As advancements in total knee arthroplasty progress at an exciting pace, two areas are of special interest, as they directly impact implant design and surgical decision making. Knee morphometry considers the three-dimensional shape of the articulating surfaces within the knee joint, and knee phenotyping provides the ability to categorize alignment into practical groupings that can be used in both clinical and research settings. This annotation discusses the details of these concepts, and the ways in which they are helping us better understand the individual subtleties of each patient’s knee. Cite this article: Bone Joint J 2024;106-B(12):1363–1368

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 Error, and Mean Absolute Error were utilized as evaluation metrics. Results. Both the models performed at excellent levels for single-label segmentations in bones (dice > 0.95), but single-label joint space performance remained considerably lower (dice < 0.87). Multi-class segmentations remained at lower performance (dice < 0.88) for both models. Combining bone and joint space labels may have introduced a class imbalance problem in multi-class models, leading to lower performance. Conclusion. It is not clear if 3D UNETR provides better performance as the selection of hyperparameters was the same across the models and was not optimized. Further evaluations will be needed with baseline UNET and nnUNET modeling architectures

Introduction. Adolescent Idiopathic Scoliosis (AIS) is a three-dimensional deformity of the spine with unclear etiology. Due to the asymmetry of lateral curves, there are differences in the muscle activation between the convex and concave sides. This study utilized a comprehensive thoracic spine and ribcage musculoskeletal model to improve the biomechanical understanding of the development of AIS deformity and approach an explanation of the condition. Methods. In this study, we implemented a motion capture model using a generic rigid-body thoracic spine and ribcage model, which is kinematically determinate and controlled by spine posture obtained, for instance, from radiographs. This model is publicly accessible via a GitHub repository. We simulated gait and standing models of two AIS (averaging 15 years old, both with left lumbar curve and right thoracic curve averaging 25 degrees) and one control subject. The marker set included extra markers on the sternum and the thoracic and lumbar spine. The study was approved by the regional Research Ethics Committee (Journal number: H17034237). Results. We investigated the difference between the muscle activation on the right and left sides including erector spinae (ES), psoas major (PS), and multifidus (MF). Results of the AIS simulations indicated that, on average throughout the gait cycle, the right ES, left PS and left MF had 46%, 44%, and 23% higher activities compared to the other side, respectively. In standing, the ratios were 28%, 40%, and 19%, respectively. However, for the control subject, the differences were under 7%, except ES throughout the gait, which was 17%. Conclusion. The musculoskeletal model revealed distinct differences in force patterns of the right and left sides of the spine, indicating an instability phenomenon, where larger curves lead to higher muscle activations for stabilization. Acknowledgement. The project is funded by the European Union's Horizon 2020 program through Marie Skłodowska-Curie grant No. [764644]

Introduction. Anterior shoulder instability results in labral and osseous glenoid injuries. With a large osseous defect, there is a risk of recurrent dislocation of the joint, and therefore the patient must undergo surgical correction. An MRI evaluation of the patient helps to assess the soft tissue injury. Currently, the volumetric three-dimensional (3D) reconstructed CT image is the standard for measuring glenoid bone loss and the glenoid index. However, it has the disadvantage of exposing the patient to radiation and additional expenses. This study aims to compare the values of the glenoid index using MRI and CT. Method. The present study was a two-year cross-sectional study of patients with shoulder pain, trauma, and dislocation in a tertiary hospital in Karnataka. The sagittal proton density (PD) section of the glenoid and enface 3D reconstructed images of the scapula were used to calculate glenoid bone loss and the glenoid index. The baseline data were analyzed using descriptive statistics, and the Chi-square test was used to test the association of various complications with selected variables of interest. Result. The glenoid index calculated in the current study using 3D volumetric CT images and MR sagittal PD images was 0.95±0.01 and 0.95±0.01, respectively. The CT and MRI glenoid bone loss was 5.41±0.65% and 5.38±0.65%, respectively. When compared, the glenoid index and bone loss calculated by MRI and CT revealed a high correlation and significance with a p-value of <0.001. Conclusions. The study concluded that MRI is a reliable method for glenoid measurement. The sagittal PD sequence combined with an enface glenoid makes it possible to identify osseous defects linked to glenohumeral joint damage and dislocation. The values derived from 3D CT are identical to the glenoid index and bone loss determined using the sagittal PD sequence in MRI

