Background. Two-stage revision arthroplasty is the standard treatment for chronic hip and knee periprosthetic joint infections (PJI). Accurate diagnosis of persistent infections at 2nd stage using established biomarkers and diagnostic criteria is of paramount importance. This study aimed to evaluate the diagnostic value of synovial calprotectin and alpha-defensin, and compare established diagnostic criteria from the International Consensus Meeting (ICM 2018) and the European Bone and Joint Infection Society (EBJIS 2021) to determine persistent PJI at the 2nd stage of a two-stage revision arthroplasty. Methods. We retrospectively analyzed 97 patients who underwent 100 two-stage revisions (hip: 39, knee: 61). Synovial fluid samples were assessed for calprotectin and alpha-defensin levels. ICM 2018 and EBJIS 2021 were applied to all patients undergoing 2nd stage revision. Receiver operating characteristic (ROC) curves and Youden Index were utilized to determine optimal cut-off values, and correlations between biomarkers were evaluated. The microbiological spectrum was analyzed at 2nd stage and re-revision surgery. Results. Calprotectin levels showed a sensitivity of 66.7%, specificity of 32.9%, and accuracy of 38.0% in predicting septic failure. Alpha-defensin showed sensitivity of 28.6%, specificity of 87.8%, and accuracy of 79.2%. Significant correlations included: calprotectin with PMN% (r = 0.471, p = 0.05) and alpha-defensin with WBC (r = 0.830, p < 0.01) in the successful cohort. For septic re-revisions, calprotectin and alpha-defensin were highly correlated (r = 0.969, p < 0.01). ICM correctly diagnosed persistent PJI in 26.7%, while EBJIS diagnosed 24.2%. The microbial spectrum shifted from gram-positive to gram-negative bacteria between reimplantation and re-revision surgeries. Conclusion. Synovial calprotectin and alpha-defensin demonstrated limited accuracy in ruling out persistent PJI at reimplantation. The low sensitivity of current diagnostic criteria, combined with the observed shift in microbial spectrum, underscores the challenges in diagnosing persistent PJI during 2nd stage of a two-stage revisions arthroplasty

Aim. Fast and accurate identification of pathogens causing periprosthetic joint infections (PJI) is essential to initiate effective antimicrobial treatment. Culture-based approaches frequently yield false negative results, despite clear signs of infection. This may be due to the use of general growth media, which do not mimic the conditions at site of infection. Possible alternative approaches include DNA-based techniques, the use of in vivo-like media and isothermal microcalorimetry (ITC). We developed a synthetic synovial fluid (SSF) medium that closely resembles the in vivo microenvironment and allows to grow and study PJI pathogens in physiologically relevant conditions. In this study we investigated whether the use of ITC in combination with the SSF medium can improve accuracy and time to detection in the context of PJI. Methods. In this study, 120 synovial fluid samples were included, aspirated from patients with clinical signs of PJI. For these samples microbiology data (obtained in the clinical microbiology lab using standard procedures) and next generation sequencing (NGS) data, were available. The samples were incubated in the SSF medium at different oxygen levels (21% O. 2. , 3% O. 2. and 0% O. 2. ) for 10 days. Every 24h, the presence of growth was checked. From positive samples, cultures were purified on Columbia blood agar and identified using MALDI-TOF. In parallel, heat produced by metabolically active microorganisms present in the samples was measured using ITC (calScreener, Symcel), (96h at 37°C, in SSF, BHI and thioglycolate). From the resulting thermograms the ‘time to activity’ could be derived. The accuracy and time to detection were compared between the different detection methods. Results. So far, seven samples were investigated. Using conventional culture-based techniques only 14.3% of the samples resulted in positive cultures, whereas NGS indicated the presence of microorganisms in 57.1% of the samples (with 3/7 samples being polymicrobial). Strikingly, 100% of the samples resulted in positive cultures after incubation in the SSF medium, with time to detection varying from 1 to 9 days. MALDI-TOF revealed all samples to be polymicrobial after cultivation in SSF, identifying organisms not detected by conventional techniques or NGS. For the samples investigated so far, signals obtained with ITC were low, probably reflecting the low microbial load in the first set of samples. Conclusion. These initial results highlight the potential of the SSF medium as an alternative culture medium to detect microorganisms in PJI context. Further studies with additional samples are ongoing; in addition, the microcalorimetry workflow is being optimized

