Gait analysis systems have enjoyed increasing usage and have been validated to provide highly accurate assessments for range of motion. Size, cost, need for marker placement and need for complex data processing have remained limiting factors in uptake outside of what remains predominantly large research institutions. Progress and advances in deep neural networks, trained on millions of clinically labelled datasets, have allowed the development of a computer vision system which enables assessment using a handheld smartphone with no markers and accurate range of motion for knee during flexion and extension. This allows clinicians and therapists to objectively track progress without the need for complex and expensive equipment or time-consuming analysis, which was concluded to be lacking during a recent systematic review of existing applications. A smartphone based computer vision system was assessed for accuracy with a gold standard comparison using a validated ‘traditional’ infra-red motion capture system which had a defined calibrated accuracy of 0.1degrees. A total of 22 subjects were assessed simultaneously using both the computer vision smartphone application and the standard motion capture system. Assessment of the handheld system was made by comparison to the motion capture system for knee flexion and extension angles through a range of motion with a simulated fixed-flexion deformity which prevented full extension to assess the accuracy of the system, repeating movements ten times. The peak extension angles and also numerous discrete angle measurements were compared between the two systems. Repeatability was assessed by comparing several sequential cycles of flexion/extension and comparison of the maximum range of motion in normal knees and in those with a simulated fixed-flexion deformity. In addition, discrete angles were also measured on both legs of three cadavers with both skin and then bone implanted fiducial markers for ground truth reliability accounting for skin movement. Data was processed quickly through an automated secure cloud system.Introduction
Method
The accurate positioning of the total knee arthroplasty affects the survival of the implants(1). Alignment of the femoral component in relation to the native knee is best determined using pre- and post-operative 3D-CT reconstruction(2). Currently, the scans are visualised on separate displays. There is a high inter- and intra-observer variability in measurements of implant rotation and translation(3). Correct alignment is required to allow a direct comparison of the pre- and post-operative surfaces. This is prevented by the presence of the prostheses, the bone shape alteration around the implant, associated metal artefacts, and possibly a segmentation noise. Create a novel method to automatically register pre- and post-operative femora for the direct comparison of the implant and the native bone.Background
Aim
Trochlear geometry of modern femoral implants is designed for the mechanical alignment (MA) technique for Total Knee Arthroplasty (TKA). The biomechanical goal is to create a proximalised and more valgus trochlea to better capture the patella and optimize tracking. In contrast, Kinematic alignment (KA) technique for TKA respects the integrity of the soft tissue envelope and therefore aims to restore native articular surfaces, either femoro-tibial or femoro-patellar. Consequently, it is possible that current implant designs are not suitable for restoring patient specific trochlea anatomy when they are implanted using the kinematic technique. This could cause patellar complications, either anterior knee pain, instability or accelerated wear or loosening. The aim of our study is therefore to explore the extent to which native trochlear geometry is restored when the Persona® implant (Zimmer, Warsaw, USA) is kinematically aligned. A retrospective study of a cohort of 15 patients with KA-TKA was performed with the Persona® prosthesis (Zimmer, Warsaw, USA). Preoperative knee MRIs and postoperative knee CTs were segmented to create 3D femoral models. MRI and CT segmentation used Materialise Mimics® and Acrobot Modeller® software, respectively. Persona® implants were laser-scanned to generate 3D implant models. Those implant models have been overlaid on the 3D femoral implant model (generated via segmentation of postoperative CTs) to replicate, BACKGROUND
METHODS
The mechanical alignment (MA) for Total Knee Arthroplasty (TKA) with neutral alignment goal has had good overall long-term outcomes. In spite of improvements in implant designs and surgical tools aiming for better accuracy and reproducibility of surgical technique, functional outcomes of MA TKA have remained insufficient. Therefore, alternative, more anatomicaloptions restoring part (adjusted MA (aMA) and adjusted kinematic alignment (aKA) techniques) or the entire constitutional frontal deformity (unicompartment knee arthroplasty (UKA) and kinematic alignment (KA) techniques) have been developed, with promising results. The kinematic alignment for TKA is a new and attractive surgical technique enabling a patient specific treatment. The growing evidence of the kinematic alignment mid-term effectiveness, safety and potential short falls are discussed in this paper. The current review describes the rationale and the evidence behind different surgical options for knee replacement, including current concepts in alignment in TKA. We also introduce two new classification systems for “implant alignments options” (Figure 1) and “osteoarthritic knees” (Figure 2) that would help surgeons to select the best surgical option for each patient. This would also be valuable for comparison between techniques in future research.