Introduction. The surgical treatment of critical-sized bone defects with complex three-dimensional (3D) geometries is a challenge for the treating surgeon. Additive manufacturing such as 3D printing enables the production of highly individualized bone implants meeting the shape of the patient's bone defect and including a tunable internal structure. In this study, we showcase the design process for patient-specific implants with critical-sized tibia defects. Methods. Two clinical cases of patients with critical tibia defects (size 63×20×21 mm and 50×24×17 mm) were chosen. Brainlab software was used for segmentation of CT data generating 3D models of the defects. The implant construction involves multiple stages. Initially, the outer shell is precisely defined. Subsequently, the specified volume is populated with internal structures using Voronoi, Gyroid, and NaCl crystal structures. Variation in pore size (1.6 mm and 1.0 mm) was accomplished by adjusting scaffold size and material thickness. Results. An algorithmic design process in Rhino and Grasshopper was successfully applied to generate model implants for the tibia from Ct data. By integrating a precise mesh into an outer shell, a scaffold with controlled porosity was designed. In terms of the internal design, both Voronoi and Gyroid form macroscopically homogeneous properties, while NaCl, exhibits irregularities in density and consequently, in the strength of the structure. Data implied that Voronoi and Gyroid structures adapt more precisely to complex and irregular outer shapes of the implants. Conclusion. In proof-of-principle studies customized tibia implants were successfully generated and printed as model implants based on resin. Further studies will include more patient data sets to refine the workflows and digital tools for a broader spectrum of bone defects. The algorithm-based design might offer a tremendous potential in terms of an automated design process for 3D printed implants which is essential for clinical application

The objective of this study was to assess the clinical and radiological results of patients who were revised using a new generation custom-made triflange acetabular component (CTAC) for component loosening and large acetabular defect (Paprosky 3A and 3B) after previous total hip arthroplasty (THA). New generation CTACs involve the use of patient-specific drill guides and incorporate three-dimensional printed bone models, enhancing precision during surgical implantation. Data were extracted from a single centre prospective database of patients with large acetabular defects who were treated with a new generation CTAC. Patients were included if they had a minimum follow-up of five years. The modified Oxford Hip Score (mOHS), EurQol EuroQoL five-dimension three-level (EQ-5D-3L) utility, and Numeric Rating Scale (NRS), including visual analogue score (VAS) for pain, were gathered at baseline, and at two- and five-year follow-up. Reasons for revision, and radiological and clinical complications were registered. Trends over time are described and tested for significance (p<0.05). A total of 49 (70%) of 70 patients with a mean age of 73.5 years (SD 7.7) had a complete follow-up of 5 years. A significant improvement was found in HOOS, mOHS, EQ-5D-3L utility and NRS, VAS pain rest and activity between baseline and final follow-up. Complications included 8 cases with loosening screws, 4 with bony fractures, 4 periprosthetic infections and 2 cases with dislocation. One patient with bilateral pelvic discontinuity had revision surgery due to recurrent dislocations. No revision surgery was performed for screw failure or implant breakage. New generation CTAC in patients with THA acetabular loosening and massive acetabular bone loss (Paprosky 3A and 3B) can result in stable constructs and significant improvement in functioning and health-related quality of life at five years’ follow-up

