Abstract. Objectives. The fidelity of a 3D model created using image segmentation must be precisely quantified and evaluated for the model to be trusted for use in subsequent biomechanical studies such as finite element analysis. The bones within the ankle joint vary significantly in size and shape. The purpose of this study was to test the hypothesis that the accuracy and reliability of a segmented bone geometry is independent of the particular bone being measured. Methods. Computed tomography (CT) scan data (slice thickness 1 mm, pixel size 808±7 µm) from three anonymous patients was used for the development of the ankle geometries (consisting of the tibia, fibula, talus, calcaneus, and navicular bones) using Simpleware Scan IP software (Synopsys, Exeter, UK). Each CT scan was segmented 4 times by an inexperienced undergraduate, resulting in a total of 12 geometry assemblies. An experienced researcher segmented each scan once, and this was used as the ‘gold standard’ to quantify the accuracy. The solid bone geometries were imported into CAD software (Inventor 2023, Autodesk, CA, USA) for measurement of the surface area and volume of each bone, and the distances between bones (tibia to talus, talus to navicular, talus to calcaneus, and tibia to fibula) were carried out. The intra-class coefficient (ICC) was used to assess intra-observer reliability. Bland Altman plots were employed as a statistical measure for criteria validity (accuracy) [1]. Results. The average ICC score was 0.93, which is regarded as a high reliability score for an inexperienced user. The talus to navicular and talus to tibia separations, which had the smallest distances, showed a slight decrease in reliability and this was observed for all separations shorter than 2 mm. According to the Bland-Altman plots, more than 95% of the data points were inside the borders of agreement, which is an excellent indication of accuracy. The bias percentage (average
Introduction. Precision
Polyethylene wear of joint replacements can cause severe clinical complications, including; osteolysis, implant loosening, inflammation and pain. Wear simulator testing is often used to assess new designs, but it is expensive and time consuming. It is possible to predict the volume of polyethylene implant wear from finite element models using a modification of Archard's classic wear law [1–2]. Typically, linear elastic isotropic, or elasto-plastic material models are used to represent the polyethylene. The purpose of this study was to investigate whether use of a viscoelastic material model would significantly alter the predicted volumetric wear of a mobile-bearing unicompartmental knee replacement. Tensile creep-recovery experiments were performed to characterise the creep and relaxation behaviour of the polyethylene (moulded GUR 4150 samples machined to 180×20×1 mm). Samples were loaded to 3 MPa stress in 4 minutes, and then held for 6 hours, the tensile stress was removed and samples were left to relax for 6 hours. The mechanical test data was used fit to a validated three–dimensional fractional Maxwell viscoelastic constitutive material model [3]. An explicit finite element model of a mobile–bearing unicompartmental knee replacement was created, which has been described previously [4]. The medial knee replacement was loaded to 1200 N over a period of 0.2 s. The bearing was meshed using quadratic tetrahedral elements (1.5 mm seeding size based on results of a mesh convergence study), and the femoral component was represented as an analytical rigid body. Wear predictions were made from the contact stress and sliding distance using Archard's law, as has been described in the literature [1–2]. A wear factor of 5.24×10−11 was used based upon the work by Netter et al. [2]. All models were created and solved using ABAQUS finite element software (version 6.14, Simulia, Dassault Systemes). The fractional viscoelastic material model predicted almost twice as much wear (0.119 mm3/million cycles) compared to the elasto-plastic model (0.069 mm3/million cycles). The higher wear prediction was due to both an increased sliding distance and higher contact pressures in the viscoelastic model. These preliminary findings indicate the simplified elasto-plastic polyethylene material representation can underestimate wear predictions from numerical simulations. Polyethylene is known to be a viscoelastic material which undergoes creep clinically, and it is not surprising that it is necessary to represent that viscoelastic behaviour to accurately predict implant wear. However, it does increase the complexity and run time of such computational studies, which may be prohibitive.
