Aims. To examine whether natural language processing (NLP) using a clinically based large language model (LLM) could be used to predict patient selection for total hip or total knee arthroplasty (THA/TKA) from routinely available free-text radiology reports. Methods. Data pre-processing and analyses were conducted according to the Artificial intelligence to Revolutionize the patient Care pathway in Hip and knEe aRthroplastY (ARCHERY) project protocol. This included use of de-identified Scottish regional clinical data of patients referred for consideration of THA/TKA, held in a secure data environment designed for artificial intelligence (AI) inference. Only preoperative radiology reports were included. NLP algorithms were based on the freely available GatorTron model, a LLM trained on over 82 billion words of de-identified clinical text. Two inference tasks were performed: assessment after model-fine tuning (50 Epochs and three cycles of k-fold cross validation), and external validation. Results. For THA, there were 5,558 patient radiology reports included, of which 4,137 were used for model training and testing, and 1,421 for external validation. Following training, model performance demonstrated average (mean across three folds) accuracy, F1 score, and area under the receiver operating curve (AUROC) values of 0.850 (95% confidence interval (CI) 0.833 to 0.867), 0.813 (95% CI 0.785 to 0.841), and 0.847 (95% CI 0.822 to 0.872), respectively. For TKA, 7,457 patient radiology reports were included, with 3,478 used for model training and testing, and 3,152 for external validation. Performance metrics included accuracy, F1 score, and AUROC values of 0.757 (95% CI 0.702 to 0.811), 0.543 (95% CI 0.479 to 0.607), and 0.717 (95% CI 0.657 to 0.778) respectively. There was a notable deterioration in performance on external validation in both cohorts. Conclusion. The use of routinely available preoperative radiology reports provides promising potential to help screen suitable candidates for THA, but not for TKA. The external validation results demonstrate the importance of further model testing and training when confronted with new clinical cohorts. Cite this article: Bone Joint J 2024;106-B(7):688–695

Aims. Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are common orthopaedic procedures requiring postoperative radiographs to confirm implant positioning and identify complications. Artificial intelligence (AI)-based image analysis has the potential to automate this postoperative surveillance. The aim of this study was to prepare a scoping review to investigate how AI is being used in the analysis of radiographs following THA and TKA, and how accurate these tools are. Methods. The Embase, MEDLINE, and PubMed libraries were systematically searched to identify relevant articles. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews and Arksey and O’Malley framework were followed. Study quality was assessed using a modified Methodological Index for Non-Randomized Studies tool. AI performance was reported using either the area under the curve (AUC) or accuracy. Results. Of the 455 studies identified, only 12 were suitable for inclusion. Nine reported implant identification and three described predicting risk of implant failure. Of the 12, three studies compared AI performance with orthopaedic surgeons. AI-based implant identification achieved AUC 0.992 to 1, and most algorithms reported an accuracy > 90%, using 550 to 320,000 training radiographs. AI prediction of dislocation risk post-THA, determined after five-year follow-up, was satisfactory (AUC 76.67; 8,500 training radiographs). Diagnosis of hip implant loosening was good (accuracy 88.3%; 420 training radiographs) and measurement of postoperative acetabular angles was comparable to humans (mean absolute difference 1.35° to 1.39°). However, 11 of the 12 studies had several methodological limitations introducing a high risk of bias. None of the studies were externally validated. Conclusion. These studies show that AI is promising. While it already has the ability to analyze images with significant precision, there is currently insufficient high-level evidence to support its widespread clinical use. Further research to design robust studies that follow standard reporting guidelines should be encouraged to develop AI models that could be easily translated into real-world conditions. Cite this article: Bone Joint J 2022;104-B(8):929–937

The Unified Classification System (UCS) was introduced because of a growing need to have a standardised universal classification system of periprosthetic fractures. It combines and simplifies many existing classification systems, and can be applied to any fracture around any partial or total joint replacement occurring during or after operation. Our goal was to assess the inter- and intra-observer reliability of the UCS in association with knee replacement when classifying fractures affecting one or more of the femur, tibia or patella. We used an international panel of ten orthopaedic surgeons with subspecialty fellowship training and expertise in adult hip and knee reconstruction (‘experts’) and ten residents of orthopaedic surgery in the last two years of training (‘pre-experts’). They each received 15 radiographs for evaluation. After six weeks they evaluated the same radiographs again but in a different order. . The reliability was assessed using the Kappa and weighted Kappa values. The Kappa values for inter-observer reliability for the experts and the pre-experts were 0.741 (95% confidence interval (CI) 0.707 to 0.774) and 0.765 (95% CI 0.733 to 0.797), respectively. The weighted Kappa values for intra-observer reliability for the experts and pre-experts were 0.898 (95% CI 0.846 to 0.950) and 0.878 (95% CI 0.815 to 0.942) respectively. The UCS has substantial inter-observer reliability and ‘near perfect’ intra-observer reliability when used for periprosthetic fractures in association with knee replacement in the hands of experienced and inexperienced users. Cite this article: Bone Joint J 2014;96-B:1669–73