Aim. Diagnosing low-grade periprosthetic joint infections (PJI) can be very challenging due to low-virulent microorganisms capable of forming biofilm. Clinical signs can be subtle and may be similar to those of aseptic failure. To minimize morbidity and mortality and to preserve quality of life, accurate diagnosis is essential. The aim of this study was to assess the performance of various diagnostic tests in diagnosing low-grade PJI. Methods. Patients undergoing revision surgery after total hip and knee arthroplasty were included in this retrospective cohort study. A standardized diagnostic workup was performed using the components of the 2021 European Bone and Joint Infection Society (EBJIS) definition of PJI. For statistical analyses, the respective test was excluded from the infection definition to eliminate incorporation bias. Receiver-operating-characteristic curves were used to calculate the diagnostic performance of each test, and their area-under-the-curves (AUC) were compared using the z-test. Results. 422 patients undergoing revision surgery after total hip and knee arthroplasty were included in this study. 208 cases (49.3%) were diagnosed as septic. Of those, 60 infections (28.8%) were defined as low-grade PJI (symptoms >4 weeks and caused by low-virulent microorganisms (e. g. coagulase-negative staphylococci, Cutibacterium spp., enterococci and Actinomyces)). Performances of the different test methods are listed in Table 1. Synovial fluid (SF) - WBC (white blood cell count) >3000G/L (0.902), SF - %PMN (percentage of polymorphonuclear neutrophils) > 65% (0.959), histology (0.948), and frozen section (0.925) showed the best AUCs. Conclusion. The confirmatory criteria according to the EBJIS definition showed almost ideal performances in ruling-in PJI (>99% specificity). Histology and synovial fluid cell count (SF-WBC and SF-%PMN) showed excellent accuracies for diagnosing low-grade PJI. However, a reduced immune reaction in these cases may necessitate lower cut-off values. Intraoperative frozen section may be valuable in cases with inconclusive preoperative diagnosis. For any tables or figures, please contact the authors directly

Aim. Accurate diagnosis is key in correctly managing prosthetic joint infection(PJI). Shoulder PJI definition and diagnosis is challenging. Current PJI definitions, based overwhelmingly in hip/knee research, may not accurately diagnose shoulder PJI. Our aim is to compare the preoperative performance of two PJI definitions comparing it to definitive postoperative classification. Method. This is a retrospective study of patients who have undergone total shoulder revision surgery for infection between 2005 and 2022. Cases were classified using two different PJI definitions: a)the European Bone and Joint Infection Society (EBJIS) and; 2)the 2018 International Consensus Meeting(ICM) PJI specific shoulder definition. Preoperative classification was based on clinical features, inflammatory markers and synovial fluid leukocyte count and definitive classification also considered microbiology and histology results. Results. Preoperative and definitive PJI classification status of the 21 patients included were evaluated and is summarized in table 1. The shoulder specific 2018 ICM definition showed the highest agreement between preoperative and definitive classification (76.2%, k=0.153, p=0.006) compared to EBJIS (52.4%, k=0.205, p=0.006). In all cases, the classification was changed because of positive intraoperative microbiology (at least two identical isolates). Microbiology findings showed coagulase negative staphylococci, Staphyloccocus aureus and Cutibacterium acnes to be the most frequent. Four patients had polymicrobial infections. Conclusions. Both the EBJIS 2021 and 2018 ICM definitions have low accuracy in predicting shoulder PJI preoperatively. Clearly further studies with larger cohorts are in dire need focusing specifically on shoulder revision arthroplasty to improve on existing definitions. Caution is advised while extrapolating of criteria/thresholds recommended for hip/knee joints. For any tables or figures, please contact the authors directly

