentation and the intraoperative X-rays have to be taken from defined positions in space in order to avoid pedicle screws’ head occlusion. This motivated us to develop advancements to the system to achieve higher levels of automation in the hope of higher clinical feasibility. In this study, we developed an automatic segmentation and X-ray adequacy assessment protocol. An artificial neural network was trained on a dataset that included a number of digitally reconstructed radiographs representing pedicle screw projections from different points of view. This model was able to segment the projection of any pedicle screw given an X-ray as its input with accuracy of 93% of the pixels. Once the pedicle screw was segmented, a number of descriptive geometric features were extracted from the isolated blob. These segmented images were manually labels as ‘adequate’ or ‘not adequate’ depending on the visibility of the screw axis. The extracted features along with their corresponding labels were used to train a decision tree model that could classify each X-ray based on its adequacy with accuracies on the order of 95%. In conclusion, we presented here a robust, fast and automated pedicle screw segmentation process, combined with an accurate and automatic algorithm for classifying views of pedicle screws as adequate or not. These tools represent a useful step towards full automation of our pedicle screw positioning assessment system.

ABSTRACT. We present an analysis of manual and computer-assisted preoperative pedicle screw placement planning. Preoperative planning of 256 pedicle screws was performed manually twice by two experienced spine surgeons (M1 and M2) and automatically once by a computer-assisted method (C) on three-dimensional computed tomography images of 17 patients with thoracic spinal deformities. Statistical analysis was performed to obtain the intraobserver and interobserver variability for the pedicle screw size (i.e. diameter and length) and insertion trajectory (i.e. pedicle crossing point, sagittal and axial inclination, and normalized screw fastening strength). In our previous study, we showed that the differences among both manual plannings (M1 and M2) and computer-assisted planning (C) are comparable to the differences between manual plannings, except for the pedicle screw inclination in the sagittal plane. In this study, however, we obtained also the intraobserver variability for both manual plannings (M1 and M2), which revealed that larger differences occurred again for the sagittal screw inclination, especially in the case of manual planning M2 with average differences of up to 18.3°. On the other hand, the interobserver variability analysis revealed that the intraobserver variability for each pedicle screw parameter was, in terms of magnitude, comparable to the interobserver variability among both manual and computer-assisted plannings. The results indicate that computer-assisted pedicle screw placement planning is not only more reproducible and faster than, but also as reliable as manual planning.

ABSTRACT. In orthopaedic spine surgery pedicle screw systems are used for stabilisation of the spine after injuries or disorders. With an percutaneous operation method surgeons are faced with huge challenges compared to an open surgery, but it`s less traumatic and the patient benefits with a faster rehabilitation and less traumatic injuries. The screw positions and the required rod dimensions for the stabilizing connection between the screws are hard to define without an open view on the operating field. Because of these facts a new smart device based system for rod shape determination was invented. Therefore, an application was developed, which integrates a localizer module to get the position data of the pedicle screws, with help of rigid bodies placed on top of the pedicle screws down-tubes. An algorithm was developed to choose the best fitting rod to connect the pedicle screws with help of calculating the rod length and the rod radius. The system was tested in a test scenario where four pedicle screws were drilled into a wooden plate. The positions of the screws were adjusted to fit a curved and a straight rod. In the test scenario the application chose always the rod correctly.

ABSTRACT. To introduce a new robot-assisted surgical system for spinal posterior fixation which called TiRobot, based on intraoperative three-dimensional images. TiRobot has three components: the planning and navigation system, optical tracking system and robotic arm system. By combining navigation and robot techniques, TiRobot can guide the screw trajectories for orthopedic surgeries. In this randomized controlled study approved by the Ethics Committee, 40 patients were involved and all has been fully informed and sign the informed consent. 17 patients were treated by free-hand fluoroscopy-guided surgery, and 23 patients were treated by robot-assisted spinal surgery. A total of 190 pedicle screws was implanted. The overall operation time were not different for both groups. None of the screws necessitated re-surgery for revised placement. In the robot-assisted group, assessment of pedicle screw accuracy showed that 102 of 102 screws (100%) were safely placed (<2 mm, category A+B). And mean deviation in entry point was 1.70 +/- 0.83mm, mean deviation in end point was 1.84 +/- 1.04mm. In the conventional freehand group, assessment of pedicle screw accuracy showed that 87 of 88 (98.9%) were safely placed (<2 mm, category A+B), 1 screw fall in category C, mean deviation in entry point was 3.73 +/- 2.28mm, mean deviation in end point was 4.11 +/- 2.31mm. This randomized controlled study verified that robot-assisted pedicle screw placement with real-time navigation is a more accuracy and safer method, and also revealed great clinical potential of robot-assisted surgery in the future.

ABSTRACT. Introduction: Pedicle screw (PS) insertion has been criticized for its risk of serious injury to neurovascular structures. Although computed tomography (CT)-based navigation has been developed to avoid such complications, perforation remains an issue, even with the aid of additional guidance. We clarify screw perforation rate and direction in 359 consecutive patients treated using CT-based PS insertion and present important considerations for more accurate screw placement.

Methods: The medical records of 359 consecutive patients who underwent PS insertion involving C2-L5 using a CT-based navigation system were reviewed. Postoperative CT images were analyzed to evaluate the accuracy of screw placement. We investigated both rate and direction of screw perforation according to vertebral level.

Results: Of the 3413 PS that were inserted, 3.0% (104/3413) were judged as Grade 3 (more than 4mm) perforations. Allover perforation rates by vertebral level were shown in Table 1. The rate of these perforations was 5.0% for C2, 7.8% for C3-5, 3.9% for C6-7, 3.4% for T1-4, 3.5% for T5-8, 1.4% for T9-12, and 1.7% for L1-5. We also analyzed the odds ratio (OR) for screw perforation in vertebrae accounting for the effects of age and disease. Multivariate analysis identified that PS insertions at C3-5 (OR 4.9, 95% CI 2.2-10.9; p<0.001) were significantly associated with Grade 3 screw perforation as compared with that of L1-5.