Hip precautions following total hip arthroplasty (THA) limits flexion, adduction and internal rotation, yet these precautions cause unnecessary psychological stress. This study aims to assess bony and implant impingement using virtual models from actual patient's bony morphology and spinopelvic parameters to deduce whether hip precautions are necessary with precise implant positioning in the Asian population. Individualized sitting and standing sacral slope data of robotic THAs performed at two tertiary referral centers in Hong Kong was inputted into the simulation system based on patients’ pre-operative sitting and standing lumbar spine X-rays. Three-dimensional dynamic models were reconstructed using the Stryker Mako THA 4.0 software to assess bony and implant impingement both anteriorly and posteriorly, with default cup placement at 40° inclination and 20° anteversion. Femoral anteversion followed individual patient's native version. A 36mm hip ball was chosen for all cups equal or above 48mm and 32mm for those below. Anterior impingement was assessed by hip flexion and posterior impingement was assessed by hip extension. 113 patients were included. At neutral rotation and adduction, no patients had anterior implant impingement at hip flexion of 100°. 1.7% had impingement at 110°, 3.5% had impingement at 120°, 9.7% had impingement at 130°. With 20° of internal rotation and adduction, 0.8% had anterior implant impingement at hip flexion of 90°, 7.1% had impingement at 100° and 18.5% had impingement at 110°. With the hip externally rotated by 20°, 0.8% of patients had posterior implant impingement, and 8.8% bony impingement at 0° extension. With enabling technology allowing accurate component positioning, hip precautions without limiting forward flexion in neutral position is safe given precise implant positioning and adequate osteophyte removal. Patients should only be cautioned about combined internal rotation, adduction with flexion

Background. Weightbearing computed tomography scans allow for better understanding of foot alignment in patients with Progressive Collapsing Foot Deformity. However, soft tissue integrity cannot be assessed via WBCT. As performing both WBCT and magnetic resonance imaging is not cost effective, we aimed to assess whether there is an association between specific WBCT and MRI findings. Methods. A cohort of 24 patients of various stages of PCFD (mean age 51±18 years) underwent WBCT scans and MRI. In addition to signs of sinus tarsi impingement, four three-dimensional measurements (talo-calcaneal overlap, talo-navicular coverage, Meary's angle axial/lateral) were obtained using a post processing software (DISIOR 2.1, Finland) on the WBCT datasets. Sinus tarsi obliteration, spring ligament complex and tibiospring ligament integrity, as well as tibialis posterior tendon degeneration were evaluated with MRI. Statistical analysis was performed for significant (P<0.05) correlation between findings. Results. None of the assessed 3D measurements correlated with spring ligament complex or tibiospring ligament tears. Age, body mass index, and TCO were associated with tibialis posterior tendon tears. 75% of patients with sinus tarsi impingement on WBCT also showed signs of sinus tarsi obliteration on MRI. Of the assessed parameters, only age and BMI were associated with sinus tarsi obliteration diagnosed on MRI, while the assessed WBCT based 3D measurements were, with the exception of MA axial, associated with sinus tarsi impingement. Conclusion. While WBCT reflects foot alignment and indicates signs of osseous impingement in PCFD patients, the association between WBCT based 3D measurements and ligament or tendon tears in MRI is limited. Partial or complete tears of the tibialis posterior tendon were only detectable in comparably older and overweight PCFD patients with an increased TCO. WBCT does not replace MRI in diagnostic value. Both imaging options add important information and may impact decision-making in the treatment of PCFD patients

Introduction. Pes cavovarus is a foot deformity that can be idiopathic (I-PC) or acquired secondary to other pathology. Charcot-Marie-Tooth disease (CMT) is the most common adult cause for acquired pes cavovarus deformity (CMT-PC). The foot morphology of these distinct patient groups has not been previously investigated. The aim of this study was to assess if morphological differences exist between CMT-PC, I-PC and normal feet (controls) using weightbearing computed tomography (WBCT). Methods. A retrospective analysis of WBCT scans performed between May 2013 and June 2017 was undertaken. WBCT scans from 17 CMT-PC, 17 I-PC and 17 healthy normally-aligned control feet (age-, side-, sex- and body mass index-matched) identified from a prospectively collected database, were analysed. Eight 2-dimensional (2D) and three 3-dimensional (3D) measurements were undertaken for each foot and mean values in the three groups were compared using one-way ANOVA with the Bonferroni correction. Results. Significant differences were observed between CMT-PC or I-PC and controls (p< 0.05). Two-dimensional measurements were similar in CMT-PC and I-PC, except for forefoot arch angle (p= 0.04). 3D measurements (foot and ankle offset, calcaneal offset and hindfoot alignment angle) demonstrated that CMT-PC exhibited more severe hindfoot varus malalignment than I-PC (p= 0.03, 0.04 and 0.02 respectively). Discussion. CMT-related cavovarus and idiopathic cavovarus feet are morphologically different from healthy feet, and CMT feet exhibit increased forefoot supination and hindfoot malalignment compared to idiopathic forms. The use of novel three-dimensional analysis may help highlight subtle structural differences in patients with similar foot morphology but aetiologically different pathology

