Goal-directed Perioperative Albumin Substitution Versus Standard of Care to Reduce Postoperative Complications: A Randomized Clinical Trial (SuperAdd Trial).

OBJECTIVE: To investigate whether goal-directed albumin substitution during surgery and postanesthesia care to maintain a serum albumin concentration >30 g/L can reduce postoperative complications. BACKGROUND: Hypoalbuminemia is associated with numerous postoperative complications. Since albumin has important physiological functions, substitution of patients with hypoalbuminemia is worth considering. METHODS: We conducted a single-center, randomized, controlled, outcome assessor-blinded clinical trial in adult patients, American Society of Anesthesiologists physical status classification 3 to 4 or undergoing high-risk surgery. Patients, whose serum albumin concentration dropped <30 g/L were randomly assigned to goal-directed albumin substitution maintaining serum concentration >30 g/L or to standard care until discharge from the postanesthesia intermediate care unit. Standard of care allowed albumin substitution in hemodynamic instable patients with serum concentration <20 g/L, only. Primary outcome was the incidence of postoperative complications ≥2 according to the Clavien-Dindo Classification in at least 1 of 9 domains (pulmonary, infectious, cardiovascular, neurological, renal, gastrointestinal, wound, pain, and hematological) until postoperative day 15. RESULTS: Of 2509 included patients, 600 (23.9%) developed serum albumin concentrations <30 g/L. Human albumin 60 g (40-80 g) was substituted to 299 (99.7%) patients in the intervention group and to 54 (18.0%) in the standard care group. At least 1 postoperative complication classified as Clavien-Dindo Classification ≥2 occurred in 254 of 300 patients (84.7%) in the intervention group and in 262 of 300 (87.3%) in the standard treatment group (risk difference -2.7%, 95% CI, -8.3% to 2.9%). CONCLUSION: Maintaining serum albumin concentration of >30 g/L perioperatively cannot generally be recommended in high-risk noncardiac surgery patients.

Impact of metadata in multimodal classification of bone tumours.

The accurate classification of bone tumours is crucial for guiding clinical decisions regarding treatment and follow-up. However, differentiating between various tumour types is challenging due to the rarity of certain entities, high intra-class variability, and limited training data in clinical practice. This study proposes a multimodal deep learning model that integrates clinical metadata and X-ray imaging to improve the classification of primary bone tumours. The dataset comprises 1,785 radiographs from 804 patients collected between 2000 and 2020, including metadata such as age, affected bone site, tumour position, and gender. Ten tumour types were selected, with histopathology or tumour board decisions serving as the reference standard. METHODS: Our model is based on the NesT image classification model and a multilayer perceptron with a joint fusion architecture. Descriptive statistics included incidence and percentage ratios for discrete parameters, and mean, standard deviation, median, and interquartile range for continuous parameters. RESULTS: The mean age of the patients was 33.62 ± 18.60 years, with 54.73% being male. Our multimodal deep learning model achieved 69.7% accuracy in classifying primary bone tumours, outperforming the Vision Transformer model by five percentage points. SHAP values indicated that age had the most substantial influence among the considered metadata. CONCLUSION: The joint fusion approach developed in this study, integrating clinical metadata and imaging data, outperformed state-of-the-art models in classifying primary bone tumours. The use of SHAP values provided insights into the impact of different metadata on the model's performance, highlighting the significant role of age. This approach has potential implications for improving diagnostic accuracy and understanding the influence of clinical factors in tumour classification.

Significant difference in femoral torsion between coronal plane alignment of the knee type 1 and 4.

PURPOSE: The purpose of this study was to find out whether the torsions of the femur and tibia are dependent on the coronal plane alignment of the knee (CPAK) type. METHODS: Five hundred patients (1000 legs) were included, who received a whole leg standing three-dimensional (3D) radiograph using EOS imaging (EOS Imaging, Paris, France). SterEOS software was used for digital reconstruction. Femoral and tibial torsions were determined by analysing 3D reconstructions of each leg. Femoral torsion was defined as the angle between the femoral neck axis (FNA) and the posterior condylar axis (PCA). Tibial torsion was defined as the angle between the axis tangent to the posterior part of the tibia plateau and the transmalleolar axis. Arithmetic hip-knee-ankle angle (aHKA) and joint-line obliquity (JLO) were also determined, allowing each leg to be assigned one of nine possible phenotypes according to CPAK. RESULTS: The mean femoral torsion in CPAK type 1 was significantly higher (+ 2.6° ± 0.8°) than in CPAK type 4 (p = 0.02). All other CPAK types did not differ in the degree of femoral torsions. No differences could be demonstrated for the tibial torsion. CONCLUSION: There is a correlation between the coronal alignment of the lower limb and femoral torsion. This may provide the basis for extending the CPAK classification beyond the coronal plane. LEVEL OF EVIDENCE: Level III.

Definition of normal, neutral, deviant and aberrant coronal knee alignment for total knee arthroplasty.