Conclusions: Even with CT-based navigation, careful insertion of PS is needed in the middle cervical spine because of a significantly higher perforation rate as compared with the lumbar region.

ABSTRACT. Introduction: Due to the increasing number of fragility fractures of the pelvis also an elevation of implant failures of sacroiliac screw fixation was observed. Nevertheless, the clinical relevance and causes for sacroiliac screw fixation failure are still unknown possibly related to the less reliable measurement techniques for the posterior pelvic ring. Moreover, measurement techniques for implant failure were not reported. Materials and Methods: A lateral compression facture of nine cadaveric body donors pelvis and fixed by sacroiliac screw fixation. A CT scan and X-rays (anterior-posterior, inlet, outlet) were made. One sacroiliac screw was approximately 2 cm manually retracted and measured by placing a K-wire parallel to the sacroiliac screw. A photograph was taken placing a reference below. X-rays (anterior-posterior, inlet, outlet) and a 3D scan were obtained. The retraction was measured three times by three independent observers using ImageJ for photographic measurements (ImageJ 1.44p, National Institutes of Health, USA) and Mimics 16 (Materialise, Leuven, Belgien) for X-ray and 3D scan measurements. The CT scans were segmented and a contour based registration of the 3D models and post-retraction X-rays was performed and retraction was measured as vector. Results: Screw retraction in the 3D scans was 19.5 ± 2.4 mm. Only the 3D registration using the X-ray module and two X-ray measurements presented a difference less than 5 mm compared to the 3D scans. The interrater reliability was except for one observer for all techniques very good (ICC(1)) ≥ 0.81. Only X-ray and photograph measurements yielded a very good interrater reliability (ICC (A,1)) ≥ 0.81.

ABSTRACT. Introduction

Successful results after knee replacement are based on accurate bone cuts. In most cases intramedullary cutting guides are used for correct distal femoral cut and are optional for the tibial cuts. The use of intramedullary guides is problematic in cases with femoral deformity and or other medullary lesions (tumor, infection etc.). For this cases computer assisted technology is enabler.

Materials and Methods

We retrospectively reviewed our computerized arthroplasty database using Orthosoft system. Between 2009 and 2016 we performed about 1400 primary knee replacements; in 134 (9.6%) we used computerized navigation assistance. 12 patients in the navigation group (8.95) had severe distal femoral deformities or other pathology precluding the use of the conventional intramedullary cutting guides ("enabler cases'). In all cases the navigation enables accurate cuts with no outliers (deviation from preplanning of less than 3ͦ . The procedure was longer than the conventional (average 10 min.), we did not observed any intraoperative or early complications in these patients. The long term outcome was simmilar to our conventional database outcome.

Conclusion

Computerized navigation was found to be improving tool in reducing outliers in knee replacement surgeries. When facing cases with severe deformity or other contraindication for intramedullary guidance It both enables the performance of the procedure and helps create a well aligned knee replacement. The accuracy that was achieved in this selective group indicates the quality and the potential of computer assisted technology.

ABSTRACT. Computer assisted external fixators in the treatment of long bone deformities have become popularity in recent years. The main advantage of the system is providing the possibility of manipulate the bone segment without any changes in the applied system. In addition to the deformity correction these systems can also be used successfully in the treatment of trauma cases.

The aim of current study is to report the use of computer assisted external fixators which are placed onto the previously placed Schanz screws in long bone fractures.

Material method Between years 2009and 2016 7 tibial, 3 femoral and 2 humeral fractures of 12 patients were treated with computer assisted external fixators adapted onto the previously applied Schanz screws. All 12 patients were treated with multiaxial unilateral fixtors initially. After a loss in the fracture reduction previously applied fixator has been disassembled and a computer assisted circular external fixator was adapted to the previously placed Schanz screws in the outpatient clinic without anesthesia. Consequently the fracture reduction has been provided by using the fixator software. Finally the fixators were removed when bone healing has been achieved. ,

Results The mean age fixator converting time was 18.4 days (range 2-35 days). The mean computer assisted external fixator time was 146 days (range 60-180 days). Bone union was achieved in each patient. There was no refracture or deformity following the treatment.

Discussion

Computer assisted external fixators are systems which can be used both in the treatment of bone deformities and acute fractures. With a proper preoperative planning the computer assisted system enables the surgeon to reduce the fractures or correct the deformities and modify the treatment during the follow-up periods. In the current study the system has been applied to the already placed Schanz screws. the software of the system allows to reduce the fragments with proper definition of the mounting parameter of the settled system. So that no additional surgical procedures are necessary to convert the treatment to the computer assisted fixator. From our results we can conclude that converting the previously applied external fixator system to a computer assisted one is a successful method for the cases with a loss of reduction in the fracture treatment.