INTRODUCTION. Simulation plays an important role in surgical education and the ability to perfect surgical performance. Simulation can be enhanced by adding various layers of realism to the experience. Haptic feedback enhances the simulation experience by providing tactile responses and virtual reality imagery provides an immersive experience and allows for greater appreciation of three-dimensional structures. In this study, we present a proof-of-concept haptic simulator to replicate key steps of a cervical laminoplasty procedure. The technology uses affordable components and is easily modifiable so that it can be used from novice through to expert level. Custom models can be easily added ensuring the simulator can be used in a wide range of orthopaedic applications from baseline education through to day of surgery pre-operative simulation. METHOD. We used the Unity Game Engine, the 3D Systems “Touch” Haptic Feedback Device (HFD), and a Meta Quest VR headset. Our system uses a number of complex algorithms to track the shape and provide haptic feedback of a virtual bone model. This allows for simulation of various tools including a high-speed burr, Kerrison rongeur and intraoperative X-rays. RESULTS. Our simulator replicates the tactile sensations of bone-burring tasks. Although we focused on the cervical laminoplasty procedure, the system can load data from CT scans, enabling the simulation of multiple other procedures. The parts cost of our system, $10,000 NZD, is a fraction of the cost of traditional surgical simulators. DISCUSSION. Our simulator reduces financial barriers to accessing orthopaedic simulators. Trainees can perform hands-on practice without compromising patient safety. The immersive nature of VR, combined with realistic haptic feedback, enables trainees to develop the dexterity and three-dimensional understanding of detailed bony work. Further refinements are needed before we can perform validation studies on our system. CONCLUSIONS. We present an affordable surgical simulator capable of simulating bony surgical procedures in a VR environment using haptic feedback technology and consumer-grade components. ACKNOWLEDGEMENTS. This research was made possible by the generosity of the Wishbone Trust

Aims

This systematic review aims to identify 3D predictors derived from biplanar reconstruction, and to describe current methods for improving curve prediction in patients with mild adolescent idiopathic scoliosis.

Aims

To systematically evaluate whether bracing can effectively achieve curve regression in patients with adolescent idiopathic scoliosis (AIS), and to identify any predictors of curve regression after bracing.

Articular cartilage (AC) and subchondral bone (SB) are intimately intertwined, forming a complex unit called the AC-SB interface. Our recent studies have shown that cartilage and bone marrow are connected by a three-dimensional network of microchannels (i.e. cartilage-bone marrow microchannel connector; CMMC), which differ microarchitecturally in number, size and morphology depending on the maturation stage of the bone and the region of the joint. However, the pathological significance of CMMC is largely unknown. Here, we quantitatively assessed how CMMC microarchitecture relates to cartilage condition and regional differences in early idiopathic osteoarthritis (OA). Two groups of cadaveric female human femoral heads (intact cartilage vs early cartilage lesions) were identified and biopsy-based high-resolution micro-CT imaging was used. Subchondral bone (SB) thickness, CMMC number, maximum and minimum CMMC size, and CMMC morphology were quantified and compared between the two groups. The effect of joint region and cartilage condition on each dependent variable was examined. The number and morphology of CMMCs were influenced by the region of the joint, but not by the cartilage condition. On the other hand, the minimum and maximum CMMC size was modified by both joint location and cartilage condition. The smallest CMMCs were consistently found in the load bearing region (LBR) of the joint. Compared to healthy subjects, the size of the microchannels was increased in early OA, most notably in the non-load bearing region (NLBR) and the peripheral rim (PR) of the femoral head. In addition, subchondral bone thinning was observed in early OA as a localized event associated with areas of partial chondral defect. Our data suggest an enlargement of the SB microchannel network and a collective structural deterioration of the SB in early idiopathic OA