PURPOSE: One of the most pertinent questions in total knee arthroplasty (TKA) is: what could be considered normal coronal alignment? This study aims to define normal, neutral, deviant and aberrant coronal alignment using large data from a computed tomography (CT)-scan database and previously published phenotypes. METHODS: Coronal alignment parameters from 11,191 knee osteoarthritis (OA) patients were measured based on three dimensional reconstructed CT data using a validated planning software. Based on these measurements, patients' coronal alignment was phenotyped according to the functional knee phenotype concept. These phenotypes represent an alignment variation of the overall hip knee ankle angle (HKA), femoral mechanical angle (FMA) and tibial mechanical angle (TMA). Each phenotype is defined by a specific mean and covers a range of ±1.5° from this mean. Coronal alignment is classified as normal, neutral, deviant and aberrant based on distribution frequency. Mean values and distribution among the phenotypes are presented and compared between two populations (OA patients in this study and non-OA patients from a previously published study). RESULTS: The arithmetic HKA (aHKA), combined normalised data of FMA and TMA, showed that 36.0% of knees were neutral within ±1 SD from the mean in both angles, 44.3% had either a TMA or a FMA within ±1-2 SD (normally aligned), 15.3% of the patients were deviant within ±2-3 SD and only 4.4% of them had an aberrant alignment (±3-4 SD in 3.4% and >4 SD in 1.0% of the patients respectively). However, combining the normalised data of HKA, FMA and TMA, 15.4% of patients were neutral in all three angles, 39.7% were at least normal, 27.7% had at least one deviant angle and 17.2% had at least one aberrant angle. For HKA, the males exhibited 1° varus and females were neutral. For FMA, the females exhibited 0.7° more valgus in mean than males and grew 1.8° per category (males grew 2.1° per category). For TMA, the males exhibited 1.3° more varus than females and both grew 2.3° and 2.4° (females) per category. Normal coronal alignment was 179.2° ± 2.8-5.6° (males) and 180.5 > ± 2.8-5.6° (females) for HKA, 93.1 > ± 2.1-4.2° (males) and 93.8 > ± 1.8-3.6° (females) for FMA and 86.7 > ± 2.3-4.6° (males) and 88 > ± 2.4-4.8° (females) for TMA. This means HKA 6.4 varus or 4.8° valgus (males) or 5.1° varus to 6.1° valgus was considered normal. For FMA HKA 1.1 varus or 7.3° valgus (males) or 0.2° valgus to 7.4° valgus was considered normal. For TMA HKA 7.9 varus or 1.3° valgus (males) or 6.8° varus to 2.8° valgus was considered normal. Aberrant coronal alignment started from 179.2° ± 8.4° (males) and 180.5 > ± 8.4° (females) for HKA, 93.1 > ± 6.3° (males) 93.8 > ± 5.4° (females) for FMA and 86.7 > ± 6.9° (males) and 88 > ± 7.2° (females) for TMA. This means HKA > 9.2° varus or 7.6° valgus (males) or 7.9° varus to 8.9° valgus was considered aberrant. CONCLUSION: Definitions of neutrality, normality, deviance as well as aberrance for coronal alignment in TKA were proposed in this study according to their distribution frequencies. This can be seen as an important first step towards a safe transition from the conventional one-size-fits-all to a more personalised coronal alignment target. There should be further definitions combining bony alignment, joint surfaces' morphology, soft tissue laxities and joint kinematics. LEVEL OF EVIDENCE: III.

SAM-X: sorting algorithm for musculoskeletal x-ray radiography.

OBJECTIVE: To develop a two-phased deep learning sorting algorithm for post-X-ray image acquisition in order to facilitate large musculoskeletal image datasets according to their anatomical entity. METHODS: In total, 42,608 unstructured and pseudonymized radiographs were retrieved from the PACS of a musculoskeletal tumor center. In phase 1, imaging data were sorted into 1000 clusters by a self-supervised model. A human-in-the-loop radiologist assigned weak, semantic labels to all clusters and clusters with the same label were merged. Three hundred thirty-two non-musculoskeletal clusters were discarded. In phase 2, the initial model was modified by "injecting" the identified labels into the self-supervised model to train a classifier. To provide statistical significance, data split and cross-validation were applied. The hold-out test set consisted of 50% external data. To gain insight into the model's predictions, Grad-CAMs were calculated. RESULTS: The self-supervised clustering resulted in a high normalized mutual information of 0.930. The expert radiologist identified 28 musculoskeletal clusters. The modified model achieved a classification accuracy of 96.2% and 96.6% for validation and hold-out test data for predicting the top class, respectively. When considering the top two predicted class labels, an accuracy of 99.7% and 99.6% was accomplished. Grad-CAMs as well as final cluster results underlined the robustness of the proposed method by showing that it focused on similar image regions a human would have considered for categorizing images. CONCLUSION: For efficient dataset building, we propose an accurate deep learning sorting algorithm for classifying radiographs according to their anatomical entity in the assessment of musculoskeletal diseases. KEY POINTS: • Classification of large radiograph datasets according to their anatomical entity. • Paramount importance of structuring vast amounts of retrospective data for modern deep learning applications. • Optimization of the radiological workflow and increase in efficiency as well as decrease of time-consuming tasks for radiologists through deep learning.