ABSTRACT. Modern operating rooms (OR) undergo a constant rise in the amount and complexity of technical systems. Due to a lack of inter-device communication and integration, each device works stand-alone resulting in redundant sensors, input-devices, monitors and – last but not least - crowded ORs and error prone human-machine-interaction. Therefore, various manufacturers such as Brainlab and Karl Storz provide proprietary integrated workstations. However, those “monolithic” solutions restrict the flexibility of the users and the clinical operators regarding integration of innovative third party devices. In view of this, the OR.NET initiative (www.ornet.org) strives to develop international open standards for secure dynamic networks of medical devices in ORs. In the scope of the OR.NET project, based on service oriented architecture (SOA), the SDC (Service-oriented Device Connectivity) approach is currently in the process of standardization under IEEE 11073 to link medical devices in the OR (short OR.NETwork). It paves the way to interoperability between various medical devices due to its independence of license holders. However, the SDC network does not suit real time (RT) requirements of a deterministic data transmission and low maximum latency e.g. for robotic applications. This paper shows an approach to extend the secure dynamic OR by a real-time capable network to allow the integration of robotic systems. Exemplarily, this paper outlines an orthopaedic robotic system that is released by a universal configurable footswitch. This significantly extends the scope of applications for integrated ORs with the IEEE 11073 standard.

ABSTRACT. To perform or not to perform patellar resurfacing (PR) during primary total knee arthroplasty is a persistent controversy. Because no evidence-based optimal decision strategies and surgical techniques exist to date, the PR remains based on intuitive reasoning. Today, PR is mostly performed by sawing and aligning the resection by hand and manually placing the implant. Therefore, this study implemented software and tools to conduct a complete computer-assisted PR. The study used both artificial patella bones (AP) and patella from cadaveric specimens (SP) to assess the accuracy of the transferred preoperative planning. Software was designed for the preoperative planning, the execution and analysis of both the experiments with AP and SP. The planning software allows a virtual patellar resection and implant positioning in compliance with the main criteria based on literature. The transfer of planning data occurs by matching the 3D bone geometry with corresponding surface palpation data using an optical measurement system (OMS). To perform the surgery, a trackable resection guide and drill template was utilised. The findings regarding the implant positioning show good similarity between the AP and SP experiments. The results regarding the patella resection show comparable accuracies for the application to AP and SP. However, it seems that it is required to utilise a state-of-the-art OMS with smaller tool geometries to improve the general accuracy and to allow the tracking tools to be closer positioned to the plane or points of interest of the resection and drilling devices, hence further decreasing rotational and translation error sources.

ABSTRACT. Success in revision total knee arthroplasty (TKA) requires the understanding of the failure mechanisms of the preceding implant. A newly developed and novel computer assisted orthopaedic surgery system (CAOS) was used not only to determine the failure modes of the primary implant, but also to establish a well aligned diaphysis-engaging prosthesis in accordance with the mechanical axis.

Tracking of the femur was accomplished with a metaphyseal clamp which allowed access to the intramedullary canal while avoiding proximal diaphyseal pinning of the bone. The tibial tracker was pinned distal to the anticipated stem. Relevant landmarks were acquired, and varus/valgus stability curves were generated to gain insight into the potential causes of implant failure. The mechanical axis was established and compared with the anatomic axes of both the femur and the tibia. Newly developed stepped femoral and tibial blocks were used to make the bone cuts perpendicular to the mechanical axis while allowing for augmentation in the case of bone loss. The system also allowed for calculation of the stem offset using virtual points acquired from the anterior cortex and the medial and lateral cortices.

Based on early clinical experience, analysis of the hip-knee-ankle angle using both the CAOS system and post-operative radiographs showed no outliers exceeding ±3° of varus/valgus deviation from the mechanical axis.

In an environment where there are limited intra-operative tools available for providing quantitative information for the reasons of previous failures, this system provides quantitative data to aid the surgeon in providing a well aligned revision TKA.

ABSTRACT. Accurate leg alignment during conventional revision total knee arthroplasty (TKA) can be challenging due to bone loss, difficulty in identifying bony landmarks, and dependency on the engagement between stem and intra-medullary canal. Computer-assisted orthopaedic surgery (CAOS) may offer assistance to overcome these limitations. The aim of this study was to report the accuracy of leg alignment based on a cohort of CAOS revision TKAs.

A retrospective review was performed on the leg alignment in 44 CAOS revision TKAs, including 25 revision TKAs and 19 complex primary TKAs. All surgeries were performed by one experienced surgeon using an imageless CAOS system. The leg alignment, namely hip-knee-ankle (HKA) angle, was measured both during the revision TKA by the CAOS system, and postoperatively based on the full-length weight bearing standing radiograph. Percentage of outliers in alignment (>3° from mechanical axis) was calculated for both measurements. Significance was defined as p < 0.05.

The HKA angle was 180.1°±1.3° (177.1°-183.6°) and 179.0°±2.0° (174.0°-183.0°) when measured from CAOS system and from post-operative radiograph; respectively. Significant difference was found between the two measurements (1.1°±1.9°, p<0.05). Percentage of outliers was 2.3% (1/44) and 9.1% (4/44) when measured from CAOS system and from post-operative radiograph; respectively.

This study supports the use of an imageless CAOS system for revision TKA in order to improve leg alignment, with the data demonstrating substantial reduction of alignment outliers compared to conventional surgery. Similar difference in HKA angle between the two measurement methods was reported previously. The difference may be due to cementing and variation in limb orientation during radiographing.

ABSTRACT. Introduction: Preoperative planning and patient-specific instruments have grown in popularity across a wide range of orthopaedic subspecialties to improve the accuracy of implant selection and positioning. Recent literature on total shoulder arthroplasty (TSA) has focused on patient specific instrumentation to improve glenoid component placement.(Hasan 2002, Iannotti 2005, Franta 2007, Verborgt 2011) However, nearly all this research has focused on the glenoid side of the joint implant, while little research has looked at the humeral head side. The purpose of our study was to compare optimal methods of preoperative planning using manufacturer and independent simulation software for selection of ideal humeral component in TSA. Methods: A retrospective analysis was completed of 15 patients who underwent a total shoulder arthroplasty with preoperative computerized tomography studies and pre-operative planning simulations. A detailed independent preoperative planning simulation (IPPS) (OrthoVis) was performed and compared to an automated manufacturer preoperative simulation (AMPS) (ArthrexVIP) recommendations. Intraoperative implant selection was recorded by a blinded investigator from preoperative simulation recommendations. Using paired statistical t tests the plans were compared to intraoperative implant selection to determine significant differences and utilize of preoperative planning software.