Aim. Evaluate if Neutrophil Extracellular Traps related biomarkers (citrullinated histone H3 [H3Cit], cellfree DNA [cfDNA], and myeloperoxidase) are increased in synovial fluid of patients with PJI and investigate the diagnostic accuracy of NET formation biomarkers for PJI. Method. Patients who underwent hip or knee revision total joint arthroplasty were categorised into two groups according to the Second International Consensus Meeting on Musculoskeletal Infection (2018) criteria. Sixteen patients were classified as infected and 16 as non-infected. cf-DNA, myeloperoxidase and H3Cit were measured in synovial fluid collected during surgery. Sensitivity, specificity, and receiver operating characteristic (ROC) curve were calculated. Results. Patients with PJI presented significantly higher levels of synovial fluid cf-DNA (105.5 ng/ml ± 58.3 vs 1.9 ± 1.2, p>0.0001), myeloperoxidase (1575 pg/ml ± 826 vs 50.16 ± 100, p<0.0001) and citrullinated histone H3 (1.688 ± 1.214 vs 13.88 ± 24.4, p < 0.0001). In the ROC curve analyses, the area under the curve for cf-DNA, myeloperoxidase and H3cit were 1 [0.89 – 1], 0.98 [0.86 – 1], and 0.94 [0.8 – 0.99], respectively. The sensitivity for detecting PJI using synovial fluid was 100% for cf-DNA, 93,7% for myeloperoxidase, and 87,5% for H3cit. The sensitivity for cf-DNA and myeloperoxidase was 100%, and 87,5 % for H3cit. Conclusions. Our results show that neutrophils within periprosthetic microenvironment release NETs as part of the bactericidal arsenal to fight infection. These results allow a better understanding of the cellular and molecular processes that occur in this microenvironment, enabling the design of more assertive strategies for the identification of new biomarkers and for a better use of the available ones. Furthermore, novel studies are needed to define whether and how NET-related biomarkers can be useful for the diagnosis of PJI

Aim. This study aimed to externally validate promising preoperative PJI prediction models in a recent, multinational European cohort. Method. Three preoperative PJI prediction models (by Tan et al., Del Toro et al., and Bülow et al.) which previously demonstrated high levels of accuracy were selected for validation. A multicenter retrospective observational analysis was performed of patients undergoing total hip arthroplasty (THA) and total knee arthroplasty (TKA) between January 2020 and December 2021 and treated at centers in the Netherlands, Portugal, and Spain. Patient characteristics were compared between our cohort and those used to develop the prediction models. Model performance was assessed through discrimination and calibration. Results. A total of 2684 patients were included of whom 60 developed a PJI (2.2%). Our patient cohort differed from the models’ original cohorts in terms of demographic variables, procedural variables, and the prevalence of comorbidities. The c-statistics for the Tan, Del Toro, and Bülow models were 0.72, 0.69, and 0.72 respectively. Calibration was reasonable, but precise percentage estimates for PJI risk were most accurate for predicted risks up to 3-4%; the Tan model overestimated risks above 4%, while the Del Toro model underestimated risks above 3%. Conclusions. In this multinational cohort study, the Tan, Del Toro, and Bülow PJI prediction models were found to be externally valid for classifying high risk patients for developing a PJI. These models hold promise for clinical application to enhance preoperative patient counseling and targeted prevention strategies. Keywords. Periprosthetic Joint Infection (PJI), High Risk Groups, Prediction Models, Validation, Infection Prevention

Introduction. With advances in artificial intelligence, the use of computer-aided detection and diagnosis in clinical imaging is gaining traction. Typically, very large datasets are required to train machine-learning models, potentially limiting use of this technology when only small datasets are available. This study investigated whether pretraining of fracture detection models on large, existing datasets could improve the performance of the model when locating and classifying wrist fractures in a small X-ray image dataset. This concept is termed “transfer learning”. Method. Firstly, three detection models, namely, the faster region-based convolutional neural network (faster R-CNN), you only look once version eight (YOLOv8), and RetinaNet, were pretrained using the large, freely available dataset, common objects in context (COCO) (330000 images). Secondly, these models were pretrained using an open-source wrist X-ray dataset called “Graz Paediatric Wrist Digital X-rays” (GRAZPEDWRI-DX) on a (1) fracture detection dataset (20327 images) and (2) fracture location and classification dataset (14390 images). An orthopaedic surgeon classified the small available dataset of 776 distal radius X-rays (Arbeidsgmeischaft für Osteosynthesefragen Foundation / Orthopaedic Trauma Association; AO/OTA), on which the models were tested. Result. Detection models without pre-training on the large datasets were the least precise when tested on the small distal radius dataset. The model with the best accuracy to detect and classify wrist fractures was the YOLOv8 model pretrained on the GRAZPEDWRI-DX fracture detection dataset (mean average precision at intersection over union of 50=59.7%). This model showed up to 33.6% improved detection precision compared to the same models with no pre-training. Conclusion. Optimisation of machine-learning models can be challenging when only relatively small datasets are available. The findings of this study support the potential of transfer learning from large datasets to improve model performance in smaller datasets. This is encouraging for wider application of machine-learning technology in medical imaging evaluation, including less common orthopaedic pathologies

