Despite well-defined criteria for radiographic diagnosis of atypical femur fractures (AFFs), missed and delayed diagnosis is common. An AFF diagnostic software could provide timely AFF detection to prevent progression of incomplete or development of contralateral AFFs. In this study, we investigated the ability for an artificial intelligence (AI)-based application, using deep learning models (DLMs), particularly convolutional neural networks (CNNs), to detect AFFs from femoral radiographs. A labelled Australian dataset of pre-operative complete AFF (cAFF), incomplete AFF (iAFF), typical femoral shaft fracture (TFF), and non-fractured femoral (NFF) X-ray images in anterior-posterior view were used for training (N = 213, 49, 394, 1359, respectively). An AFFnet model was developed using a pretrained (ImageNet dataset) ResNet-50 backbone, and a novel Box Attention Guide (BAG) module to guide the model's scanning patterns to enhance its learning. All images were used to train and internally test the model using a 5-fold cross validation approach, and further validated by an external dataset. External validation of the model's performance was conducted on a Sweden dataset comprising 733 TFF and 290 AFF images. Precision, sensitivity, specificity, F1-score and AUC were measured and compared between AFFnet and a global approach with ResNet-50. Excellent diagnostic performance was recorded in both models (all AUC >0.97), however AFFnet recorded lower number of prediction errors, and improved sensitivity, F1-score and precision compared to ResNet-50 in both internal and external testing. Sensitivity in the detection of iAFF was higher for AFFnet than ResNet-50 (82 % vs 56 %). In conclusion, AFFnet achieved excellent diagnostic performance on internal and external validation, which was superior to a pre-existing model. Accurate AI-based AFF diagnostic software has the potential to improve AFF diagnosis, reduce radiologist error, and allow urgent intervention, thus improving patient outcomes.

Atypical femur fractures (AFF) represent a very rare type of fracture that can be difficult to discriminate radiologically from normal femur fractures (NFF). AFFs are associated with drugs that are administered to prevent osteoporosis-related fragility fractures, which are highly prevalent in the elderly population. Given that these fractures are rare and the radiologic changes are subtle currently only 7% of AFFs are correctly identified, which hinders adequate treatment for most patients with AFF. Deep learning models could be trained to classify automatically a fracture as AFF or NFF, thereby assisting radiologists in detecting these rare fractures. Historically, for this classification task, only imaging data have been used, using convolutional neural networks (CNN) or vision transformers applied to radiographs. However, to mimic situations in which all available data are used to arrive at a diagnosis, we adopted an approach of deep learning that is based on the integration of image data and tabular data (from electronic health records) for 159 patients with AFF and 914 patients with NFF. We hypothesized that the combinatorial data, compiled from all the radiology departments of 72 hospitals in Sweden and the Swedish National Patient Register, would improve classification accuracy, as compared to using only one modality. At the patient level, the area under the ROC curve (AUC) increased from 0.966 to 0.987 when using the integrated set of imaging data and seven pre-selected variables, as compared to only using imaging data. More importantly, the sensitivity increased from 0.796 to 0.903. We found a greater impact of data fusion when only a randomly selected subset of available images was used to make the image and tabular data more balanced for each patient. The AUC then increased from 0.949 to 0.984, and the sensitivity increased from 0.727 to 0.849.

These AUC improvements are not large, mainly because of the already excellent performance of the CNN (AUC of 0.966) when only images are used. However, the improvement is clinically highly relevant considering the importance of accuracy in medical diagnostics. We expect an even greater effect when imaging data from a clinical workflow, comprising a more diverse set of diagnostic images, are used.

Background: Operative treatment is widely used for acute proximal hamstring avulsions, but its effectiveness compared with that of nonoperative treatment has not been shown in randomized trials.

Methods: In this noninferiority trial at 10 centers in Sweden and Norway, we enrolled patients 30 to 70 years of age with a proximal hamstring avulsion in a randomized trial and a parallel observational cohort. Treatments were operative reinsertion of the tendons or nonoperative management. The primary end point was the Perth Hamstring Assessment Tool (PHAT) at 2 years of follow-up. Secondary outcomes included scores on the Lower Extremity Functional Scale (LEFS).

Results: A total of 119 patients were enrolled in the randomized trial and 97 patients in the observational cohort. In the per-protocol analysis of the randomized trial, the mean (±standard deviation) PHAT scores were 79.9±19.5 and 78.5±19.4 in the operative and nonoperative groups, respectively (PHAT scores range from 0 to 100, with higher scores indicating higher function). The prespecified noninferiority limit of 10 points was not crossed (mean difference, -1.2; 95% confidence interval [CI], -8.6 to 6.2; P=0.009 for noninferiority). Analyses of secondary outcomes, including a mean difference in the LEFS score of -1.6 (95% CI, -5.2 to 2.0), aligned with the primary outcome. The observed numbers of adverse events in the randomized trial were nine in the operative group versus three in the nonoperative group (odds ratio, 0.3; 95% CI, 0.1 to 1.2). In the analysis of the observational cohort, the mean PHAT score difference between the nonoperative and operative treatment groups was -2.6 (95% CI, -9.9 to 4.6).