Results: There was no significant difference between average IPPS humeral best-fit sphere (50.0mm) compared to recommended AMPS plan size of 50.1mm (p=0.757). Both plans measured best-fit sizes (IPPS vs. AMPS) had excellent prediction of average intraoperative implant size, 50.5mm (p=0.367; p=0.185). (p=0.508). Overall there was a high correlation between both methods of preoperative humeral head size measurements (IPPS vs AMPS) and intraoperative implant selection for humeral head size (R=0.843, R=0.867) but little correlation between IPPS humeral head thickness and intraoperative humeral head thickness (R=-0.066).

Discussion: Our results support the use of either commercially available or independent preoperative planning simulations software to guide intraoperative humeral head implant size but not thickness. Consideration still needs to be made in assessing soft tissue tensioning when selecting optimal humeral head thickness intraoperatively

ABSTRACT. The aim of this article is to present an integrated approach based on new applications for orthopaedic trauma treatment and rehabilitation. The integrated approach that we are proposing aims to provide a continuity of care from hospital admission to social and work re-entry and it is based on the following phases: (1) pre-operative planning based on i) clinical evaluation, ii) multimodal sensor data acquisition and processing and c) an accurate biomechanical model of the joint(s)/segment(s) to be treated; (2) surgical treatment (even using robotic devices); (3) intra-operative assessment using high-quality online 2D and 3D images; (4) post-operative assessment (using mechanical, ultrasound, electromagnetic, optical techniques,…), where pre-operative data acquisition and processing provide additional inputs; (5) precision rehabilitation, where a planning phase based on the outputs of the previous phase provides data and information needed during the hospital rehabilitation (using conventional and robot-assisted treatments). The entire process is intended to lead to an optimal functional recovery and to increase the quality of life of patients. The name of such integrated approach is "Precision Orthopaedic Surgery - Precision Orthopaedic Rehabilitation" (POS-POR). The success of the entire process depends heavily on the success of each of the phases. Nonetheless, one crucial aspect is represented by the biomechanical data recorded during the surgical treatment phase representing essential inputs to the musculoskeletal model to be used during the robot-assisted hospital rehabilitation. The proposed POS-POR integrated approach may provide potential advantages to the achievement of optimal functional recovery that represents the final objective of the entire surgical and rehabilitation orthopaedic process.

ABSTRACT. INTRODUCTION In total hip arthroplasty (THA), the orientation of the acetabular cup is an important factor in preventing postoperative complications such as dislocation, accelerated wear and loosening. Computed tomography (CT)-based navigation is a well-documented assist tool for optimal implant placement, and several reports have described clinical outcomes in terms of the precision of the cup angle. In our institution, CT-based navigation system (Stryker) is used from 2004. The objective of this study is to investigate the accuracy of cup placement in THA using CT-based navigation system. MATERIALS AND METHODS From January 2012 to June 2014, a total of 363 hips in 335 patients underwent THA at our institution, using a CT-based navigation system (Stryker CT-Hip System, Stryker-Leibinger, Freiberg, Germany). Among them, in 336 hips (92.5%) the navigation procedure was carried out successfully without loosening of trackers, and 336 hips were analysised so far (54 men, 282 women) . The mean age of the patients was 63 years (range: 24–90 years) and mean body mass index (BMI) was 24 (range: 15–40). Using two personal computers, we adjusted the differences in pelvic orientation and inclination between preoperative and postoperative CT images, according to the method described by Kyo, et al. Cup alignments (inclination and anteversion) were measured from the postoperative CT images. Then clinical accuracy (difference between preoperative planning and postoperative CT measurement) and measurement error (difference between intraoperative record and postoperative CT measurement) were computed and verified. Furthermore, approach (anterior and posterolateral approach groups) and surgeons (expert and beginner groups) compared, respectively. RESULTS The clinical accuracy of cup inclination and anteversion was 1.4°±1.4° (range; -6° to 7° ) and 1.9° ± 1.9° (range; -7° to 8° ), respectively. The measurement error of cup inclination and anteversion was 1.0±0.9° (range; -3° to 4°) and 1.1±0.8° (range; -3° to 3°), respectively (Table 1). No significant differences in accuracy were identified between anterior and posterolateral approaches or between expert and beginner surgeon groups. In conclusion, clinical accuracy and measurement error of cup alignment using CT-based navigation system for 336 cases of THA was excellent.

ABSTRACT. Biomechanical models of the knee have the potential to predict joint mechanics and improve the therapeutic outcome. In case of total knee arthroplasty (TKA) knowledge on post-operative acting joint forces and its consideration in pre-operative planning might increase implant longevity and patient satisfaction. The goal of this study was to evaluate 2D biomechanical models of the knee suitable for pre-operative TKA planning and to validate them based on in vivo measurements of 9 patients treated with instrumented knee implants.

An extensive literature research on 2D biomechanical models of the knee with the requirement that patient-specific model adaption can be performed with data available in the conventional clinical workflow revealed the models of Maquet and Minns. Both models require one-leg stand AP long-leg radiographs and they output the resultant tibiofemoral joint force. For model validation, data from the freely accessible OrthoLoad database (www.orthoload.com) containing forces acting in orthopaedic implants have been used.