Introduction. Accurate assessment of alignment in pre-operative and post-operative knee radiographs is important for planning and evaluating knee replacement surgery. Existing methods predominantly rely on manual measurements using long-leg radiographs, which are time-consuming to perform and are prone to reliability errors. In this study, we propose a machine-learning-based approach to automatically measure anatomical varus/valgus alignment in pre-operative and post-operative standard AP knee radiographs. Method. We collected a training dataset of 816 pre-operative and 457 one-year post-operative AP knee radiographs of patients who underwent knee replacement surgery. Further, we have collected a separate distinct test dataset with both pre-operative and one-year post-operative radiographs for 376 patients. We manually outlined the distal femur and the proximal tibia/fibula with points to capture the knee joint (including implants in the post-operative images). This included point positions used to permit calculation of the anatomical tibiofemoral angle. We defined varus/valgus as negative/positive deviations from zero. Ground truth measurements were obtained from the manually placed points. We used the training dataset to develop a machine-learning-based automatic system to locate the point positions and derive the automatic measurements. Agreement between the automatic and manual measurements for the test dataset was assessed by intra-class correlation coefficient (ICC), mean absolute difference (MAD) and Bland-Altman analysis. Result. Analysing the agreement between the manual and automated measurements, ICC values were excellent pre-/post-operatively (0.96, CI: 0.94-0.96) / (0.95, CI: 0.95-0.96). Pre-/post-operative MAD values were 1.3°±1.4°SD / 0.7°±0.6°SD. The Bland-Altman analysis showed a pre-/post-operative mean difference (bias) of 0.3°±1.9°SD/-0.02°±0.9°SD, with pre-/post-operative 95% limits of agreement of ±3.7°/±1.8°, respectively. Conclusion. The developed machine-learning-based system demonstrates high accuracy and reliability in automatically measuring anatomical varus/valgus alignment in pre-operative and post-operative knee radiographs. It provides a promising approach for automating the measurement of anatomical alignment without the need for long-leg radiographs. Acknowledgements. This research was funded by the Wellcome Trust [223267/Z/21/Z]

Introduction. Experimental bone research often generates large amounts of histology and histomorphometry data, and the analysis of these data can be time-consuming and trivial. Machine learning offers a viable alternative to manual analysis for measuring e.g. bone volume versus total volume. The objective was to develop a neural network for image segmentation, and to assess the accuracy of this network when applied to ectopic bone formation samples compared to a ground truth. Method. Thirteen tissue slides totaling 114 megapixels of ectopic bone formation were selected for model building. Slides were split into training, validation, and test data, with the test data reserved and only used for the final model assessment. We developed a neural network resembling U-Net that takes 512×512 pixel tiles. To improve model robustness, images were augmented online during training. The network was trained for 3 days on a NVidia Tesla K80 provided by a free online learning platform against ground truth masks annotated by an experienced researcher. Result. During training, the validation accuracy improved and stabilised at approx. 95%. The test accuracy was 96.1 %. Conclusion. Most experiments using ectopic bone formation will yield an inter-observer or inter-method variance of far more than 5%, so the current approach may be a valid and feasible technique for automated image segmentation for large datasets. More data or a consensus-based ground truth may improve training stability and validation accuracy. The code and data of this project are available upon request and will be available online as part of our publication