The impact of different alignment strategies on bone cuts in total knee arthroplasty for varus knee phenotypes.

PURPOSE: The purpose of this study was to visualise the influence of alignment strategy on bone resection in varus knee phenotypes. The hypothesis was that different amounts of bone resection would be required depending on the alignment strategy chosen. Through visualisation of the corresponding bone sections, it was hypothesised, it would be possible to assess which of the different alignment strategies would require the least amount of change to the soft tissues for the chosen phenotype, whilst still ensuring acceptable alignment of the components, and thus could be considered the most ideal alignment strategy. METHODS: Simulations of the different alignment strategies (mechanical, anatomical, constrained kinematic and unconstrained kinematic) in relation to their bone resections were performed on five common exemplary varus knee phenotypes. VARHKA174° VARFMA87° VARTMA84°, VARHKA174° VARFMA90° NEUTMA87°, VARHKA174° NEUFMA93° VARTMA84°, VARHKA177° NEUFMA93° NEUTMA87° and VARHKA177° VALFMA96° VARTMA81°. The phenotype system used categorises knees based on overall limb alignment (i.e. hip knee angle) but also takes into account joint line obliquity (i.e. TKA and FMA) and has been applied in the global orthopaedic community since its introduction in 2019. The simulations are based on long-leg radiographs under load. It is assumed that a change of 1° in the alignment of the joint line corresponds to a displacement of the distal condyle by 1 mm. RESULTS: In the most common phenotype VARHKA174° NEUFMA93° VARTMA84°, a mechanical alignment would result in an asymmetric elevation of the tibial medial joint line by 6 mm and a lateral distalisation of the femoral condyle by 3 mm, an anatomical alignment only by 0 and 3 mm, a restricted by 3 and 3 mm, respectively, whilst a kinematic alignment would result in no change in joint line obliquity. In the similarly common phenotype 2 VARHKA174° VARFMA90° NEUTMA87° with the same HKA, the changes are considerably less with only 3 mm asymmetric height change on one joint side, respectively, and no change in restricted or kinematic alignment. CONCLUSION: This study shows that significantly different amounts of bone resection are required depending on the varus phenotype and the alignment strategy chosen. Based on the simulations performed, it can, therefore, be assumed that an individual decision for the respective phenotype is more important than the dogmatically correct alignment strategy. By including such simulations, the modern orthopaedic surgeon can now avoid biomechanically inferior alignments and still obtain the most natural possible knee alignment for the patient.

Postoperative full leg radiographs exhibit less residual coronal varus deformity compared to intraoperative measurements in robotic arm-assisted total knee arthroplasty with the MAKO™ system.

PURPOSE: Robotic arm-assisted total knee arthroplasty (raTKA), currently a major trend in knee arthroplasty, aims to improve the accuracy of implant positioning and limb alignment. However, it is unclear whether and to what extent manual radiographic and navigation measurements with the MAKO™ system correlate. Nonetheless, a high agreement would be crucial to reliably achieve the desired limb alignment. METHODS: Thirty-six consecutive patients with osteoarthritis and a slight-to-moderate varus deformity undergoing raTKA were prospectively included in this study. Prior to surgery and at follow-up, a full leg radiograph (FLR) under weight-bearing conditions was performed. In addition, a computed tomography (CT) scan was conducted for preoperative planning. The hip-knee-ankle angle (HKA), mechanical lateral distal femur angle (mLDFA), mechanical medial proximal tibial angle (mMPTA) and joint line convergence angle (JLCA) were measured in the preoperative and follow-up FLR as well as in the CT scout (without weight-bearing) by three independent raters. Furthermore, the HKA was intraoperatively assessed with the MAKO™ system before and after raTKA. RESULTS: Significantly higher HKA values were identified for intraoperative deformity assessment using the MAKO system compared to the preoperative FLR and CT scouts (p = 0.006; p = 0.05). Intraoperative assessment of the HKA with final implants showed a mean residual varus deformity of 3.2° ± 1.9°, whereas a significantly lower residual varus deformity of 1.4° ± 1.9° was identified in the postoperative FLR (p < 0.001). The mMPTA was significantly higher in the preoperative FLR than in the CT scouts (p < 0.001). Intraoperatively, the mMPTA was adjusted to a mean of 87.5° ± 0.9° with final implants, while significantly higher values were measured in postoperative FLRs (p < 0.001). Concerning the mLDFA, no significant differences could be identified. CONCLUSION: The clinical importance of this study lies in the finding that there is a difference between residual varus deformity measured intraoperatively with the MAKO™ system and those measured in postoperative FLRs. This has implications for preoperative planning as well as intraoperative fine-tuning of the implant position during raTKA to avoid overcorrection of knees with slight-to-moderate varus osteoarthritis. LEVEL OF EVIDENCE: Level IV.

The impact of different alignment strategies on bone cuts for neutral knee phenotypes in total knee arthroplasty.