Aims

A substantial fraction of patients undergoing knee arthroplasty (KA) or hip arthroplasty (HA) do not achieve an improvement as high as the minimal clinically important difference (MCID), i.e. do not achieve a meaningful improvement. Using three patient-reported outcome measures (PROMs), our aim was: 1) to assess machine learning (ML), the simple pre-surgery PROM score, and logistic-regression (LR)-derived performance in their prediction of whether patients undergoing HA or KA achieve an improvement as high or higher than a calculated MCID; and 2) to test whether ML is able to outperform LR or pre-surgery PROM scores in predictive performance.

Balancing service provision and surgical training is a challenging issue that affects all healthcare systems. A multicentre prospective study of 1501 total hip replacements was undertaken to investigate whether there is an association between surgical outcome and the grade of the operating surgeon, and whether there is any difference in outcome if surgeons’ assistants assist with the operation, rather than orthopaedic trainees. The primary outcome measure was the change in the Oxford hip score (OHS) at five years. Secondary outcomes included the rate of revision and dislocation, operating time, and length of hospital stay. There was no significant difference in ΔOHS or complication rates between operations undertaken by trainers and trainees, or those at which surgeons’ assistants and trainees were the assistant. However, there was a significant difference in the duration of surgery, with a mean reduction of 28 minutes in those in which a surgeons’ assistant was the assistant. This study provides evidence that total hip replacements can be performed safely and effectively by appropriately trained surgeons in training, and that there are potential benefits of using surgeons’ assistants in orthopaedic surgery

Aims

To review the evidence and reach consensus on recommendations for follow-up after total hip and knee arthroplasty.

Aims

Gram-negative periprosthetic joint infection (PJI) has been poorly studied despite its rapidly increasing incidence. Treatment with one-stage revision using intra-articular (IA) infusion of antibiotics may offer a reasonable alternative with a distinct advantage of providing a means of delivering the drug in high concentrations. Carbapenems are regarded as the last line of defense against severe Gram-negative or polymicrobial infection. This study presents the results of one-stage revision using intra-articular carbapenem infusion for treating Gram-negative PJI, and analyzes the characteristics of bacteria distribution and drug sensitivity.

Aims

The COVID-19 pandemic has disrupted the provision of arthroplasty services in England, Wales, and Northern Ireland. This study aimed to quantify the backlog, analyze national trends, and predict time to recovery.

Aims

Revision total knee arthroplasty (rTKA) and revision total hip arthroplasty (rTHA) are complex procedures with higher rates of re-revision, complications, and mortality compared to primary TKA and THA. We report the effects of the establishment of a revision arthroplasty network (the East Midlands Specialist Orthopaedic Network; EMSON) on outcomes of rTKA and rTHA.

Aims

To determine the trajectories of patient reported pain and functional disability over five years following total hip arthroplasty (THA) or total knee arthroplasty (TKA).

Aims

This study aimed to develop a virtual clinic for the purpose of reducing face-to-face orthopaedic consultations.

Aims

The aim of this study was to give estimates of the incidence of component incompatibility in hip and knee arthroplasty and to test the effect of an online, real-time compatibility check.

Aims

Increasing demand for total hip and knee arthroplasty (THA/TKA) and associated follow-up has placed huge demands on orthopaedic services. Feasible follow-up mechanisms are therefore essential.

Objectives

We wanted to investigate regional variations in the organisms reported to be causing peri-prosthetic infections and to report on prophylaxis regimens currently in use across England.

Objectives

Wear debris released from bearing surfaces has been shown to provoke negative immune responses in the recipient. Excessive wear has been linked to early failure of prostheses. Analysis using coordinate measuring machines (CMMs) can provide estimates of total volumetric material loss of explanted prostheses and can help to understand device failure. The accuracy of volumetric testing has been debated, with some investigators stating that only protocols involving hundreds of thousands of measurement points are sufficient. We looked to examine this assumption and to apply the findings to the clinical arena.