Surgical planning of long bone surgery often takes place using outdated 2D axes on 2D images such as long leg standing X-rays. This leads to errors and great variation between intra- and inter- observers due to differing frames of reference. With the advent of 3D planning software, researchers developed 3D axes of the knee such as the Flexion Facet Axis (FFAx) and Trochlear Axis (TrAx), and these proved easy to derive and reliable. Unlike 2D axes, clinicians and scientists can use a single 3D axis to obtain measurements relative to other 3D axes, in all three planes Deriving a 3D axis also does not require an initial frame of reference, such as in trying to derive the 2D Posterior Condylar Axis (PCAx), whereby a slight change in slice orientation will affect its position. However, there is no 3D axis derived for the tibial plateau yet. Measurements of tibial joint line obliquity are with a 2D axis drawn on AP long leg standing X-rays. The same applies to tibial plateau rotation, as measured by 2D axes drawn on axial CT/MRI slices. this study aimed to to develop a novel new 3D axis for the tibial plateau to quantify both tibial plateau joint line obliquity and axial rotation. Materialise software version 8.0 (Materialise Inc., Belgium) handled segmentation of CT data and for analysis of bony morphology. A line joining the centroids of the medial and lateral tibial plateaus formed the TCAx (Fig1). A line joining the middle coordinate of the TCAx, to the centre of the best-fit sphere between the medial and lateral malleolus formed the Tibial Mechanical Axis (TMAx). A standard frame of reference aligned 72 tibias with the TCAx horizontal in the axial view, and the TMAx aligned parallel to the global reference coordinate system vertical axis. Tibial joint line obliquity was the angle between the TCAx and TMAx on the medial side, also known as the Medial Tibial Plateau Angle (MPTA)(Fig2). The authors compared reliability and accuracy of the TCAx against three other rotational axes of the tibia as described in the literature.Background
Methods
Constitutional knee varus increases the risk of medial OA disease due to increase in the knee adduction moment and shifting of the mechanical axis medially. Hueter-Volkmann's law states that the amount of load experienced by the growth plate during development influences the bone morphology. For this reason, heightened sports activity during growth is associated with constitutional varus due to added knee adduction moment. In early OA, X-rays often show a flattened medial femoral condyle extension facet (EF). However, it is unknown whether this is a result of osteoarthritic wear, creep deformation over decades of use, or an outcome of Hueter-Volkmann's law during development. A larger and flattened medial EF can bear more weight, due to increased load distribution. However, a flattened EF may also extrude the meniscus, leading meniscus degeneration and joint failure. Therefore, this study aimed to investigate whether varus knees have flattened medial EFs of both femur and tibia in a cohort of patients with no signs yet of bony attrition. Segmentation and morphology analysis was conducted using Materialise software (version 8.0, Materialise Inc., Belgium). This study excluded knees with bony attrition of the EFs based on Ahlbäck criteria, intraoperative findings, and operation notes history. Standard reference frames were used for both the femur and tibia to ensure reliable and repeatable measurements. The hip-knee-angle (HKA) angle defined varus or valgus knee alignment. Femur: The femoral EFs and flexion facets (FFs) had best-fit spheres fitted with 6 repetitions. (Fig1) Tibia: The slopes of the antero-medial medial tibial plateau were approximated using lines. (fig2)Background
Methods
Conventional TKA surgery attempts to restore patients to a neutral alignment, and devices are designed with this in mind. Neutral alignment may not be natural for many patients, and may cause dissatisfaction [1]. To solve this, kinematical alignment (KA) attempts to restore the native pre-arthritic joint-line of the knee, with the goal of improving knee kinematics and therefore patient's function and satisfaction [1]. Proper prosthetic trochlea alignment is important to prevent patella complications such as instability or loosening. However, available TKA components have been designed for mechanical implantation, and concerns remain relating the orientation of the prosthetic trochlea when implants are kinematically positioned. The