PURPOSE: The purpose of this study was to simulate and visualise the influence of the alignment strategy on bone resection in neutral knee phenotypes. It was hypothesised that different amounts of bone resection would be required depending on the alignment strategy chosen. The hypothesis was that by visualising the corresponding bone cuts, it would be possible to assess which of the different alignment strategies required the least change to the soft tissues for the chosen phenotype but still ensured acceptable component alignment and could, therefore, be considered the most ideal alignment strategy. METHODS: Simulations of the different alignment strategies (mechanical, anatomical, restricted kinematic and unrestricted kinematic) regarding their bone resections were performed on four common exemplary neutral knee phenotypes. NEUHKA0° VARFMA 90° VALTMA90°, NEUHKA0° NEUFMA 93° NEUTMA87°, NEUHKA0° VALFMA 96° NEUTMA87° and NEUHKA0° VALFMA 99° VARTMA84°. The phenotype system used categorises knees based on overall limb alignment (i.e. hip knee angle) but also considers joint line obliquity (i.e. TKA and FMA) and has been used globally since its introduction in 2019. These simulations are based on long leg weightbearing radiographs. It is assumed that a change of 1° in the alignment of the joint line corresponds to correspond to 1 mm of distal condyle offset. RESULTS: In the most common neutral phenotype NEUHKA0° NEUFMA 93° NEUTMA87°, with a prevalence of 30%, bone cuts remain below 4 mm regardless of alignment strategy. The greatest changes in the obliquity of the joint line can be expected for the mechanical alignment of the phenotype NEUHKA0° VALFMA 99° VARTMA84° where the medial tibia is raised by 6 mm and the lateral femur is shifted distally by 9 mm. In contrast, the NEUHKA0° VARFMA 90° VALTMA90° phenotype requires no change in joint line obliquity if the mechanical alignment strategy is used. CONCLUSION: Illustrations of alignment strategies help the treating surgeon to estimate the postoperative joint line obliquity. When considering the alignment strategy, it seems reasonable to prefer a strategy where the joint line obliquity is changed as little as possible. Although for the most common neutral knee phenotype the choice of alignment strategy seems to be of negligible importance, in general, even for neutral phenotypes, large differences in bone cuts can be observed depending on the choice of alignment strategy.

Prediction of complications and surgery duration in primary TKA with high accuracy using machine learning with arthroplasty-specific data.

PURPOSE: The number of primary total knee arthroplasties (TKA) is expected to rise constantly. For patients and healthcare providers, the early identification of risk factors therefore becomes increasingly fundamental in the context of precision medicine. Others have already investigated the detection of risk factors by conducting literature reviews and applying conventional statistical methods. Since the prediction of events has been moderately accurate, a more comprehensive approach is needed. Machine learning (ML) algorithms have had ample success in many disciplines. However, these methods have not yet had a significant impact in orthopaedic research. The selection of a data source as well as the inclusion of relevant parameters is of utmost importance in this context. In this study, a standardized approach for ML in TKA to predict complications during surgery and an irregular surgery duration using data from two German arthroplasty-specific registries was evaluated. METHODS: The dataset is based on two initiatives of the German Society for Orthopaedics and Orthopaedic Surgery. A problem statement and initial parameters were defined. After screening, cleaning and preparation of these datasets, 864 cases of primary TKA (2016-2019) were gathered. The XGBoost algorithm was chosen and applied with a hyperparameter search, a cross validation and a loss weighting to cope with class imbalance. For final evaluation, several metrics (accuracy, sensitivity, specificity, AUC) were calculated. RESULTS: An accuracy of 92.0%, sensitivity of 34.8%, specificity of 95.8%, and AUC of 78.0% were achieved for predicting complications in primary TKA and 93.4%, 74.0%, 96.3%, and 91.6% for predicting irregular surgery duration, respectively. While traditional statistics (correlation coefficient) could not find any relevant correlation between any two parameters, the feature importance revealed several non-linear outcomes. CONCLUSION: In this study, a feasible ML model to predict outcomes of primary TKA with very promising results was built. Complex correlations between parameters were detected, which could not be recognized by conventional statistical analysis. Arthroplasty-specific data were identified as relevant by the ML model and should be included in future clinical applications. Furthermore, an interdisciplinary interpretation as well as evaluation of the results by a data scientist and an orthopaedic surgeon are of paramount importance. LEVEL OF EVIDENCE: Level IV.

The synovial fluid calprotectin lateral flow test for the diagnosis of chronic prosthetic joint infection in failed primary and revision total hip and knee arthroplasty.