Currently implemented accuracy metrics in open-source libraries for segmentation by supervised machine learning are typically one-dimensional scores [1]. While extremely relevant to evaluate applicability in clinics, anatomical location of segmentation
Current evidence suggests that superior surgical team performance is linked to fewer intra-operative
Evidence supporting the use of virtual reality (VR) training in orthopaedic procedures is rapidly growing. However, the impact of the timing of delivery of this training is yet to be tested. We aimed to investigate whether spaced VR training is more effective than massed VR training. 24 medical students with no hip arthroplasty experience were randomised to learning the direct anterior approach total hip arthroplasty using the same VR simulation, training either once-weekly or once-daily for four sessions. Participants underwent a baseline physical world assessment on a saw bone pelvis. The VR program recorded procedural
Gait analysis is an indispensable tool for scientific assessment and treatment of individuals whose ability to walk is impaired. The high cost of installation and operation are a major limitation for wide-spread use in clinical routine. Advances in Artificial Intelligence (AI) could significantly reduce the required instrumentation. A mobile phone could be all equipment necessary for 3D gait analysis. MediaPipe Pose provided by Google Research is such a Machine Learning approach for human body tracking from monocular RGB video frames that is detecting 3D-landmarks of the human body. Aim of this study was to analyze the accuracy of gait phase detection based on the joint landmarks identified by the AI system. Motion data from 10 healthy volunteers walking on a treadmill with a fixed speed of 4.5km/h (Callis, Sprintex, Germany) was sampled with a mobile phone (iPhone SE 2nd Generation, Apple). The video was processed with Mediapipe Pose (Version 0.9.1.0) using custom python software. Gait phases (Initial Contact - IC and Toe Off - TO) were detected from the angular velocities of the lower legs. For the determination of ground truth, the movement was simultaneously recorded with the AS-200 System (LaiTronic GmbH, Innsbruck, Austria). The number of detected strides, the
This study aims to create a novel computational workflow for frontal plane laxity evaluation which combines a rigid body knee joint model with a non-linear implicit finite-element model wherein collateral ligaments are anisotropically modelled using subject-specific, experimentally calibrated Holzpfel-Gasser-Ogden (HGO) models. The framework was developed based on CT and MRI data of three cadaveric post-TKA knees. Bones were segmented from CT-scans and modelled as rigid bodies in a multibody dynamics simulation software (MSC Adams/view, MSC Software, USA). Medial collateral and lateral collateral ligaments were segmented based on MRI-scans and are modelled as finite elements using the HGO model in Abaqus (Simulia, USA). All specimens were submitted varus/valgus loading (0-10Nm) while being rigidly fixed on a testing bench to prevent knee flexion. In subsequent computer simulations of the experimental testing, rigid bodies kinematics and the associated soft-tissue force response were computed at each time step. Ligament properties were optimised using a gradient descent approach by minimising the
Introduction. Three-dimensional (3D) morphological understanding of the hip joint, specifically the joint space and surrounding anatomy, including the proximal femur and the pelvis bone, is crucial for a range of orthopedic diagnoses and surgical planning. While deep learning algorithms can provide higher accuracy for segmenting bony structures, delineating hip joint space formed by cartilage layers is often left for subjective manual evaluation. This study compared the performance of two state-of-the-art 3D deep learning architectures (3D UNET and 3D UNETR) for automated segmentation of proximal femur bone, pelvis bone, and hip joint space with single and multi-class label segmentation strategies. Method. A dataset of 56 3D CT images covering the hip joint was used for the study. Two bones and hip joint space were manually segmented for training and evaluation. Deep learning models were trained and evaluated for a single-class approach for each label (proximal femur, pelvis, and the joint space) separately, and for a multi-class approach to segment all three labels simultaneously. A consistent training configuration of hyperparameters was used across all models by implementing the AdamW optimizer and Dice Loss as the primary loss function. Dice score, Root Mean Squared
Abstract. Objectives. Dual mobility (DM) hip implants whereby the polyethylene liner is “free-floating” are being used increasingly clinically. The motion of the liner is not well understood and this may provide insight into failure mechanisms; however, there are no published methods on tracking liner motion while testing under clinically relevant conditions. The aim was to develop and evaluate a bespoke inertial tracking system for DM implants that could operate submerged in lubricant without line-of-sight and provide 3D orientation information. Methods. Trackers (n=5) adhered to DM liners were evaluated using a robotic arm and a six-degree of freedom anatomical hip simulator. Before each set of testing the onboard sensor suites were calibrated to account for steady-state and non-linearity