A comprehensive understanding of the self-repair abilities of menisci and their overall function in the knee joint requires three-dimensional information. However, previous investigations of the meniscal blood supply have been limited to two-dimensional imaging methods, which fail to accurately capture tissue complexity. In this study, micro-CT was used to analyse the 3D microvascular structure of the meniscus, providing a detailed visualization and precise quantification of the vascular network. A contrast agent (μAngiofil®) was injected directly into the femoral artery of cadaver legs to provide the proper contrast enhancement. First, the entire knee joint was analysed with micro-CT, then to increase the applicable resolution the lateral and medial menisci were excised and investigated with a maximum resolution of up to 4 μm. The resulting micro-CT datasets were analysed both qualitatively and quantitatively. Key parameters of the vascular network, such as vascular volume fraction, vessel radius, vessel length density, and tortuosity, were separately determined for the lateral and medial meniscus, and their four circumferential zones defined by Cooper. In accordance with previous literature, the quantitative micro-CT data confirm a decrease in vascular volume fraction along the meniscal zones. The highest concentration of blood vessels was measured in the meniscocapsular region 0, which is characterized by vascular segments with a significantly larger average radius. Furthermore, the highest vessel length density observed in zone 0 suggests a more rapid delivery of oxygen and nutrients compared to other regions. Vascular tortuosity was detected in all circumferential regions, indicating the occurrence of vascular remodelling in all tissue areas. In conclusion, micro-CT is a non-invasive imaging technique that allows for the visualization of the internal structure of an object in three dimensions. These advanced 3D vascular analyses have the potential to establish new surgical approaches that rely on the healing potential of specific areas of the meniscus. Acknowledgements: The authors acknowledge R. Hlushchuk, S. Halm, and O. Khoma from the University of Bern for their help with contrast agent perfusions

Regenerative medicine (RM) promises to restore both the mechanical functionality and the biological composition of tissues after damage. Three-dimensional scaffolds are used in RM to host cells and let them produce proteins that are the building blocks of the native tissues. While regenerating tissues evolve over time through dynamic biomechanical and biochemical changes, current scaffolds’ generation are passive causing mechanical mismatch, suboptimal growth, and pain. Furthermore, current scaffolds ignore the complexity of the reciprocal bio-mechanics regulation, hindering the design of the next-gen scaffolds. To regenerate tissues and organs, biofabrication strategies that impart spatiotemporal control over cell-cell and cell-extracellular matrix communication, often through control over cell and material deposition and placement, are being developed. To achieve these targets, the spatiotemporal control over biological signals at the interface between cells and materials is often aimed for. Alternatively, biological activity can be triggered through the control of mechanical cues, harnessing more fundamental know-how in mechanobiology that could be combined with biofabrication strategies. Here, I present some of our most recent advancements in merging mechanobiology with biofabrication that enabled the control of cell activity, moving towards enhanced tissue regeneration as well as the possibility to create more complex 3D in vitro models to study biological processes

Biofabrication is a popular technique to produce personalized constructs for tissue engineering. In this study we combined laponite (Lap), gellan gum (GG) with platelet-rich plasma (PRP) aiming to enhance the endothelial regeneration through the synergistic effects of their individual properties. Laponite has the ability to form porous three-dimensional networks mimicking the extracellular matrix structure, and PRP delivery of growth factors stimulates the endothelial cell proliferation and migration, offering a composite bioink for cell growth and support. The sustained release of these growth factors from the GG-laponite-PRP composite material over time provides a continuous source of stimulation for the cells, leading to more effective tissue engineering strategies for endothelial tissue regeneration. Four blend compositions comprising 1% w/v GG and 0.5 or 1% w/v Lap and 25% v/v PRP were combined with Wharton jelly mesenchymal stem cells (WJ-MSCs) and bioprinted into vessel-like structures with an inner diameter of 3 mm and a wall thickness of 1 mm. Stress/strain analysis revealed the elastomeric properties of the hydrogels with Young modulus values of 10 MPa. Increasing the Lap concentration led to a non-significant decrease of swelling ratio from 93 to 91%. Live/dead assay revealed cell viability of at least 76%, with the 0.5%Lap-GG viability exceeding 99% on day 21. Gradual increase of glycosaminoglycans accumulation and collagen production indicate promotion of ECM formation. The expression and membranous localization of PECAM-1 from day 7 and the granular intracellular localization of vWF after 2 weeks demonstrate in vitro endothelial functionality. In vivo subcutaneous implantation indicated the absence of any adverse immunological reactions. The results reveal the expression of both vWF and PECAM-1 by WJ-MSCs entrapped in all four construct compositions with significantly higher expression of vWF in the presence of PRP