PURPOSE: The diagnostic criteria of prosthetic joint infection (PJI) recommended by the most commonly used diagnostic algorithms can be obscured or distorted by other inflammatory processes or aseptic pathology. Furthermore, the most reliable diagnostic criteria are garnered during revision surgery. A robust, reliable addition to the preoperative diagnostic cascade is warranted. Calprotectin has been shown to be an excellent diagnostic marker for PJI. In this study, we aimed to evaluate a lateral flow test (LFT) in the challenging patient cohort of a national referral centre for revision arthroplasty. METHODS: Beginning in March 2019, we prospectively included patients scheduled for arthroplasty exchange of a total hip (THA) or knee arthroplasty (TKA). Synovial fluid samples were collected intra-operatively. We used the International Consensus Meeting of 2018 (ICM) score as the gold standard. We then compared the pre-operative ICM score with the LFT result to calculate its diagnostic accuracy as a standalone pre-operative marker and in combination with the ICM score as part of an expanded diagnostic workup. RESULTS: A total of 137 patients with a mean age of 67 (± 13) years with 53 THA and 84 TKA were included. Ninety-nine patients (72.8%) were not infected, 34 (25.0) were infected, and four (2.9%) had an inconclusive final score and could not be classified after surgery. The calprotectin LFT had a sensitivity (95% confidence interval) of 0.94 (0.80-0.99) and a specificity of 0.87 (0.79-0.93). The area under the receiver operating characteristic curve (AUC) for the calprotectin LFT was 0.94 (0.89-0.99). In nine cases with an inconclusive pre-operative ICM score, the calprotectin LFT would have led to the correct diagnosis of PJI. CONCLUSIONS: The synovial fluid calprotectin LFT shows excellent diagnostic metrics both as a rule-in and a rule-out test, even in a challenging patient cohort with cases of severe osteolysis, wear disease, numerous preceding surgeries, and poor soft tissue conditions, which can impair the common diagnostic criteria. As it is available pre-operatively, this test might prove to be a very useful addition to the diagnostic algorithm.

Topical vancomycin powder does not affect patella cartilage degeneration in primary total knee arthroplasty and conversion rate for secondary patella resurfacing.

INTRODUCTION: Vancomycin powder (VP) is an antibiotic first introduced in pediatric spinal surgery to prevent surgical site infections (SSI). Recently its topical application was expanded to total hip and knee arthroplasty (THA, TKA) and anterior cruciate ligament reconstruction (ACLR). Toxicity to cartilage is the subject of current research. The aim of this study was to prove the hypothesis that topical application of VP in TKA does not result in a degeneration of patella cartilage. We propagate that the conversion rate for secondary patella resurfacing is not influenced by its use. MATERIALS AND METHODS: Between 2014 and 2021, 4292 joints were included in this monocentric retrospective cohort study. All patients underwent TKA without primary patella resurfacing. After a change of the procedure in the hospital, one group (VPG) was administered VP intraoperatively. The other group (nVPG) received no VP during surgery (nVPG). The remaining perioperative procedure was constant over the investigation period. Conversion rates for secondary patella resurfacing for both groups were determined without making distinctions in the indication. A second cohort was composed of patients presenting for follow-up examination 12 months after TKA and included 210 joints. Retrospective radiographic evaluations were performed preoperatively, before discharge and at follow-up examination. Patella axial radiographs were analyzed for patella tracking (lateral patellar tilt, patellar displacement) and patella degeneration (Sperner classification, patellofemoral joint space). RESULTS: There was no significant difference in the conversion rate for secondary patella resurfacing (4.24% VPG, 4.97% nVPG). Patella tracking and patella degeneration did not differ significantly between both groups. CONCLUSIONS: The topical application of VP does not influence the conversion rate for secondary patella resurfacing. Moreover, it does not result in a degeneration of patella cartilage in TK. LEVEL OF EVIDENCE: Retrospective case series, Level III.

Less soft tissue release in total knee arthroplasty for anteromedial compared to posteromedial knee osteoarthritis.

In total knee arthroplasty (TKA), the aim of achieving a mechanically straight leg axis as well as symmetrical and equally wide gaps has become established as the gold standard in terms of surgical technique. In contrast to TKA unicompartmental knee arthroplasty (UKA) is performed in anteromedial osteoarthritis (AMOA) and does not normally require releases. This raises the hypothesis whether the type of osteoarthritis (AMOA vs. posteromedial osteoarthritis (PMOA)) determines the requirement for soft tissue releases in TKA.In this retrospective study, 114 patients with medial osteoarthritis of the knee who had been treated with a navigated total knee replacement were consecutively included. On the basis of the preoperative lateral radiographs, the patients were divided into two groups: AMOA and PMOA. The incidence and the extent of releases performed were recorded using the navigation records.Patient-specific data (gender, age) did not differ between the groups (NS). Knees with AMOA presented an overall varus alignment of 5.3 ± 3.5°, knees with PMOA 8.0 ± 4.0° (p < 0.001). 30 cases (44%) had to be released in the AMOA group, compared with 33 cases (72%) in the PMOA group (p = 0.004). In the case of medial release, the extension gap increased 3.3 ± 2.4 mm in the AMOA compared to 5.3 ± 3.7 mm in the PMOA group (p = 0.006). The medial flexion gap was released 2.2 ± 2.6 mm in the AMOA and 2.9 ± 3.0 mm in the PMOA group (p = 0.008).To achieve a neutral mechanical alignment, a release has to be performed due to asymmetry of the extension gap more often if PMOA is present than in AMOA. Surgeons should be prepared to perform more frequent and extensive medial releases in PMOA. Higher constrained implants should be available in case of unintended over release in PMOA.

"Koehlers teardrop is not a reliable landmark for assessing the centre of rotation after Total hip arthroplasty" - a retrospective radiological study.