Introduction. The increased prevalence of osteoporosis in the patient population undergoing reverse shoulder arthroplasty (RSA) results in significantly increased complication rates. Mainly demographic and clinical predictors are currently taken into the preoperative assessment for risk stratification without quantification of preoperative computed tomography (CT) data (e.g. bone density). It was hypothesized that preoperative CT bone density measures would provide objective quantification with subsequent classification of the patients’ humeral bone quality. Methods. Thirteen bone density parameters from 345 preoperative CT scans of a clinical RSA cohort represented the data set in this study. The data set was divided into testing (30%) and training data (70%), latter included an 8-fold cross validation. Variable selection was performed by choosing the variables with the highest descriptive value for each correlation clustered variables. Machine learning models were used to improve the clustering (Hierarchical Ward) and classification (Support Vector Machine (SVM)) of bone densities at risk for complications and were compared to a conventional statistical model (Logistic Regression (LR)). Results. Clustering partitioned this cohort (training data set) into a high bone density subgroup consisting of 96 patients and a low bone density subgroup consisting of 146 patients. The optimal number of clusters (n = 2) was determined based on optimization metrics. Discrimination of the cross validated classification model showed comparable performance for the training (accuracy=91.2%; AUC=0.967) and testing data (accuracy=90.5 %; AUC=0.958) while outperforming the conventional statistical model (Logistic Regression (LR)). Local interpretable model-agnostic explanations (LIME) were created for each patient to explain how the predicted output was achieved. Conclusion. The trained and tested model provides preoperative information for surgeons treating patients with potentially poor bone quality. The use of machine learning and patient-specific calibration showed that multiple 3D bone density scores improved accuracy for objective preoperative bone quality assessment

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. Knee Osteoarthritis (KOA) is a prevalent joint disease requiring accurate diagnosis and prompt management. The condition occurs due to cartilage deterioration and bone remodeling. Ultrasonography has emerged as a promising modality for diagnosing KOA. Medial meniscus extrusion (MME), characterized by displacement of medial meniscus beyond the joint line has been recognized as a significant marker of KOA progression. This study aimed to explore potentials Ultrasound findings in timely detection of MME and compare it to magnetic resonance imaging (MRI) as a reference standard. Method. A comprehensive literature search was performed in 4 databases from inception to May 1 2024. Two independent reviewers, initiated screening protocols and selected the articles based on inclusion and exclusion criteria and then extracted the data. Meta-analysis was conducted using R 4.3.2 packages mada and metafor. Result. A total of 2500 articles from 4 databases was retrieved; however, following the application of inclusion and exclusion criteria 23 articles were finally extracted. These studies collectively encompassed a total of 777 patients with mean age of 53.2±7.4. The mean BMI calculated for patients was 28.31 ± 2.45. All patients underwent non-weight bearing knee ultrasonography in supine position with 0° flexion. The reported medial meniscus extrusion was 2.58 mm for articles using MRI and 2.65 mm for those using Ultrasound (MD: 0.05 ± 0.12, P= 0.65, I. 2. : 54%). Our meta-analysis revealed insignificant difference between US and MRI. (SMD: 0.03, 95% CI: -0.18 _0.23, P= 0.77, I. 2. : 56%) Meta analysis for diagnostic accuracy measures yielded a pooled sensitivity and specificity of 90.8% and 77% (95% CI: 84.2% – 94.8%, 35.5% – 95.3%, respectively, I. 2. : 44%). Conclusion. Our results indicate a close alignment in the accuracy of measurements obtained using Ultrasound modality. The narrow range suggests a minimal discrepancy in MME values between MRI and ultrasound, highlighting their comparable precision in diagnostic assessments