goal of this study is to investigate how a currently available femoral component restores the native trochlear geometry of healthy knees when virtually placed in kinematic alignment. The healthy knee OAI (Osteoarthritis Initiative) MRI dataset was used. 36 MRI scans of healthy knees were segmented to produce models of the bone and cartilage surfaces of the distal femur. A set of commercially available femoral components was laser scanned. Custom 3D planning software aligned these components with the anatomical models: distal and posterior condyle surfaces of implants were coincident with distal and posterior condyle surfaces of the cartilage; the anterior flange of the implant sat on the anterior cortex; the largest implant that fitted with minimal overhang was used, performing ‘virtual surgery’ on healthy subjects. Software developed in-house fitted circles to the deepest points in the trochlear grooves of the implant and the cartilage. The centre of the cartilage trochlear circle was found and planes, rotated from horizontal (0%, approximately cutting through the proximal trochlea) through to vertical (100%, cutting through the distal trochlea) rotated around this, with the axis of rotation parallel to the flexion facet axis. These planes cut through the trochlea allowing comparison of cartilage and implant surfaces at 1 degree increments - (fig.1). Trochlear groove geometry was quantified with (1) groove radial distance from centre of rotation cylinder (2) medial facet radial distance (3) lateral facet radial distance and (4) sulcus angle, along the length of the trochlea. Data were normalised to the mean trochlear radius. The orientation of the groove was measured in the coronal and axial plane relative to the flexion facet axis. Inter- and intra-observer reliability was measured.BACKGROUND
METHODS
Patellofemoral joint (PFJ) arthroplasty is traditionally performed using mechanical jigs to align the components, and it is hard to fine tune implant placement for the individual patient. These replacements have not had the same success rate as other forms of total or partial knee replacement surgery1. Our team have developed a computer assisted planning tool that allows alignment of the implant based on measurements of the patient's anatomy from MRI data with the aim of improving the success of patellofemoral joint arthroplasty. When planning a patellofemoral joint arthroplasty, one must start from the premise that the original joint is either damaged as a result of osteoarthritis, or is dysplastic in some way, deviating from a normal joint. The research aimed to plan PFJ arthroplasty using knowledge of the relationship between a normal PFJ (trochlear groove, trochlea axis and articular surfaces) and other aspects of the knee2, allowing the plan to be estimated from unaffected bone surfaces, within the constraints of the available trochlea. In order to establish a patient specific trochlea model a method was developed to automatically compute an average shape of the distal femur from normal distal femur STL files (Fig.1). For that MRI scans of 50 normal knees from osteoarthritis initiative (OAI) study were used. Mimics and 3-matic software (Materialise) packages were used for segmentation and analysis of 3D models. Spheres were fitted to the medial and lateral flexion facets for both average knee model and patient knee model. The average knee was rescaled and registered in order to match flexion facet axis (FFA) distance and FFA midpoint of the patient (Fig.2). The difference between the patient surface and the average knee surface allow to plan the patella groove alteration. The Patella cut is planned parallel to the plane fitted to the anterior surface of the patella. The patella width/thickness ratio (W/T=2) is used to predict the post reconstruction thickness3. The position of the patella component (and its orientation if a component with a median ridge is used) is also planned. The plan is next fine-tuned to achieve satisfactory PFJ kinematics4 (Fig.3). This will be complemented by intraoperative PFJ tracking which assists with soft tissue releases. PFJ kinematics is evaluated in terms of patella shift, tilt and deviation from the previously described circular path of the centre of the patella. The effect of preoperative planning on PFJ tracking and soft tissue releases is being examined. Additional study is needed to evaluate whether planning and intraoperative kinematic measurements improve the clinical outcome of PFJ arthroplasty.