Looking at the results, Minns’s model was in the same range as the reference forces (deviation: -55 %BW to +80 %BW) as well as Maquet’s (deviation: -80 %BW to +50 %BW). Overall, the root-mean-square error (RMSE) was smallest for Minns’ model (48.36 %BW against 51.48 %BW). The patient-specific adaptation process plus simulation took 7 to 8 minutes making the models applicable in the conventional clinical workflow with acceptable costs.

In terms of clinical application, the models’ output might be consulted, e.g. to minimize the magnitude of the resultant joint force, target a specific orientation or load distribution in order to consider mechanical factors which might be associated with premature implant failure.

ABSTRACT. INTRODUCTION Iliosacral and transsacral screw fixation under fluoroscopy guidance has become a popular technique to stabilize unstable pelvic ring fractures. It has been reported that dysmorphic sacrum and S2 screw are risk factors of malposition of screws.　The purpose of the present study was to retrospectively assess if 3D fluoroscopic navigation combined with the preoperative CT-based plan could assist surgeons to perform safe and reliable screw insertion regardless of screw site or sacral morphology.

MATERIAL AND METHODS We treated 27 patients with pelvic ring fracture by percutaneous iliosacral or transsacral screw fixation. In 24 patients, screws were inserted into both S1 and S2, respectively. In three patients, only S1 screw was inserted because screw corridor for S2 screw was quite narrow in preoperative planning. Twelve sacra (44%) were classified as dysmorphic. In order to assess the clinical accuracy of the navigation system, postoperative CT images were matched to preoperative ones, and planned virtual screws were superimposed on the postoperative CT images.

RESULTS One S2 screw for normal sacrum showed grade 1 (< 2mm) perforation. One of two transsacral screws inserted into S1 vertebra with normal sacrum showed grade 2 (2-4 mm) perforation. There was no postoperative complication including nerve palsy. The mean deviation between the planned and the inserted screw position was 3.3±2.3 mm at the entry point of the screw, 2.3±1.4 mm at the area around the nerve root tunnels and 2.5±1.3 mm at the tip of the screw. The accuracy of screw position was not affected by sacral morphology or screw site.

DISCUSSION Percutaneous iliosacral and transsacral screw insertion with the 3D-fluoroscopic navigation combined with CT-based plan was clinical feasible, in which both S1 and S2 screws could be inserted with sufficient safety margin based on a precise preoperatively constructed CT-based plan regardless of sacral morphology.

ABSTRACT. Dental implants are in 96% positioned free hand including a risk of suboptimal place-ment. A miniaturized navigation system has been developed to overcome these limita-tions by fixing two stereo cameras to the drill to guide the surgeon during implant inser-tion. The aim of the present study was to develop and analyse a dual-mode marker with a highly precise, high contrast optical pattern, visible by the camera system, on a pre-cise substrate clearly identifiable on Cone Beam CT (CBCT) images and sterilisable. The marker was developed with a laser engraved optical pattern brought to a ceramic substrate containing two holes to ensure clear identification of the position of the marker within CBCT images. The substrate dimensions were verified with a micrometer gauge and compared to the expected tolerances. The position and angular error of the optical pattern on the sub-strate was analysed on optical microscope images. Six markers were exposed to 20 cleaning and steam sterilization cycles and changes were analysed on optical and scanning electron microscope (SEM) images. Energy dispersive X-ray spectroscopy (EDX) analysis was performed to study the percentage composition of the different ele-ments. Substrate dimensions and angular errors remained always within the defined toleranc-es. Positioning errors were higher than the tolerances for some of the measurements. The error distribution indicates a systematic error. Results were reproducible over the samples. No changes could be observed visually on the optical microscope images between the initial marker and the 20th sterilization cycle. EDX results show slight per-centage variations of the different elements between the different cycles. Fluctuations were most important for the carbon-content. A dual mode marker could be developed with a very accurate optical pattern very pre-cisely positioned on a ceramic substrate making any marker calibration procedure un-necessary. Precision analysis confirmed a reproducible position of the pattern on the substrate, indicating a stable and robust process. Systematic errors can be corrected directly at the used Laser system. Cleaning and sterilization tests confirmed the stability of the dual-mode marker after cleaning and sterilization. The present study has shown that highly accurate, high contrast and robust markers can be manufactured using laser-engraving technology on ceramic substrates.

ABSTRACT. In pelvic fracture surgeries, percutaneous screws must be placed accurately for effective fixation and to prevent damage to surrounding tissue structures. Fluoroscopy is currently used to image the pelvis to provide guidance, but this produces harmful ionising radiation and does not allow three-dimensional (3D) visualisation. Ultrasound offers three-dimensional, non-ionising, real-time, and inexpensive imaging. It has thus emerged as an alternative to fluoroscopy for intraoperative imaging in computer-assisted orthopaedic surgery (CAOS). However, ultrasound-based surgical guidance is challenging because ultrasound produces inherently noisy images with limited field-of-view. While several techniques have been proposed to improve bone clarity in ultrasound scans, there is limited work on enhancing ultrasound’s field-of-view for CAOS. In particular, improving the field-of-view for surgical guidance for pelvic fracture surgeries would be needed to achieve accurate and reliable registration to preoperative data, and accurate screw placement in the pelvis. We propose and evaluate the feasibility of a trackerless method for stitching volumetric ultrasound to achieve an extended field-of-view. Stitching is performed using corresponding features in the overlap between three ultrasound volumes, extracted using an implementation of the 3D scale-invariant feature transform. The volumes are processed using confidence-map weighted phase symmetry detection. Alignment between the volumes is calculated using coherent point drift rigid registration. We succeeded in extending the field-of-view of 3D ultrasound by creating a 39×43×115mm volume from three initial overlapping volumes, with reasonable overall accuracy. We show a mean post-registration surface error of 0.54mm, compared to 0.33mm achieved by previous tracking-based stitching. Our method achieved a mean distance error of 5.1%, compared to 2% in a similar tracked and 3D SIFT-based technique. Our stitching method does not use tracking, thus contributing towards simpler surgical navigation.