Conclusions: In patients 30 to 70 years of age with proximal hamstring avulsions, nonoperative treatment was noninferior to operative treatment. (Funded by Afa Försäkring and others; ClinicalTrials.gov number, NCT03311997.).

Background: Early tensile loading improves material properties of healing Achilles tendon ruptures in animal models and in surgically treated human ruptures. However, the effect of such rehabilitation in patients who are nonsurgically treated remains unknown. Hypothesis: In nonsurgically treated Achilles tendon ruptures, early tensile loading would lead to higher elastic modulus 19 weeks after the injury compared with controls. Study Design: Randomized controlled trial; Level of evidence, 2. Methods: Between October 2015 and November 2018, a total of 40 nonsurgically treated patients with acute Achilles tendon rupture were randomized to an early tensile loading (loaded group) or control group. Tantalum bead markers were inserted percutaneously into the tendon stumps 2 weeks after the injury to allow high-precision measurements of callus deformation under mechanical testing. The loaded group used a training pedal twice daily to produce a gradual increase in tensile load during the following 5 weeks. Both groups were allowed full weightbearing in an ankle orthosis and unloaded range of motion exercises. Patients were followed clinically and via roentgen stereophotogrammetric analysis and computed tomography at 7, 19, and 52 weeks after the injury. Results: The mean +/- standard deviation elastic modulus at 19 weeks was 95.6 +/- 38.2 MPa in the loaded group and 108 +/- 45.2 MPa in controls (P = .37). The elastic modulus increased in both groups, although it was lower in the loaded group at all time points. Tendon cross-sectional area increased from 7 weeks to 19 weeks, from 231 +/- 99.5 to 388 +/- 142 mm(2) in the loaded group and from 188 +/- 65.4 to 335 +/- 87.2 mm(2) in controls (P < .001 for the effect of time). Cross-sectional area for the loaded group versus controls at 52 weeks was 302 +/- 62.4 mm(2) versus 252 +/- 49.2 mm(2), respectively (P = .03). Gap elongation was 7.35 +/- 13.9 mm in the loaded group versus 2.86 +/- 5.52 mm in controls (P = .27). Conclusion: Early tensile loading in nonsurgically treated Achilles tendon ruptures did not lead to higher elastic modulus in the healing tendon but altered the structural properties of the tendon via an increased tendon thickness. Registration: NCT0280575 (ClinicalTrials.gov identifier).

Bisphosphonates (BPs)-associated atypical femoral fractures (AFFs) present with impaired fracture healing, yet the underlying mechanism is unclear, which prevents the development of effective therapy. Peripheral sensory nerve has been shown to regulate fracture healing via releasing neuropeptides. Here we show that long-term BPs pre-treatment leads to fracture non-union in rats, characterized by reduced expression of calcitonin gene-related peptide (CGRP, a predominant type of neuropeptides) and abundant fibrous tissues in the non-bridged fracture gap, mimicking clinical AFFs. By using single-cell RNA-sequencing, long-term BPs treatment was identified to promote transition of progenitor cells into a specific cluster of fibroblasts that actively deposit dense extracellular matrix (ECM) to prevent fracture callus bridging. Administration of exogenous CGRP at early stages of fracture repair, in contrast, eliminates the ECM-secreting fibroblast cluster, attenuates fibrogenesis, and facilitates callus bridging, suggesting CGRP is a promising agent to facilitate AFF healing. Accordingly, we have developed an innovative magnesium (Mg) containing hybrid intramedullary nail fixation system (Mg-IMN) to effectively rescue BPs-impaired fracture healing via elevating CGRP synthesis and release. Such device optimizes the fracture healing in BPs-pretreated rats, comparable to direct administration of CGRP. These findings address the indispensable role of CGRP in advancing the healing of AFFs and develop translational strategies to accelerate AFF healing by taking advantage of the CGRP-stimulating effect of Mg-based biodegradable orthopedic implant. The study also indicates fibrosis could be targeted by augmenting CGRP expression to accelerate fracture healing even under challenging scenarios where fibroblasts are aberrantly activated.

Background and purpose Evidence guiding the decision on whether to treat proximal hamstring tendon avulsions (PHA) operatively or non-operatively is very limited. The aim of this study was to identify the current practices and the rationale behind PHA treatment decisions in the Nordic countries. Methods A survey was sent to orthopaedic surgeons in Sweden, Norway, Finland and Denmark. The study population consisted of responding surgeons with exposure to surgical treatment of PHA (n = 125). The questions covered surgeon and unit characteristics, and surgeons understanding of the evidence for treatment, and they explored which patient and injury factors influence treatment allocation. Results Although some surgeons indicated a preference for one of the treatments, 84% stated that the treatment decision was based on patient and injury-related factors. Severe obesity, drug abuse, a sedentary lifestyle, age > 60 years and delayed diagnosis (> 6 weeks) were considered contraindications to surgical treatment. Also, there was agreement that patients expressing a preference for non-operative treatment should not be operated. Complete avulsions with tendon dislocation >= 2-3 cm on MRI were relative indications for surgical treatment. The majority of surgeons did not believe that operatively treated patients did better than non-operatively treated patients and experienced that patients, generally, were satisfied with the treatment result, regardless of the type of treatment. Most surgeons had experienced significant complications to operative treatment. Conclusion Current practices varied among different units, and despite the lack of evidence for their prognostic value, several factors were inconsistently being used as decision modifiers when selecting patients for surgical treatment.