PURPOSE: Various anatomical landmarks have become established in radiography for the assessment of cup positioning after total hip arthroplasty (THA). The most important one is Koehler's teardrop figure (KTF). However, there is a lack of data on the validity of this landmark, which is widely used clinically for assessing the centre of rotation of the hip. METHOD: A retrospective measurement of the lateral and cranial distance of the KTF to the centre of hip rotation was performed on the basis of 250 X-ray images of patients who had undergone THA. In addition, the dependence of these distances on pelvic tilt was determined in 16 patients by means of virtual X-ray projections based on pelvic CTs. RESULTS: It was shown that the distance of the KTF from the centre of hip rotation in the horizontal plane is gender-dependent (men: 42.8 ± 6.0 mm vs. women: 37.4 ± 4.7 mm; p < 0.001) and age-dependent (Pearson correlation - 0.114; p < 0.05). Furthermore, the vertical and horizontal distances are subject to variation depending on height (Pearson correlation 0.14; p < 0.05 and 0.40; p < 0.001, respectively) and weight (Pearson correlation 0.158; p < 0.05). The distance between the KTF and the centre of hip rotation varies slightly depending on pelvic tilt. CONCLUSION: The KTF is not a sufficiently valid landmark for assessing the centre of rotation after THA. It is influenced by many different disturbance variables. However, it is largely robust against changes in pelvic tilt, so that it can be used as a reference point when comparing different intraindividual radiographs to assess the change in the centre of rotation due to implantation or to detect cup migration.

Applications of machine learning for imaging-driven diagnosis of musculoskeletal malignancies-a scoping review.

Musculoskeletal malignancies are a rare type of cancer. Consequently, sufficient imaging data for machine learning (ML) applications is difficult to obtain. The main purpose of this review was to investigate whether ML is already having an impact on imaging-driven diagnosis of musculoskeletal malignancies and what the respective reasons for this might be. A scoping review was conducted by a radiologist, an orthopaedic surgeon and a data scientist to identify suitable articles based on the PRISMA statement. Studies meeting the following criteria were included: primary malignant musculoskeletal tumours, machine/deep learning application, imaging data or data retrieved from images, human/preclinical, English language and original research. Initially, 480 articles were found and 38 met the eligibility criteria. Several continuous and discrete parameters related to publication, patient distribution, tumour specificities, ML methods, data and metrics were extracted from the final articles. For the synthesis, diagnosis-oriented studies were further examined by retrieving the number of patients and labels and metric scores. No significant correlations between metrics and mean number of samples were found. Several studies presented that ML could support imaging-driven diagnosis of musculoskeletal malignancies in distinct cases. However, data quality and quantity must be increased to achieve clinically relevant results. Compared to the experience of an expert radiologist, the studies used small datasets and mostly included only one type of data. Key to critical advancement of ML models for rare diseases such as musculoskeletal malignancies is a systematic, structured data collection and the establishment of (inter)national networks to obtain substantial datasets in the future. KEY POINTS: • Machine learning does not yet significantly impact imaging-driven diagnosis for musculoskeletal malignancies compared to other disciplines such as lung, breast or CNS cancer. • Research in the area of musculoskeletal tumour imaging and machine learning is still very limited. • Machine learning in musculoskeletal tumour imaging is impeded by insufficient availability of data and rarity of the disease.

Diagnosis of joint invasion in patients with malignant bone tumors: value and reproducibility of direct and indirect signs on MR imaging.

OBJECTIVES: To evaluate the performance and reproducibility of MR imaging features in the diagnosis of joint invasion (JI) by malignant bone tumors. METHODS: MR images of patients with and without JI (n = 24 each), who underwent surgical resection at our institution, were read by three radiologists. Direct (intrasynovial tumor tissue (ITT), intraarticular destruction of cartilage/bone, invasion of capsular/ligamentous insertions) and indirect (tumor size, signal alterations of epiphyseal/transarticular bone (bone marrow replacement/edema-like), synovial contrast enhancement, joint effusion) signs of JI were assessed. Odds ratios, sensitivity, specificity, PPV, NPV, and reproducibilities (Cohen's and Fleiss' κ) were calculated for each feature. Moreover, the diagnostic performance of combinations of direct features was assessed. RESULTS: Forty-eight patients (28.7 ± 21.4 years, 26 men) were evaluated. All readers reliably assessed the presence of JI (sensitivity = 92-100 %; specificity = 88-100%, respectively). Best predictors for JI were direct visualization of ITT (OR = 186-229, p < 0.001) and destruction of intraarticular bone (69-324, p < 0.001). Direct visualization of ITT was also highly reliable in assessing JI (sensitivity, specificity, PPV, NPV = 92-100 %), with excellent reproducibility (κ = 0.83). Epiphyseal bone marrow replacement and synovial contrast enhancement were the most sensitive indirect signs, but lacked specificity (29-54%). By combining direct signs with high specificity, sensitivity was increased (96 %) and specificity (100 %) was maintained. CONCLUSION: JI by malignant bone tumors can reliably be assessed on preoperative MR images with high sensitivity, specificity, and reproducibility. Particularly direct visualization of ITT, destruction of intraarticular bone, and a combination of highly specific direct signs were valuable, while indirect signs were less predictive and specific. KEY POINTS: • Direct visualization of intrasynovial tumor was the single most sensitive and specific (92-100%) MR imaging sign of joint invasion. • Indirect signs of joint invasion, such as joint effusion or synovial enhancement, were less sensitive and specific compared to direct signs. • A combination of the most specific direct signs of joint invasion showed best results with perfect specificity and PPV (both 100%) and excellent sensitivity and NPV (both 96 %).