It is known that the gait dynamics of elderly substantially differs from that of young people. However, it has not been well studied how this age-related gait dynamics affects the knee biomechanics, e.g., cartilage mechanical response. In this study, we investigated how aging affects knee biomechanics in a female population using subject-specific computational models. Two female subjects (ages of 23 and 69) with no musculoskeletal disorders were recruited. Korea National Institute for Bioethics Policy Review Board approved the study. Participants walked at a self-selected speed (SWS), 110% of SWS, and 120% of SWS on 10 m flat ground. Three-dimensional marker trajectories and ground reaction forces (Motion Analysis, USA), and lower limbs’ muscle activities were measured (EMG, Noraxon USA). Knee cartilage and menisci geometries were obtained from subjects’ magnetic resonance images (3T, GE Health Care). An EMG-assisted musculoskeletal finite element modeling workflow was used to estimate knee cartilage tissue mechanics in walking trials. Knee cartilage and menisci were modeled using a transversely isotropic poroviscoelastic material model. Walking speed in SWS, 110%, and 120% of SWS were 1.38 m/s, 1.51 m/s, and 1.65 m/s for the young, and 1.21 m/s, 1.34 m/s and 1.46 m/s for the elderly, respectively. The maximum tensile stress in the elderly tibial cartilage was ~25%, ~33%, and ~32% lower than the young at SWS, 110%, and 120% of SWS, respectively. These preliminary results suggest that the cartilage in the elderly may not have enough stimulation even at 20% increases in walking speed, which may be one reason for tissue degeneration. To enhance these findings, further study with more subjects and different genders will investigate how age-related gait dynamics affects knee biomechanics. Acknowledgments: Australian NHMRC Ideas Grant (APP2001734), KITECH (JE220006)

Acute syndesmotic ankle injuries continue to impose a diagnostic dilemma and it remains unclear whether weighbearing or external rotation should be exerted rotation during the imaging process. Therefore, we aimed to implement both axial load (weightbearing) and external rotation in the assessment of a clinical cohort of patients with syndesmotic ankle injuries syndesmotic using weightbearing CT imaging. In this retrospective comparative cohort study, patients with an acute syndesmotic ankle injury were analyzed using a WBCT (N= 20; Mean age= 31,64 years; SD= 14,07. Inclusion criteria were an MRI confirmed syndesmotic ankle injury imaged by a bilateral WBCT of the ankle during weightbearing and combined weightbearing-external rotation. Exclusion criteria consisted of fracture associated syndesmotic ankle injuries. Three-dimensional (3D) models were generated from the CT slices. Tibiofibular displacement and Talar Rotation was quantified using automated3D measurements (Anterior TibioFibular Distance (ATFD), Alpha Angle, Posterior TibioFibular Distance (PTFD) and Talar Rotation (TR) Angle) in comparison to a cohort of non-injured ankles. Results. The difference in neutral-stressed Alpha° and ATFD showed a significant difference between patients with a syndesmotic ankle lesion and healthy ankles (P = 0.046 and P = 0.039, respectively) The difference in neutral-stressed PTFD and TR° did not show a significant difference between patients with a syndesmotic ankle lesion and healthy ankles (P = 0.492; P = 0.152, respectively). Conclusion. Application of combined weightbearing-external rotation reveals a dynamic anterior tibiofibular widening in patients with syndesmotic ankle injuries. This study provides the first insights based on 3D measurements to support the potential relevance of applying external rotation during WBCT imaging. However, to what extent certain displacement patterns are associated with syndesmotic instability and thus require operative treatment strategies has yet to be determined in future studies