Background and purpose - To continuously assess the incidence of atypical femoral fractures (AFFs) in the population is important, to allow the evaluation of the risks and benefits associated with osteoporosis treatment. Therefore, we investigated the possibility to use the Swedish Fracture Register (SFR) as a surveillance tool for AFFs in the population and to explore means of improvement. Patients and methods - All AFF registrations in the SFR from January 1, 2015 to December 31, 2018 were enrolled in the study. For these patients, radiographs were obtained and combined with radiographs from 176 patients with normal femoral fractures, to form the study cohort. All images were reviewed and classified into AFFs or normal femur fractures by 2 experts in the field (gold-standard classification) and 1 orthopedic resident educated on the specific radiographic features of AFF (educated-user classification). Furthermore, we estimated the incidence rate of AFFs in the population captured by the register through comparison with a previous cohort and calculated the positive predictive value (PPV) and, where possible, the inter-observer agreement (Cohen's kappa) between the different classifications. Results - Of the 178 available patients with AFF in the SFR, 104 patients were classified as AFF using the goldstandard classification, and 89 using the educated-user classification. The PPV increased from 0.58 in the SFR classification to 0.93 in the educated-user classification. The interobserver agreement between the gold-standard classification and the educated-user classification was 0.81. Interpretation - With a positive predictive value of 0.58 the Swedish Fracture Register outperforms radiology reports and reports to the Swedish Medical Products Agency on adverse drug reactions as a diagnostic tool to identify atypical femoral fractures.

Supraphysiological loading induced by unstable orthopedic implants initiates osteoclast formation, which results in bone degradation. We aimed to investigate which mechanosensitive cells in the peri-implant environment produce osteoclast-stimulating factors and how the production of these factors is stimulated by supraphysiological loading. The release of osteoclast-stimulating factors by different types of isolated bone marrow-derived hematopoietic and mesenchymal stem cells from six osteoarthritic patients was analyzed after one hour of supraphysiological loading (3.0 ± 0.2 Pa, 1 Hz) by adding their conditioned medium to osteoclast precursors. Monocytes produced factors that enhanced osteoclastogenesis by 1.6 ± 0.07-fold and mesenchymal stem cells by 1.4 ± 0.07-fold. Medium from osteoprogenitors and pre-osteoblasts enhanced osteoclastogenesis by 1.3 ± 0.09-fold and 1.4 ± 0.03-fold, respectively, where medium from four patients elicited a response and two did not. Next generation sequencing analysis of osteoprogenitors revealed that genes encoding for inflammation-related pathways and cytoskeletal rearrangements were regulated differently between responders and non-responders. Our data suggest that released osteoclast-stimulating soluble factors by progenitor cells in the bone marrow after supraphysiological loading may be related to cytoskeletal arrangement in an inflammatory environment. This connection could be relevant to better understand the aseptic loosening process of orthopedic implants.

Background and purpose - A correct diagnosis is essential for the appropriate treatment of patients with atypical femoral fractures (AFFs). The diagnostic accuracy of radiographs with standard radiology reports is very poor. We derived a diagnostic algorithm that uses deep neural networks to enable clinicians to discriminate AFFs from normal femur fractures (NFFs) on conventional radiographs. Patients and methods - We entered 433 radiographs from 149 patients with complete AFF and 549 radiographs from 224 patients with NFF into a convolutional neural network (CNN) that acts as a core classifier in an automated pathway and a manual intervention pathway (manual improvement of image orientation). We tested several deep neural network structures (i.e., VGG19, InceptionV3, and ResNet) to identify the network with the highest diagnostic accuracy for distinguishing AFF from NFF. We applied a transfer learning technique and used 5-fold cross-validation and class activation mapping to evaluate the diagnostic accuracy.Results - In the automated pathway, ResNet50 had the highest diagnostic accuracy, with a mean of 91% (SD 1.3), as compared with 83% (SD 1.6) for VGG19, and 89% (SD 2.5) for InceptionV3. The corresponding accuracy levels for the intervention pathway were 94% (SD 2.0), 92% (2.7), and 93% (3.7), respectively. With regards to sensitivity and specificity, ResNet outperformed the other networks with a mean AUC (area under the curve) value of 0.94 (SD 0.01) and surpassed the accuracy of clinical diagnostics.Interpretation - Artificial intelligence systems show excellent diagnostic accuracies for the rare fracture type of AFF in an experimental setting.