ABSTRACT. INTRODUCTION Interlocked intramedullary nailing is the treatment of choice for femoral shaft fractures. However, distal locking is a technically challenging portion of the procedure that can result in distal femoral malrotation and high radiation exposure.

MATERIAL AND METHODS We have tested a robotic procedure for robotic distal locking based on the computation of a drilling trajectory on two calibrated fluoroscopic images. Twenty distal holes were attempted in ten cadaveric femur specimens.

RESULTS Successful screw hole drilling was achieved in the first attempt in each of the ten specimens (20 drill holes in total). No failures were recorded. The average total number of images needed was 6.5 +/- 3.6. The average computation time was 16.5+/- 16.0 seconds.

DISCUSSION Robotic distal locking was feasible in this test and can be integrated into a fully robotic intramedullary nailing procedure.

ABSTRACT. Patient Specific Instrumentation (PSI) may contribute to reduce blood loss related to total knee arthoplasty (TKA) by avoiding violation of the medullary canal. The purpose of the study was to compare the estimated haemoglobin (Hb) and red blood cell volume (RBC) losses in two groups of patients undergoing TKA with PSI and conventional instrumentation. Pre- and post-operative blood samples were collected and drain losses recorded for 22 patients randomly assigned to receive a PSI-assisted or conventional TKA. Post- to preoperative Hb difference was calculated and RBC loss was estimated according to Mercuriali et al. A significant difference in Hb reduction in favour of the PSI group was registered on the last day of stay (p = 0.0084) but not on the previous post-operative days. No significant differences were found in mean RBC loss at any of the post-operative measurements. A significant treatment effect (p = 0.027) on Hb reduction after intervention was found with a regression model for longitudinal measurements. The analyses of transfusions, drain losses, ischemia time and total surgical time revealed no significant differences between the two groups. PSI lead to a significant trend in earlier Hb re-gain, although this does not reflect a significant RBC loss. These promising results suggest a beneficial effect of PSI in blood loss reduction. PSI may hence be considered among the strategies available to control and reduce blood loss related to TKA.

ABSTRACT. We hypothesized that the anterior trochanteric bump (ATB) at the supero-lateral part of the intertrochanteric line which exists on the frontal aspect of the femoral trochanter could be an anatomical landmark to indicate the level of femoral neck osteotomy during hip arthroplasty. The purpose of this study is to clarify the positional relationship between the lesser trochanter and the osteotomy line indicated by the ATB using osteotomy CT simulation. Subjects are femoral computed tomography (CT) images of normal hip or non-collapse osteonecrosis of femoral head. CT images of 84 hip joints from 20 females (n=40 joints) and 27 males (n=44 joints) were used. Using 3D viewer software, the entire femur was estimated by multi planar reconstruction (MPR) image. The most prominent point of the bump at the supero-lateral part of the intertrochanteric line on the frontal aspect of the femoral trochanter was identified. The most depressed point of the proximal lesser trochanter was identified as LT base. On the antero-posterior view of the Digital Reconstructed Radiograph (DRR), similar to femoral neck osteotomy technique, two patterns of neck osteotomy lines were traced that inclined at 30 degrees and 40 degrees to the mediolateral axis from the ATB toward medial cortex of femoral neck. The intersection point of osteotomy line and the medial cortex were plotted as ATB 30 and ATB 40, respectively. On the antero-posterior view of DRR, the distances between ATB 30, ATB 40 and LT base were measured as ATB 30-LT and ATB 40-LT, respectively. ATB 30-LT in female was 15.1 ± 3.9 (6.1-26.0) and that in male was14.9 ± 2.8 (9.3-21.1). ATB 40-LT in female was 9.8 ± 3.7 (1.6-20), and that in male was 8.9 ± 2.9 (3.0-16.1). In 91.7% of the subjects, ATB30-LT was within 10 to 20 mm. In 83.3% of the subjects, ATB40-LT was within 5 to 15 mm. The gender difference in ATB 30-LT, ATB 40-LT, and the distribution of these distances were not significant. In more than 90% of the Japanese subjects, an osteotomy line that inclined at 30 degrees from the ATB in the proximal femoral plane passed through the region of 10 to 20 mm measured from the most depressed point of the proximal lesser trochanter, regardless gender, height and femoral length. It was suggested that ATB can be used as an anatomical landmark to indicate the level of femoral neck osteotomy during hip arthroplasty.

ABSTRACT. Operators sometimes experienced difficulties to obtain acceptable pelvic reference points on bi-planar X-ray images for EOS (EOS imaging, Paris, France) measurements when the patient’s pelvis rotated or inclined. The aim of this study was to clarify the effects of pelvic Z-axis rotation (rotation) and X-axis rotation (inclination) on the measurement values of sacral slope (SS), pelvic tilt (PT), and anterior pelvic plane (APP) angle by using of EOS system. We obtained anteroposterior and lateral X-ray images of a trunk phantom, which was rotated and inclined in standing position, by using of EOS system. SS and PT (parameters to judge patient’s pelvic alignments for spinal surgeons), and APP angle (reference to decide patient’s pelvic alignments for hip surgeons) were measured with the EOS software. We compared SS, PT and APP angle when the phantom was rotated and inclined. PT showed small changes when the phantom was rotated, whereas SS decreased and APP angle increased. The values of SS and PT dispersed at 30⁰ rotation among each inclination. APP angle dispersed at -15º and 30º of rotation among each inclination. Measurement variations of SS and PT were small amount under 15 degrees of rotation and inclination. On the contrary, APP angles showed greater variations when the phantom was rotated and inclined. Proper positioning of subjects, by reducing rotational effects, is important to get constant reproducible measurements, even if we use EOS system.