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 errors is often neglected. This study aims to include the three-dimensional (3D) spatial information in the development of a novel framework for segmentation accuracy evaluation and comparison between different methods. Predicted and ground truth (manually segmented) segmentation masks are meshed into 3D surfaces. A template mesh of the same anatomical structure is then registered to all ground truth 3D surfaces. This ensures all surface points on the ground truth meshes to be in the same anatomically homologous order. Next, point-wise surface deviations between the registered ground truth mesh and the meshed segmentation prediction are calculated and allow for color plotting of point-wise descriptive statistics. Statistical parametric mapping includes point-wise false discovery rate (FDR) adjusted p-values (also referred to as q-values). The framework reads volumetric image data containing the segmentation masks of both ground truth and segmentation prediction. 3D color plots containing descriptive statistics (mean absolute value, maximal value,…) on point-wise segmentation errors are rendered. As an example, we compared segmentation results of nnUNet [2], UNet++ [3] and UNETR [4] by visualizing the mean absolute error (surface deviation from ground truth) as a color plot on the 3D model of bone and cartilage of the mean distal femur. A novel framework to evaluate segmentation accuracy is presented. Output includes anatomical information on the segmentation errors, as well as point-wise comparative statistics on different segmentation algorithms. Clearly, this allows for a better informed decision-making process when selecting the best algorithm for a specific clinical application

Abstract. Objectives. The syndesmosis joint, located between the tibia and fibula, is critical to maintaining the stability and function of the ankle joint. Damage to the ligaments that support this joint can lead to ankle instability, chronic pain, and a range of other debilitating conditions. Understanding the kinematics of a healthy joint is critical to better quantify the effects of instability and pathology. However, measuring this movement is challenging due to the anatomical structure of the syndesmosis joint. Biplane Video Xray (BVX) combined with Magnetic Resonance Imaging (MRI) allows direct measurement of the bones but the accuracy of this technique is unknown. The primary objective is to quantify this accuracy for measuring tibia and fibula bone poses by comparing with a gold standard implanted bead method. Methods. Written informed consent was given by one participant who had five tantalum beads implanted into their distal tibia and three into their distal fibula from a previous study. Three-dimensional (3D) models of the tibia and fibula were segmented (Simpleware Scan IP, Synopsis) from an MRI scan (Magnetom 3T Prisma, Siemens). The beads were segmented from a previous CT and co-registered with the MRI bone models to calculate their positions. BVX (125 FPS, 1.25ms pulse width) was recorded whilst the participant performed level gait across a raised platform. The beads were tracked, and the bone position of the tibia and fibula were calculated at each frame (DSX Suite, C-Motion Inc.). The beads were digitally removed from the X-rays (MATLAB, MathWorks) allowing for blinded image-registration of the MRI models to the radiographs. The mean difference and standard deviation (STD) between bead-generated and image-registered bone poses were calculated for all degrees of freedom (DOF) for both bones. Results. The absolute mean tibia and fibula bone position differences (Table 1) between the bead and BVX poses were found to be less than 0.5 mm for both bones. The bone rotation differences were found to be less than 1° for all axes except for the fibula Z axis rotation which was found to be 1.46°. One study. 1. has reported the kinematics of the syndesmosis joint and reported maximum ranges of motion of 9.3°and translations of 3.3mm for the fibula. The results show that the accuracy of the methodology is sufficient to quantify these small movements. Conclusions. BVX combined with MRI can be used to accurately measure the syndesmosis joint. Future work will look at quantifying the accuracy of the talus to provide further understanding of normal ankle kinematics and to quantify the kinematics across a healthy population to act as a comparator for future patient studies. Declaration of Interest. (b) declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported:I declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research project