The trochlea of a typical patellofemoral replacement or anterior flange of a total knee replacement usually extends past the natural trochlea and continues onto the femoral anterior cortex. One reason for this is that it allows a simple patella button to be permanently engaged in the trochlea groove in an attempt to ensure stability. On the natural patella, the apex helps to guide it into the trochlea groove as the knee moves from full extension into flexion. The aim is to study whether a generalised patella can be created that is close in form to a healthy patella. MRI scans were taken of 30 patellae. Characteristics of these patellae (height, width, thickness, apex angle) were measured. The apex angle was found to be similar between patellae (mean=126 degrees, sd = 8.8), as were the ratios between height and width (mean width/height = 1.05, sd = 0.07) and between thickness and width (mean width/thickness = 1.8, sd = 0.19). These patellae were then segmented to create a surface including cartilage, resulting in 30 STL (stereolithography) files in which the surfaces are represented by triangle meshes. To design the average patella the individual patellae were aligned to a standard frame of reference by placing a set of landmarks on the proximal/distal, medial/lateral and anterior/posterior extents of each (fig.1). The vertical axis was defined as passing parallel to the proximal/distal points and the horizontal as passing parallel to the medial/lateral points when looking along the computed vertical axis. The origin centre of the frame of reference was chosen to be mid-way between these points. The mean width was then computed and each patella scaled linearly around the origin to give them all equal width. All the aligned patellae were then averaged together to provide a composite cartilaginous patella. The averaging process was achieved by taking one patella as a seed. The patella chosen for seed was that whose parameters were closest to the average width, height and thickness. An approximately normal vector was passed a point ‘P’ on the seeds, and the points at which these intersected the other models were then determined. The closest intersection point to ‘P’ on each model was chosen and these averaged together. ‘P’ is then replaced in the model with this average point. The averaging process then continues with all the remaining points on the seed model in the same manner to build the average models.Introduction
Method
Anatomical referencing, component positioning, limb alignments and correction of mechanical axes are essential first steps in successful computer assisted navigation. However, apart from basic gap balancing and quantification of ranges of motion, routine navigation technique usually fails to use the full potential of the registered information. Enhanced dynamic assessment using an upgraded navigation system (Brainlab V. 2.2) is now capable of producing enhanced ‘range of motion’ analysis, ‘tracking curves’ and ‘contact point observations’. ‘Range of motion analysis’ was performed simultaneously for both tibio-femoral and patella-femoral joints. Other dynamic information including epicondylar axis motion, valgus and varus alignments, antero-posterior tibio-femoral shifts, as well as flexion and extension gaps were simultaneously stored as a series of ‘tracking curves’ throughout a full range of motion. Simultaneous tracking values for both tibiofemoral and patellofemoral motion was also obtained after performing registration of the prosthetic trochlea. However, there seems to be little point in carrying out such observations without fully assessing joint stability by applying controlled force to the prosthetic joint. Therefore, in order to fully assess ‘potential envelopes of motion’, observations have been made using a set of standardised simple dynamic tests during insertion and after final positioning of trial components. Also, such tests have been carried out before and after any necessary ligament balancing. Firstly, the lower leg was placed in neutral alignment and the knee put through a flexion-extension cycle. Secondly the test was repeated but with the lower leg being placed into varus and internal rotation. The third test was performed with the lower leg in valgus and external rotation. Force applied was up to the point where resistance occurred without any gross elastic deformation of capsule or ligament in a manner typical of any surgeon assessing the stability of the construct. Also a passive technique of using gravity to ‘Drop-Test’ the limb into flexion and extension gave useful information regarding potential problems such as blocks to extension, over-stuffing of the extensor mechanism and tightness of the flexion gap. All the definitive tests were performed after temporary medial capsular closure. Ten total knee arthroplasties have been studied using this technique with particular reference to the patterns of instability found before, during and after adjustments to component positioning and ligament balancing. Marked intra-operative variation in the stability characteristics of the trial implanted joints has been quantified before correction. These corrections have been analysed in terms of change in translations, rotations and contact points induced by any such adjustments to components and ligament. Certain major typical patterns of instability have begun to be identified including excessive rotational and translational movements. Instability to valgus and external rotational stress was found in two cases and to varus and internal rotational stress in one case before correction. In particular, surprising amounts of edge loading in mid-flexion under stress testing has been identified and corrective measures carried out. Reductions in paradoxical tibio-femoral antero-posterior motion were also observed. Global instability and conversely tightness were also observed in early stages of surgery. Adjustments to component sizes, rotations, tibial slope angles and insert thickness were found to be necessary to optimise range of motion and stability characterisitics on an almost case-by-case basis. Two cases were identified where use of more congruent or stabilised components was necessary. Observation of quite marked loss of contact between tibia and femur was seen on the lateral side of the knee in deep flexion in several cases. Patellar tracking was also being observed during such dynamic tests and in two cases staged partial lateral retinacular releases were carried out to centre patellar tracking on the prosthetic trochlea. Although numbers in this case series are small, it has been possible to begin to observe, classify and quantify patterns of instability intra-operatively using simple stress tests. Such enhanced intra-operative information may in future make it possible to create algorithms for logical and precise adjustments to ligaments and components in order to optimise range of motion, contact areas and stability in TKR.