Dysregulation of Wnt signaling in bone of type 2 diabetes mellitus and diabetic Charcot arthropathy.

BACKGROUND: Type 2 diabetes mellitus (T2DM) patients show a markedly higher fracture risk and impaired fracture healing when compared to non-diabetic patients. However in contrast to type 1 diabetes mellitus, bone mineral density in T2DM is known to be normal or even regionally elevated, also known as diabetic bone disease. Charcot arthropathy is a severe and challenging complication leading to bone destruction and mutilating bone deformities. Wnt signaling is involved in increasing bone mineral density, bone homeostasis and apoptotic processes. It has been shown that type 2 diabetes mellitus is strongly associated with gene variants of the Wnt signaling pathway, specifically polymorphisms of TCF7L2 (transcription factor 7 like 2), which is an effector transcription factor of this pathway. METHODS: Bone samples of 19 T2DM patients and 7 T2DM patients with additional Charcot arthropathy were compared to 19 non-diabetic controls. qPCR analysis for selected members of the Wnt-signaling pathway (WNT3A, WNT5A, catenin beta, TCF7L2) and bone gamma-carboxyglutamate (BGLAP, Osteocalcin) was performed and analyzed using the 2-ΔΔCt- Method. Statistical analysis comprised one-way analysis of variance (ANOVA). RESULTS: In T2DM patients who had developed Charcot arthropathy WNT3A and WNT5A gene expression was down-regulated by 89 and 58% compared to healthy controls (p < 0.0001). TCF7L2 gene expression showed a significant reduction by 63% (p < 0.0001) and 18% (p = 0.0136) in diabetic Charcot arthropathy. In all diabetic patients BGLAP (Osteocalcin) was significantly decreased by at least 59% (p = 0.0019). CONCLUSIONS: For the first time with this study downregulation of members of the Wnt-signaling pathway has been shown in the bone of diabetic patients with and without Charcot arthropathy. This may serve as future therapeutic target for this severe disease.

Treatment of vascular graft infections: gentamicin-coated ePTFE grafts reveals strong antibacterial properties in vitro.

Vascular graft infections (VGI) are severe complications in prosthetic vascular surgery with an incidence ranging from 1 to 6%. In these cases, synthetic grafts are commonly used in combination with antimicrobial agents. Expanded polytetrafluoroethylene (ePTFE) is in clinical use as a synthetic graft material and shows promising results by influencing bacterial adhesion. However, the literature on antibiotic-bound ePTFE grafts is scarce. Gentamicin is a frequently used antibiotic for local treatment of surgical site infections, but has not been evaluated as antimicrobial agent on ePTFE grafts. In this study, we examine the antimicrobial efficacy and biocompatibility of novel types of gentamicin-coated ePTFE grafts in vitro. ePTFE grafts coated with gentamicin salt formulations with covalently-bound palmitate were evaluated in two drug concentrations (GP1.75% and GP3.5%). To investigate effects from types of formulations, also suspensions of gentamicin in palmitate as well as polylactide were used at comparable levels (GS + PA and GS + R203). Antibacterial efficacies were estimated by employing a zone of inhibition, growth inhibition and bacterial adhesion assay against Staphylococcus aureus (SA). Cytotoxicity was determined with murine fibroblasts according to the ISO standard 10993-5. Gentamicin-coated ePTFE grafts show low bacterial adherence and strong antibacterial properties in vitro against SA. Bactericidal inhibition lasted until day 11. Highest biocompatibility was achieved using gentamicin palmitate GP1.75% coated ePTFE grafts. ePTFE grafts with gentamicin-coating are effective in vitro against SA growth and adherence. Most promising results regarding antimicrobial properties and biocompatibility were shown with chemically bounded gentamicin palmitate GP1.75% coatings. Graphical abstract.

Machine learning in knee arthroplasty: specific data are key-a systematic review.