Introduction. The objective assessment of shoulder function is important for personalized diagnosis, therapies and evidence-based practice but has been limited by specialized equipment and dedicated movement laboratories. Advances in AI-driven computer vision (CV) using consumer RGB cameras (red-blue-green) and open-source CV models offer the potential for routine clinical use. However, key concepts, evidence, and research gaps have not yet been synthesized to drive clinical translation. This scoping review aims to map related literature. Method. Following the JBI Manual for Evidence Synthesis, a scoping review was conducted on PubMed and Scholar using search terms including “shoulder,” “pose estimation,” “camera”, and others. From 146 initial results, 27 papers focusing on clinical applicability and using consumer cameras were included. Analysis employed a Grounded Theory approach guided iterative refinement. Result. Studies primarily used Microsoft Kinect (infrared-based depth sensing, RGB camera; discontinued) or monocular consumer cameras with open-source CV-models, sometimes supplemented by LiDAR (laser-based depth sensing), wearables or markers. Technical validation studies against gold standards were scarce and too inconsistent for comparison. Larger range of motion (RoM) movements were accurately recorded, but smaller movements, rotations and scapula tracking remained challenging. For instance, one larger validation study comparing shoulder angles during arm raises to a marker-based gold-standard reported Pearson's R = 0.98 and a standard error of 2.4deg. OpenPose and Mediapipe were the most used CV-models. Recent efforts try to improve model performance by training with shoulder specific movements. Conclusion. Low-cost, routine clinical movement analysis to assess shoulder function using consumer cameras and CV seems feasible. It can provide acceptable accuracy for certain movement tasks and larger RoM. Capturing small, hidden or the entirety of shoulder movement requires improvements such as via training models with shoulder specific data or using dual cameras. Technical validation studies require methodological standardization, and clinical validation against established constructs is needed for translation into practice

Introduction. The identification of biological markers associated to implant failure in THA (total hip arthroplasty) patients remains a challenge in orthopedic surgery. In this search, previous studies have been mainly focused on typical mediators associated to bone metabolism and inflammation. Our group has evaluated changes in serum levels of insulin-like growth factor binding protein-1 (IGFBP-1), a protein which is not directly related to bone homeostasis, in patients undergoing THA. Method. We assessed IGFBP-1 levels in serum obtained from 131 patients (58% female, 42 % male; age: 68 ± 13 years) who underwent THA in the Orthopedic Surgery and Traumatology Department of our institution. In this cohort, 57% of patients had metal on polyethylene (MoP) as hip-bearing surface combination, 17 % had ceramic on ceramic (CoC) and 26% of them did not have any prosthesis. A test based on an enzyme-linked immunosorbent assay (ELISA) was used to determine IGFBP-1 levels in serum obtained from these patients. Result. Our results showed a significant increase in IGFBP- 1 levels in MoP group as compared to CoC and control groups, in which no differences in quantified levels were detected. Further analysis revealed no significant differences in IGFBP-1 between cemented and non-cemented MoP bearings. We performed a ROC curve to evaluate the accuracy of serum IGFBP-1 in discriminating MoP from the rest of patients (area under the curve: 0.7; 95% confidence interval: 0,6-0.8; p<0.05) and established a cut-off value of 10.2 ng/ml, according to the Youden´s Index. Logistic regression analysis showed that patients with MoP bearing surfaces had a higher risk of increased IGFBP- 1 levels in serum (p<0.05, Odds Ratio: 6.7, 95% Confidence Interval 3.1 to 14.8). Conclusion. IGFBP- 1 levels are significantly elevated in THA patients with MoP bearing surfaces, suggesting that this protein might be a reliable biomarker for the outcome of patients implanted with MoP

Introduction. Patient-specific biomechanical modeling using Finite Element Analysis (FEA) is pivotal for understanding the structural health of bones, optimizing surgical procedures, assessing outcomes, and validating medical devices, aligning with guidance issued by standards and regulatory bodies. Accurate mapping of image-to-mesh-material is crucial given bone's heterogeneous composition. This study aims to rigorously assess mesh convergence and evaluate the sensitivity of material grouping strategies in quantifying bone strength. Method. Subject-specific geometry and nonlinear material properties were derived from computed tomography (CT) scan data of one cadaveric human vertebral body. Linear tetrahedral elements with varying edge lengths between 2mm and 0.9mm were then generated to study the mesh convergence. To compare the effectiveness of different grouping strategies, three approaches were used: Modulus Gaping (a user-defined absolute threshold of Young's modulus ranging from 500 MPa to 1 MPa), Percentual Thresholding (relative parameter thresholds ranging from 50% to 1%), and Adaptive clustering (unsupervised k-means-based clustering ranging from 10 to 200 clusters). Adaptive clustering enables a constant number of unique material properties in cross-specimen studies, improving the validity of results. Result. Mesh convergence was evaluated via fracture load and reached at a 1mm mesh size across grouping strategies. All strategies exhibit minimal deviation (within 5%) from individually assigned material parameters, except Modulus Gaping, with a 500 MPa threshold (32% difference). Computational efficiency, measured by runtime, significantly improved with grouping strategies, reducing computational cost by 82 to 94% and unique material count by up to 99%. Conclusion. Different grouping strategies offer comparable mesh convergence, highlighting their potential to reduce computational complexity while maintaining accuracy in the biomechanical modeling of bones and suggesting a more efficient approach than individual element materials. The higher efficiency of FEA may increase its applicability in clinical settings with limited computational resources. Further studies are needed to refine grouping parameters and assess their suitability across different subjects