ABSTRACT. Total knee arthroplasty (TKA) typically addresses end-stage osteoarthritis. While TKA procedures have demonstrated clinical success, occasionally intraoperative complications can occur. This cadaver study examined the potential benefits of soft tissue protection in robotic-arm assisted TKA (RATKA). Six cadaver knees were prepared using RATKA by a single surgeon from a high-volume TKA joint center with no former clinical robotic experience, and compared to seven manually performed cases as a control. The presence of soft tissue disruption was assessed by having the surgeon perform visual evaluation and palpation of the medial collateral ligament (MCL), lateral collateral ligament (LCL), posterior cruciate ligament (PCL), and the patellar ligament after the procedures. The amount of tibial subluxation and patellar eversion was recorded for each case. For all RATKA cases, there was no visible evidence of disruption of any of the ligaments. All RATKA cases were successfully left with a bone island on the tibial plateau, which protected the PCL. Tibial subluxation and patellar eversion were not required for visualization. In two of the seven MTKA cases, there was slight disruption noted of the PCL, although this did not lead to any apparent change in the functional integrity of the ligament. All MTKA cases required tibial subluxation and patellar eversion to achieve optimal visualization. Aspects of soft tissue protection were noted in this cadaver study for RATKA. Standard retraction techniques during cutting are recommended. This is the first study to have soft tissue injury parameters assessed for RATKA, and may serve as a platform for future studies.

ABSTRACT. High tibial osteotomy was done using patient-specific instruments (PSI) to remove isolated chondral defects in the knee joint. The aim of using PSI for osteotomy was to predetermine the force and the contact pressure before surgery and to estimate postoperative limb alignment which must mimic the physiologic loading to secure favorable clinical outcome and cartilage repair. This regenerative process began immediately after correcting the angular deformity and improving the alignment. Range of movement and type/degree of deformity were recorded along with knee scores.

ABSTRACT. Abstract Bone allograft after tumor resection exists since the 1940s in orthopedic reconstructive surgery; allograft geometric selection process was performed using X-rays. Using advanced virtual simulators we can detect accurately the best allografts. Measuring the tumor margin and determining a planar configuration in both patient and allograft, we establish a positive - negative model. Using a surgical navigator we execute the plan carefully, allowing the surgeon to prepare the allograft, negative mold, and simultaneously other surgeons perform the surgical approach and resects the tumor, positive mold. Preoperative virtual planning and surgical navigation are potentially useful tools for reconstructive oncologic orthopedic surgery.

ABSTRACT. Introduction: Augmented glenoid implants are a novel solution for significant posterior bone loss that occurs in patients with glenohumeral osteoarthritis. These novel implant designs have limited evidence of their performance or biomechanical characteristics over time. The aim of the study was to evaluate the effect of glenohumeral mismatch for two commercially available augmented glenoid designs (wedged and stepped) to predict implant loosening and failure. Materials and Methods: A constant axial compressive load of 750 N was applied to the model and an increasing vertical (superior) shearing component was introduced linearly to the point of humeral subluxation. The different glenohumeral radial mismatches at 3.5 and 10 mm were modeled and simulated. The force ratio, implant micromotion, implant stress and cement stress were analyzed and compared between two designs during posterior and superior translations for a range of glenohumeral curvatures.

Results: The force ratio, an indicator of joint stability, was higher in the superior and inferior vectors compared to the anterior and posterior vectors for both designs at regardless of amount of radial mismatch. There was an inverse relationship between the force ratio and the amount of radial mismatch for each design. The force ratio was higher for the stepped design compared to the wedged design in all four vectors for both radial mismatches. The wedged model had a lower constraint angle (35.3 degrees) compared to the stepped model (45.3 degrees). The stepped implant experienced less backside shear stress compared to the wedged design. Interestingly, the glenoid distraction was 3-20 magnitudes higher for the wedged model in all four vectors. Both designs had the highest distraction in the center to posterior vector. Discussion: For both radial mismatches the stepped design had the highest values which indicated that it was a more stable joint. Our study showed that the stepped design experienced less back side shear stress compared to wedged design in both mismatches. Interestingly, the amount of shear stress for the stepped design was directly proportional to the amount of radial mismatch. Whereas the wedged design had an inverse relationship. Overall, the glenoid compression and distraction values indicate that the wedged model experiences more micromotion in all four vectors, which may lead to implant loosening.

ABSTRACT. Introduction Scaphoid non-union results the typical humpback deformity, pronation of the distal fragment, and a bone defect in the non-union site with shortening. Bone grafting, whether open or arthroscopic, relies on fluoroscopic and direct visual assessment of reduction. However, because of the bone defect and irregular geometry, it is difficult to determine the precise width of the bone gap and restore the original bone length, and to correct interfragmentary rotation. Correction of alignment can be performed by computer-assisted planning and intraoperative guidance. The use of computer navigation in guiding reduction in scaphoid non-unions and displaced fractures have not been reported. Objective We propose a method of anatomical reconstruction in scaphoid non-union by computer-assisted preoperative planning combined with intraoperative computer navigation. This could be done in conjunction with a minimally invasive, arthroscopic bone grafting technique. Methods A model consisting of a scaphoid bone with a simulated fracture, a forearm model, and an attached patient tracker was used. 2 titanium K-wires were inserted into the distal scaphoid fragment. 3D images were acquired and matched to those from a computed tomography (CT) scan. In an image processing software, the non-union was reduced and pin tracts were planned into the proximal fragment. The K-wires were driven into the proximal fragment under computer navigation. Reduction was assessed by direct measurement. These steps were repeated in a cadaveric upper limb. A scaphoid fracture was created and a patient tracker was inserted into the radial shaft. A post-fixation CT was obtained to assess reduction. Results and Discussion In both models, satisfactory alignment was obtained. There were minimal displacement and articular stepping, and scaphoid length was restored with less than 1mm discrepancy. This study demonstrated that an accurate reduction of the scaphoid in non-unions and displaced fractures can be accurately performed using computed navigation and computer-assisted planning. It is the first report on the use of computer navigation in correction of alignment in the wrist.