PURPOSE: Artificial intelligence (AI) in healthcare is rapidly growing and offers novel options of data analysis. Machine learning (ML) represents a distinct application of AI, which is capable of generating predictions and has already been tested in different medical specialties with various approaches such as diagnostic applications, cost predictions or identification of risk factors. In orthopaedics, this technology has only recently been introduced and the literature on ML in knee arthroplasty is scarce. In this review, we aim to investigate which predictions are already feasible using ML models in knee arthroplasty to identify prerequisites for the effective use of this novel approach. For this reason, we conducted a systematic review of ML algorithms for outcome prediction in knee arthroplasty. METHODS: A comprehensive search of PubMed, Medline database and the Cochrane Library was conducted to find ML applications for knee arthroplasty. All relevant articles were systematically retrieved and evaluated by an orthopaedic surgeon and a data scientist on the basis of the PRISMA statement. The search strategy yielded 225 articles of which 19 were finally assessed as eligible. A modified Coleman Methodology Score (mCMS) was applied to account for a methodological evaluation. RESULTS: The studies presented in this review demonstrated fair to good results (AUC median 0.76/range 0.57-0.98), while heterogeneous prediction models were analysed: complications (6), costs (4), functional outcome (3), revision (2), postoperative satisfaction (2), surgical technique (1) and biomechanical properties (1) were investigated. The median mCMS was 65 (range 40-80) points. CONCLUSION: The prediction of distinct outcomes with ML models applying specific data is already feasible; however, the prediction of more complex outcomes is still inaccurate. Registry data on knee arthroplasty have not been fully analysed yet so that specific parameters have not been sufficiently evaluated. The inclusion of specific input data as well as the collaboration of orthopaedic surgeons and data scientists are essential prerequisites to fully utilize the capacity of ML in knee arthroplasty. Future studies should investigate prospective data with specific and longitudinally recorded parameters. LEVEL OF EVIDENCE: III.

Synovial calprotectin is a reliable biomarker for periprosthetic joint infections in acute-phase inflammation - a prospective cohort study.

PURPOSE: Diagnosing periprosthetic joint infections (PJI) are challenging and may be hampered by the presence of other causes of local inflammation. Conventional synovial and serum markers are not reliable under these circumstances. Synovial calprotectin has been recently shown as a promising biomarker for PJI in total hip (THA) and total knee arthroplasty (TKA). The aim of this study is to investigate if calprotectin is reliable for PJI diagnosis in cases with accompanying inflammation due to recent surgery, dislocation or implant breakage in primary and revision TKA and THA. METHODS: Thirty-three patients were included in this prospective study between July 2019 and October 2021 (17 patients undergoing surgery < 9 months, 11 dislocations, five implant breakage, respectively). Synovial white blood cell count (WBC), percentage of polymorphonuclear neutrophils (PMC), serum C-reactive protein (CRP) and synovial calprotectin, using a lateral-flow-assay, were analysed. These parameters were tested against a modified European-Bone-and-Joint-Infection-Society (EBJIS) definition with adjusted thresholds to account for the local inflammation. Statistic quality criteria were calculated and compared using a binary classification test. RESULTS: Seventeen patients were classified as confirmed infections according to the modified EBJIS definition (13 THA and 4 TKA). The calprotectin assay yielded a sensitivity of 0.88 (0.64, 0.99), a specificity of 0.81 (0.54, 0.96), a positive predictive value (PPV) of 0.83 (0.59, 0.96) and a negative predictive value (NPV) of 0.87 (0.60, 0.98). CONCLUSIONS: Even in the presence of local inflammation due to other, non-infectious causes, calprotectin is a reliable diagnostic parameter for the detection of a PJI in primary and revision THA and TKA.

Multitask Deep Learning for Segmentation and Classification of Primary Bone Tumors on Radiographs.

Background An artificial intelligence model that assesses primary bone tumors on radiographs may assist in the diagnostic workflow. Purpose To develop a multitask deep learning (DL) model for simultaneous bounding box placement, segmentation, and classification of primary bone tumors on radiographs. Materials and Methods This retrospective study analyzed bone tumors on radiographs acquired prior to treatment and obtained from patient data from January 2000 to June 2020. Benign or malignant bone tumors were diagnosed in all patients by using the histopathologic findings as the reference standard. By using split-sample validation, 70% of the patients were assigned to the training set, 15% were assigned to the validation set, and 15% were assigned to the test set. The final performance was evaluated on an external test set by using geographic validation, with accuracy, sensitivity, specificity, and 95% CIs being used for classification, the intersection over union (IoU) being used for bounding box placements, and the Dice score being used for segmentations. Results Radiographs from 934 patients (mean age, 33 years ± 19 [standard deviation]; 419 women) were evaluated in the internal data set, which included 667 benign bone tumors and 267 malignant bone tumors. Six hundred fifty-four patients were in the training set, 140 were in the validation set, and 140 were in the test set. One hundred eleven patients were in the external test set. The multitask DL model achieved 80.2% (89 of 111; 95% CI: 72.8, 87.6) accuracy, 62.9% (22 of 35; 95% CI: 47, 79) sensitivity, and 88.2% (67 of 76; CI: 81, 96) specificity in the classification of bone tumors as malignant or benign. The model achieved an IoU of 0.52 ± 0.34 for bounding box placements and a mean Dice score of 0.60 ± 0.37 for segmentations. The model accuracy was higher than that of two radiologic residents (71.2% and 64.9%; P = .002 and P < .001, respectively) and was comparable with that of two musculoskeletal fellowship-trained radiologists (83.8% and 82.9%; P = .13 and P = .25, respectively) in classifying a tumor as malignant or benign. Conclusion The developed multitask deep learning model allowed for accurate and simultaneous bounding box placement, segmentation, and classification of primary bone tumors on radiographs. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Carrino in this issue.