Introduction. Shoulder arthroplasty (SA) has been performed with different types of implants, each requiring different replacement systems. However, data on previously utilized implant types are not always available before revision surgery, which is paramount to determining the appropriate equipment and procedure. Therefore, this meta-analysis aimed to evaluate the accuracy of the AI models in classifying SA implant types. Methods. This systematic review was conducted in Pubmed, Embase, SCOPUS, and Web of Science from inception to December 2023, according to PRISMA guidelines. Peer-reviewed research evaluating the accuracy of AI-based tools on upper-limb X-rays for recognizing and categorizing SA implants was included. In addition to the overall meta-analysis, subgroup analysis was performed according to the type of AI model applied (CNN (Convolutional neural network), non-CNN, or Combination of both) and the similarity of utilized datasets between studies. Results. 13 articles were eligible for inclusion in this meta-analysis (including 138 different tests assessing models’ efficacy). Our meta-analysis demonstrated an overall sensitivity and specificity of 0.891 (95% CI:0.866-0.912) and 0.549 (95% CI:0.532,0.566) for classifying implants in SA, respectively. The results of our subgroup analyses were as follows: CNN-subgroup: a sensitivity of 0.898 (95% CI:0.873-0.919) and a specificity of 0.554 (95% CI:0.537,0.570), Non-CNN subgroup: a sensitivity of 0.809 (95% CI:0.665-0.900) and specificity of 0.522 (95% CI:0.440,0.603), combined subgroup: a sensitivity of 0.891 (95% CI:0.752-0.957) and a specificity of 0.547 (95% CI:0.463,0.629). Studies using the same dataset demonstrated an overall sensitivity and specificity of 0.881 (95% CI:0.856-0.903) and 0.542 (95% CI:0.53,0.554), respectively. Studies that used other datasets showed an overall sensitivity and specificity of 0.995 (95% CI:969,0.999) and 0.678 (95% CI:0.234, 0.936), respectively. Conclusion. AI-based classification of shoulder implant types can be considered a sensitive method. Our study showed the potential role of using CNN-based models and different datasets to enhance accuracy, which could be investigated in future studies

Introduction. The recent introduction of Chatbots has provided an interactive medium to answer patient questions. The accuracy of responses with these programs in limb lengthening and reconstruction surgery has not previously been determined. Therefore, the purpose of this study was to assess the accuracy of answers from 3 free AI chatbot platforms to 23 common questions regarding treatment for limb lengthening and reconstruction. Method. We generated a list of 23 common questions asked by parents before their child's limb lengthening and reconstruction surgery. Each question was posed to three different AI chatbots (ChatGPT 3.5 [OpenAI], Google Bard, and Microsoft Copilot [Bing!]) by three different answer retrievers on separate computers between November 17 and November 18, 2023. Responses were only asked one time to each chatbot by each answer retriever. Nine answers (3 answer retrievers × 3 chatbots) were randomized and platform-blinded prior to rating by three orthopedic surgeons. The 4-point rating system reported by Mika et al. was used to grade all responses. Result. ChatGPT had the best response accuracy score (RAS) with a mean score of 1.73 ± 0.88 across all three raters (range of means for all three raters – 1.62 – 1.81) and a median score of 2. The mean response accuracy scores for Google Bard and Microsoft Copilot were 2.32 ± 0.97 and 3.14 ± 0.82, respectively. This ranged from 2.10 – 2.48 and 2.86 – 3.54 for Google Bard and Microsoft Copilot, respectively. The differences between the mean RAS scores were statistically significant (p < 0.0001). The median scores for Google Bard and Microsoft Copilot were 2 and 3, respectively. Conclusion. Using the Response Accuracy Score, the responses from ChatGPT were determined to be satisfactory, requiring minimal clarification, while the responses from Microsoft Copilot were either satisfactory, requiring moderate clarification, or unsatisfactory, requiring substantial clarification