ABSTRACT. Achieving precise open reduction and fixation of acetabular fractures by using a plate osteosynthesis is a complex procedure. Increasing availability of affordable 3D printing devices and services now allow to actually print physical models of the patient's anatomy by segmenting the patient’s CT image. The data processing and printing of the model however still take too much time and usually the resulting model is rigid and doesn't allow fracture reduction on the model itself. Our proposed solution automatically detects relevant structures such as the fracture gaps and cortical bone while eliminating irrelevant structures such as debris and cancellous bone. This is done by approximating a sphere to the exterior surface of a classic segmented STL model. Stepwise, these approximated vertices are projected deeper into any structure such as the acetabular socket or fractures, following a specific set of rules. The resulting surface model finally is adapted precisely to the primary segmented model. Creating an enhanced surface reconstruction model from the primary model took a median time of 42 sec. The whole workflow from DICOM to enhanced printable 3D file took a median time of 13:25 min. The median time and material needed for the prints without the process was 32:25:36 h and 241,04 g, with the process 09:41:33 h and 65,89 g, which is 70% faster. The price of material was very low with a median of 2,18€ per case. Moreover, fracture reduction becomes possible, allowing a dry-run of the procedure and allowing more precise plate placement. Pre-contouring of osteosynthesis plates by using these 3D printouts was done for eleven patients prior to surgery. These printouts were validated to be accurate by three experiences surgeons and compared to classic segmented models regarding printing time, material cost and reduction ability. The pre-contouring of the plates was safely achievable. Our results show that improving the operative treatment with the help of enhanced 3D printed fracture models seems feasible and needs comparably little time and cost, thus making it a technique that can easily integrated into the clinical workflow.

ABSTRACT. The treatment of joint-fractures is a common task in orthopaedic surgery causing considerable health costs and patient disabilities. Percutaneous techniques have been developed to mitigate the problems related to open surgery (e.g. soft tissue damage), although their application to joint-fractures is limited by the sub-optimal intra-operative imaging (2D-fluoroscopy) and by the high forces involved. Our earlier research toward improving percutaneous reduction of intra-articular fractures has resulted in the creation of a robotic system prototype, i.e. RAFS (Robot-Assisted Fracture Surgery) system. We propose a robot-bone attachment device for percutaneous bone manipulation, which can be anchored to the bone fragment through one small incision, ensuring the required stability and reducing the “biological cost” of the procedure. It consists of a custom-designed orthopaedic pin, an anchoring system (AS secures the pin to the bone), and a gripping system (GS connects the pin and the robot). This configuration ensures that the force/torque applied by the robot is fully transferred to the bone fragment to achieve the desired anatomical reduction. The device has been evaluated through the reduction of 9 distal femur fractures on human cadavers using the RAFS system. The devices allowed the reduction of 7 fractures with clinical acceptable accuracy. 2 fractures were not reduced: in one case the GS failed and was not able to keep the pin stationary inside the robot (pin rotates inside the GS). The other fracture was too dislocated (beyond the operational workspace capability of the robot). A more stable GS will be designed to avoid displacements between the pin and the robot.

ABSTRACT. With the advancement of the virtual technologies, three-dimensional surgical simulators are now possible. In this article, we describe an immersive simulation platform, allowing students in orthopedic surgery to learn how to deal with a sample diaphyseal fracture of the femur using LC-DCP plate hole, cortical screw and verbrugge forceps. To reach certain realism, weight of the objects and force feedback are used in addition to the visual scene and the 3D sound. The students feel the weight, the strength of the bone when they pierce the holes, and the vibration of the drill. The simulation is implemented by using CAVE, the CyberGlove, CyberGrasp, and 3D sound system.

ABSTRACT. Several preoperative planning tools in computer-assisted surgery in acetabular fractures have been proposed. Moreover, all these preoperative planning tools are based on geometrical repositioning with their own limitations. The aim of this study was to evaluate the value of our prototype virtual planning tool using a rigid biomechanical model to predict failure in fracture reduction. Between November of 2015 and June of 2016, 10 patients were operated by the main author for acetabular fracture in our institution. To validate our biomechanical model planning tool, biomechanical simulation was performed for each patient immediately after the surgery. Reduction quality was assessed on post-operative CT scans. A 3D model of the acetabular fracture was build out of the CT images using the non-commercial software Itksnap. Then a biomechanical model implemented within the non-commercial Artisynth framework was used to perform virtual reduction. Surgical approach and surgical strategy according to the operative report were simulated. The simulated reductions and the surgical reductions were compared. The same reductions were obtained during surgery and biomechanical simulation in the 10 cases. For 7 cases, reduction was achieved by anterior surgical approach and so was the simulation. For 3 cases, reduction was achieved by posterior surgical approach and so was the simulation. The biomechanical simulation found similar results using the same surgical strategy with 9 anatomical reductions (90%) and one imperfect reduction (10%). The mean duration to perform acetabular planning surgery was 24 +/- 9 min [16-38]. Our virtual planning tool using a rigid biomechanical model can predict success or failure in fracture reduction according to the surgical approach and the surgical strategy.