Between 1988 and 1998 we implanted 318 total hip replacements (THRs) in 287 patients using the Plasmacup (B. Braun Ltd, Sheffield, United Kingdom) and a conventional metal-on-polyethylene articulation. The main indications for THR were primary or secondary osteoarthritis.

At follow-up after a mean 11.6 years (7.6 to 18.4) 17 patients had died and 20 could not be traced leaving a final series of 280 THRs in 250 patients. There were 62 revisions (22.1%) in 59 patients. A total of 43 acetabular shells (15.4%) had been revised and 13 (4.6%) had undergone exchange of the liner. The most frequent indications for revision were osteolysis and aseptic loosening, followed by polyethylene wear. The mean Kaplan-Meier survival of the Plasmacup was 91% at ten years and 58% at 14 years. Osteolysis was found around 36 (17.1%) of the 211 surviving shells. The median annual rate of linear wear in the surviving shells was 0.12 mm/year and 0.25 mm/year in those which had been revised (p < 0.001). Polyethylene wear was a strong independent risk factor for osteolysis and aseptic loosening. The percentage of patients with osteolysis increased proportionately with each quintile of wear-rate.

There is a high late rate of failure of the Plasmacup. Patients with the combination of this prosthesis and bearing should be closely monitored after ten years.

We have investigated the accuracy of the templating of digital radiographs in planning total hip replacement using two common object-based calibration methods with the ball placed laterally (method 1) or medially (method 2) and compared them with two non-object-based methods. The latter comprised the application of a fixed magnification of 121% (method 3) and calculation of magnification based on the object-film-distance (method 4). We studied the post-operative radiographs of 57 patients (19 men, 38 women, mean age 73 years (53 to 89)) using the measured diameter of the prosthetic femoral head and comparing it with the true value.

Both object-based methods (1 and 2) produced large errors (mean/maximum: 2.55%/17.4% and 2.04%/6.46%, respectively). Method 3 applying a fixed magnification and method 4 (object-film-distance) produced smaller errors (mean/maximum 1.42%/5.22% and 1.57%/4.24%, respectively; p < 0.01). The latter results were clinically relevant and acceptable when planning was allowed to within one implant size. Object-based calibration (methods 1 and 2) has fundamental problems with the correct placement of the calibration ball. The accuracy of the fixed magnification (method 3) matched that of object-film-distance (method 4) and was the most reliable and efficient calibration method in digital templating.

We have developed a novel method of calculating the radiological magnification of the hip using two separate radio-opaque markers. We recruited 74 patients undergoing radiological assessment following total hip replacement. Both the new double marker and a conventional single marker were used by the radiographer at the time of x-ray. The predicted magnification according to each marker was calculated, as was the true radiological magnification of the components. The correlation between true and predicted magnification was good using the double marker (r = 0.90, n = 74, p < 0.001), but only moderate for the single marker (r = 0.50, n = 63, p < 0.001). The median error was significantly less for the double marker than for the single (1.1% vs 4.8%, p < 0.001). The double marker method demonstrated excellent validity (intraclass correlation coefficient = 0.89), in contrast to the single marker (0.32).

The double marker method appears to be superior to the single marker method when used in the clinical environment.

We wished to estimate the incidence of surgical-site infection (SSI) after total hip replacement (THR) and hemiarthroplasty and its strength of association with major risk factors. The SSI surveillance service prospectively gathered clinical, operative and infection data on inpatients from 102 hospitals in England during a four-year period.

The overall incidence of SSI was 2.23% for 16 291 THRs, 4.97% for 5769 hemiarthroplasty procedures, 3.68% for 2550 revision THRs and 7.6% for 198 revision hemiarthroplasties. Staphylococcus aureus was identified in 50% of SSIs; 59% of these isolates were methicillin-resistant (MRSA). In the single variable analysis of THRs, age, female gender, American Society of Anesthesiologists (ASA) score, body mass index, trauma, duration of operation and pre-operative stay were significantly associated with the risk of SSI (p < 0.05). For hemiarthroplasty, the ASA score and age were significant factors. In revision THRs male gender, ASA score, trauma, wound class, duration of operation and pre-operative stay were significant risk factors. The median time to detection of SSI was eight days for superficial incisional, 11 days for deep incisional and 11 days for joint/bone infections. For each procedure the mean length of stay doubled for patients with SSI. The multivariate analysis identified age group, trauma, duration of operation and ASA score as significant, independent risk factors for SSI. There was significant interhospital variation in the rates of SSI. MRSA was the most common pathogen to cause SSI in hip arthroplasty, especially in patients undergoing hemiarthroplasty, but coagulase-negative Staph. aureus may be more important in deep infections involving the joint.