Introduction. Intraoperative navigation systems for lumbar spine surgery allow to perform preoperative planning and visualize the real-time trajectory of pedicle screws. The aim of this study was to evaluate the deviation from preoperative planning and the correlations between screw deviation and accuracy. Method. Patients affected by degenerative spondylolisthesis who underwent posterior lumbar interbody fusion using intraoperative 3D navigation since April 2022 were included. Intraoperative cone-beam computed tomography (CBCT) was performed before screw planning and following implantation. The deviation from planning was calculated as linear, angular, and 3D discrepancies between planned and implanted screws. Accuracy and facet joint violation (FJV) were evaluated using Gertzbein-Robbins system (GRS) and Yson classification, respectively. Statistical analysis was performed using SPSS version28. One-way ANOVA followed by Bonferroni post-hoc tests were performed to evaluate the association between GRS, screw deviation and vertebral level. Statistical significance was set at p<0.05. Result. This study involved 34 patients, for a total of 154 pedicle screws. Mean age was 62.6±8.9 years. The mean two-dimensional screw tip deviation in mediolateral (ML), craniocaudal (CC), and anteroposterior (AP) was 2.6±2.45mm, 1.6±1.7mm, and 3.07±2.9mm, respectively. The mean screw tip 3D deviation was 5±3.3mm. The mean two-dimensional screw head deviation in ML, CC and AP was 1.83±1.8mm, 1.7±1.67mm and 3.6±3.1mm, respectively. The mean screw head 3D deviation was 4.94±3.2mm. 98% of screws were clinically acceptable (grade A+B), and grade 0 for FJV. Significant results were found between GRS and ML (p=0.005), AP (p=0.01) and 3D (p=0.003) tip deviations, and between GRS and AP and 3D head deviations (both p=0). Moreover, a significant correlation was found between GRS and vertebral level (p=0). Conclusion. Our results showed a reasonable rate of discrepancy between planned and positioned screws. However, accuracy was clinically acceptable in almost all cases. Therefore, pedicle screw fixation using intraoperative CBCT, 3D navigation and screw planning is safe and accurate

Introduction. Inaccurate identification of implants on X-rays may lead to prolonged surgical duration as well as increased complexity and costs during implant removal. Deep learning models may help to address this problem, although they typically require large datasets to effectively train models in detecting and classifying objects, e.g. implants. This can limit applicability for instances when only smaller datasets are available. Transfer learning can be used to overcome this limitation by leveraging large, publicly available datasets to pre-train detection and classification models. The aim of this study was to assess the effectiveness of deep learning models in implant localisation and classification on a lower limb X-ray dataset. Method. Firstly, detection models were evaluated on their ability to localise four categories of implants, e.g. plates, screws, pins, and intramedullary nails. Detection models (Faster R-CNN, YOLOv5, EfficientDet) were pre-trained on the large, freely available COCO dataset (330000 images). Secondly, classification models (DenseNet121, Inception V3, ResNet18, ResNet101) were evaluated on their ability to classify five types of intramedullary nails. Localisation and classification accuracy were evaluated on a smaller image dataset (204 images). Result. The YOLOv5s model showed the best capacity to detect and distinguish between different types of implants (accuracy: plate=82.1%, screw=72.3%, intramedullary nail=86.9%, pin=79.9%). Screw implants were the most difficult implant to detect, likely due to overlapping screw implants visible in the image dataset. The DenseNet121 classification model showed the best performance in classifying different types of intramedullary nails (accuracy=73.7%). Therefore, a deep learning model pipeline with the YOLOv5s and DenseNet121 was proposed for the most optimal performance of automating implants localisation and classification for a relatively small dataset. Conclusion. These findings support the potential of deep learning techniques in enhancing implant detection accuracy. With further development, AI-based implant identification may benefit patients, surgeons and hospitals through improved surgical planning and efficient use of theatre time

Aims

This Delphi study assessed the challenges of diagnosing soft-tissue knee injuries (STKIs) in acute settings among orthopaedic healthcare stakeholders.