Abstract. Objectives. Non-technical skills including teamwork play a pivotal role in surgical outcomes. Virtual reality is effective at improving technical skills, however there is a paucity of evidence on team-based virtual reality (VR) training. This study aimed to assess if multiplayer virtual reality training was superior to solo training for acquisition of both technical and non-technical skills in learning the complex anterior approach total hip arthroplasty operation. Methods. 10 novice surgeons and 10 novice scrub nurses, were randomised to solo or team virtual reality training to perform anterior approach total hip arthroplasty. Solo participants trained with virtual avatar counterparts, whilst teams trained in pairs (surgeon and scrub nurse). Both groups underwent 5 VR training sessions over 6 weeks. Then, they underwent a real-life assessment in which they performed AA-THA on a high-fidelity model with real equipment in a simulated operating theatre. Teams performed together and solo participants were randomly paired up with a solo player of the opposite role. Videos of the assessment were marked by two blinded expert assessors. Outcomes were procedure time, procedural
Abstract. Objectives. Evidence supporting the use of immersive virtual reality (iVR) training in orthopaedic procedures is rapidly growing. However, the impact of the timing of delivery of this training is yet to be tested. This study investigated whether spaced iVR training is more effective than massed iVR training for novices learning hip arthroplasty. Methods. 24 medical students with no hip arthroplasty experience were randomised to learning total hip arthroplasty using the same iVR simulation training either once-weekly or once-daily for four sessions. Participants underwent a baseline physical world assessment to orientate an acetabular component on a saw bone pelvis, and a baseline knowledge test. In iVR, we recorded procedural
Osteotomies in the musculoskeletal system are joint preserving procedures to correct the alignment of the patient. In the lower limb, most of the pre-operative planning is performed on full leg weightbearing radiographs. However, these images contain a 2-dimensional projection of a 3-dimensional deformity, lack a clear visualization of the joint surface and are prone to rotational
During shoulder arthroplasty the native functionality of the diseased shoulder joint is restored, this functionality is strongly dependent upon the native anatomy of the pre-diseased shoulder joint. Therefore, surgeons often use the healthy contralateral scapula to plan the surgery, however in bilateral diseases such as osteoarthritis this is not always feasible. Virtual reconstructions are then used to reconstruct the pre-diseased anatomy and plan surgery or subject-specific implants. In this project, we develop and validate a statistical shape modeling method to reconstruct the pre-diseased anatomy of eroded scapulae with the aim to investigate the existence of predisposing anatomy for certain shoulder conditions. The training dataset for the statistical shape model consisted of 110 CT images from patients without observable scapulae pathologies as judged by an experienced shoulder surgeon. 3D scapulae models were constructed from the segmented images. An open-source non-rigid B-spline-based registration algorithm was used to obtain point-to-point correspondences between the models. The statistical shape model was then constructed from the dataset using principle component analysis. The cross-validation was performed similarly to the procedure described by Plessers et al. Virtual defects were created on each of the training set models, which closely resemble the morphology of glenoid defects according to the Wallace classification method. The statistical shape model was reconstructed using the leave-one-out method, so the corresponding training set model is no longer incorporated in the shape model. Scapula reconstruction was performed using a Monte Carlo Markov chain algorithm, random walk proposals included both shape and pose parameters, the closest fitting proposal was selected for the virtual reconstruction. Automatic 3D measurements were performed on both the training and reconstructed 3D models, including glenoid version, critical shoulder angle, glenoid offset and glenoid center position. The root-mean-square
High tibial osteotomy (HTO) is an effective surgical treatment for isolated medial compartment knee osteoarthritis; however, widespread adoption is limited due to difficulty in achieving the planned correction, and patient dissatisfaction due to soft tissue irritation. A new HTO system – Tailored Osteotomy Knee Alignment (TOKA®, 3D Metal Printing Ltd, Bath, UK) could potentially address these barriers having a custom titanium plate and titanium surgical guides featuring a unique mechanism for precise osteotomy opening as well as saw cutting and drilling guides. The aim of this study was to assess the accuracy of this novel HTO system using cadaveric specimens; a preclinical testing stage ahead of first-in-human surgery according to the ‘IDEAL-D’ framework for device innovation. Local ethics committee approval was obtained. The novel opening wedge HTO procedure was performed on eight cadaver leg specimens. Whole lower limb CT scans pre- and post-operatively provided geometrical assessment quantifying the discrepancy between pre-planned and post-operative measurements for key variables: the gap opening angle and the patient specific surgical instrumentation positioning and rotation - assessed using the implanted plate. The average discrepancy between the pre-operative plan and the post-operative osteotomy correction angle was: 0.0 ± 0.2°. The R2 value for the regression correlation was 0.95. The average
A number of classification systems exist for posterior malleolus fractures of the ankle. The reliability of these classification systems remains unclear. The primary aim of this study was to evaluate the reliability of three commonly utilised fracture classification systems of the posterior malleolus. 60 patients across 2 hospitals sustaining an unstable ankle fracture with a posterior malleolus fragment were identified. All patients underwent radiographs and computed tomography of their injured ankle. 9 surgeons including pre-ST3 level, ST3-8 level, and consultant level applied the Haraguchi, Rammelt, and Mason & Molloy classifications to these patients, at two timepoints, at least 4 weeks apart. The order was randomised between assessments. Inter-rater reliability was assessed using Fleiss’ kappa and 95% confidence intervals (CI). Intra-rater reliability was assessed using Cohen's Kappa and standard
Introduction. The current methods for measuring femoral torsion have limitations, including variability and inaccuracies. Existing 3D methods are not reliable for abnormal femoral anteversion measurement. A new 3D method is needed for accurate measurement and planning of proximal femoral osteotomies. Currently available software for viewing and modelling CT data lacks measurement capabilities. The MSK Hip planner aims to address these limitations by combining measurement, planning, and analysis functionalities into one tool. We aim to answer 5 key questions: Is there a difference between 2D measurement methods? Is there a difference between 3D measurement methods? Is there a difference between 2D and 3D measurement methods? Are any of the measurement methods affected by the presence of osteoarthritis or a CAM deformity?. Method. After segmentation was carried out on 42 femoral CT scans using Osirix, 3D bone models were landmarked in the MSK lab hip planning software. Murphy's, Reikeras’, McBryde, and the novel MSK lab method were used to measure femoral anteversion. Result. Murphy's method had the lowest mean femoral neck anteversion (FNA) at 24.98°, while the MSK method had the highest at 28.55°. Bland-Altman plots showed systematic
Little information exists when using cell viability assays to evaluate cells within whole tissue, particularly specific types such as the intervertebral disc (IVD). When comparing the reported methodologies and the protocols issued by manufacturers, the processing, working times, and dye concentrations vary significantly, making the assay's reproducibility a costly and time-consuming trial and
Abstract. Background. 1. 63,284 patients presented with neck of femur fractures in England in 2020 (NHFD report 2021)2. To maximise theatre efficiency during the first wave of COVID-19, NHSE guidance recommended the use of HA for most patients requiring arthroplasty.3. The literature reports an incidence of Hemiarthroplasty dislocations of 1–15%. Aims. 1. To study the number and possible causes of dislocations in patients with Primary hemiarthroplasty for fracture neck of femur2. To compare our data with national and international data in terms of dislocation and revision rates for Hemiarthroplasty. Methods. Retrospective study Duration- 1st April 2021–31st March 20223. Inclusion criteria- Patients with neck of femur fracture treated with Hemiarthroplasty. Exclusion criteria- Patients treated with other surgical options for neck of femur fractures. Results. 1. No. of neck of femur patients- 4442. No. of patients treated with Hemiarthroplasty- 2143. No. of dislocations- 44. 75% were female, 75% had AMTS>7, 50% were operated within 36 hours of trauma, 75% dislocated within a month of surgery, 75% of the dislocations were revised. 5. One dislocation was due to >72 hour delay to surgery, second dislocation was due to smaller offset and shortening, third was due to acetabular dysplasia, fourth was due to larger head used. Discussion-1. Our 1.86% dislocation rate matches that in the literature of 1–6%2. 75% dislocated within a month of surgery matches that in the literature that maximum dislocations occur within one month. 3. Closed reduction as definitive method of management of dislocation 25% matches that in the literature of 22–25%4. 75% dislocations revised similar to literature of 75–80% revision rate. Conclusion. Pre-operative templating can reduce surgical
Lower back pain (LBP) is a global problem. Countless in vitro studies have attempted to understand LBP and inform treatment protocols such as disc replacement devices (DRDs). A common method of reporting results is applying a linear fit to load-displacement behaviour, reporting the gradient as the specimen stiffness in that axis. This is favoured for speed, simplicity and repeatability but neglects key aspects including stiffening and hysteresis. Other fits such as polynomials and double sigmoids better address these characteristics, but solution parameters lack physical representation. The aim of this study was to implement an automated method to fit spinal load-displacement behaviour using viscoelastic models. Six porcine lumbar spinal motion segments were dissected to produce isolated disc specimens. These were potted in Wood's metal, ensuring the disc midplane remained horizontal, sprayed with 0.9% saline and wrapped in saline-soaked tissue and plastic wrap to prevent dehydration. Specimens were tested using the University of Bath spine simulator operating under position control with a 400N axial preload. Specimens were approximated using representative viscoelastic elements. These models were constructed in MATLAB Simulink R2020b using the SimScape library. Solution coefficients were determined by minimizing the sum of squared