Despite the vast quantities of published artificial intelligence (AI) algorithms that target trauma and orthopaedic applications, very few progress to inform clinical practice. One key reason for this is the lack of a clear pathway from development to deployment. In order to assist with this process, we have developed the Clinical Practice Integration of Artificial Intelligence (CPI-AI) framework – a five-stage approach to the clinical practice adoption of AI in the setting of trauma and orthopaedics, based on the IDEAL principles (https://www.ideal-collaboration.net/). Adherence to the framework would provide a robust evidence-based mechanism for developing trust in AI applications, where the underlying algorithms are unlikely to be fully understood by clinical teams.

Cite this article: Bone Joint Res 2024;13(9):507–512.

Total hip arthroplasty (THA) is often performed in symptomatic patients with developmental dysplasia of the hip (DDH) who do not qualify for periacetabular osteotomy (PAO). The impact of osteoarthritis (OA) severity on postoperative outcomes in DDH patients who undergo THA is not well described. We hypothesized that DDH patients who undergo THA with mild OA have slower initial recovery postoperatively, but similar one-year patient reported outcome measures (PROMs) when compared to DDH patients with severe OA.

We performed a retrospective review at a single academic institution over a six-year period of patients with DDH who underwent primary THA and compared them to patients without DDH who underwent THA. Within the DDH cohort, we compared PROMs stratified by DDH severity and OA severity. Diagnosis of DDH was verified using radiographic lateral center edge angle (LCEA). Minimum one-year follow-up was required. PROMs were collected through one-year postoperatively. Logistic and linear regression models were used adjusting for age, sex, body mass index, and Charlson Comorbidity Index.

263 patients with DDH were compared to 1,225 THA patients without DDH. No significant differences were found in postoperative PROMs or revision rates (p=0.49). When stratified by DDH severity, patients with LCEA<10° had worse preoperative pain (p=0.01), mental health (p<0.01) and physical function (p=0.03) scores but no significant difference in postoperative PROMs. Within the DDH group, when stratified by OA severity, patients with Grade 3 Tonnis score had worse preoperative pain (p=0.04) but no significant difference in postoperative PROMs. Recovery curves in DDH patients based upon severity of DDH and OA were not significantly different at 2-weeks, 6-weeks and 1-year.

DDH patients who have mild OA have similar recovery curves compared to those with severe OA. THA is reasonable in symptomatic DDH patients who have mild arthritis and do not qualify for PAO.

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.

Introduction

The Te Whatu Ora Southern catchment area covers the largest geographical region in New Zealand (over 62,000 km2) creating logistical challenges in providing timely access to emergency neck of femur (NOF) fracture surgery. Current Australian and New Zealand guidelines recommend that NoF surgery be performed within 48 hours of presentation. The purpose of this study was to compare the outcomes for patients with NoF fractures who present directly to a referral hospital (Southland Hospital) compared to those are transferred from rural peripheral centres.

Introduction

AlloStem/Cellular Bone Allograft and autologous bone graft are accepted methods for managing hindfoot degenerative arthritis. The purpose was to evaluate outcomes of AlloStem and autograft in subtalar arthrodesis and compare overall fusion rates.

Dual Mobility (DM) Total Hip Replacements (THRs), are becoming widely used but function in-vivo is not fully understood.

The aim of this study was to compare the incidence of impingement of a modular dual mobility with that of a standard cup.

A geometrical model of one subject's bony anatomy \[1\] was developed, a THR was implanted with the cup at a range of inclination and anteversion positions (Corail® stem, Pinnacle® cup (DePuy Synthes)). Two DM variants and one STD acetabular cup were modelled. Joint motions were taken from kinematic data of activities of daily living associated with dislocation \[2\] and walking. The occurrence of impingement was assessed for each component combination, orientation and activity. Implant-implant impingement can occur between the femoral neck and the metal or PE liner (DM or STD constructs respectively) or neck-PE mobile liner (DM only).

The results comprise a colour coded matrix which sums the number of impingement events for each cup position and activity and for each implant variant.

Neck-PE mobile liner impingement, occurred for both DM sizes, for all activities, and most cup placement positions indicating that the PE mobile liner is likely to move at the start of all activities including walking.

For all constructs no placement positions avoided neck-metal (DM) or neck-PE liner (STD) impingementevents in all activities. The least number of events occurred at higher inclination and anteversion component positions. In addition to implant-implant impingement, some instances of bone-bone and implant-bone impingement were also observed.

Consistent with DM philosophy, neck-PE mobile liner impingement and liner motion occurred for all activities including walking. Neck-liner impingement frequency was comparable between both DM sizes (metal liner) and a standard cup (PE liner).

To examine whether Natural Language Processing (NLP) using a state-of-the-art clinically based Large Language Model (LLM) could predict patient selection for Total Hip Arthroplasty (THA), across a range of routinely available clinical text sources.

Data pre-processing and analyses were conducted according to the Ai to Revolutionise the patient Care pathway in Hip and Knee arthroplasty (ARCHERY) project protocol (https://www.researchprotocols.org/2022/5/e37092/). Three types of deidentified Scottish regional clinical free text data were assessed: Referral letters, radiology reports and clinic letters. NLP algorithms were based on the GatorTron model, a Bidirectional Encoder Representations from Transformers (BERT) based LLM trained on 82 billion words of de-identified clinical text. Three specific inference tasks were performed: assessment of the base GatorTron model, assessment after model-fine tuning, and external validation.

There were 3911, 1621 and 1503 patient text documents included from the sources of referral letters, radiology reports and clinic letters respectively. All letter sources displayed significant class imbalance, with only 15.8%, 24.9%, and 5.9% of patients linked to the respective text source documentation having undergone surgery. Untrained model performance was poor, with F1 scores (harmonic mean of precision and recall) of 0.02, 0.38 and 0.09 respectively. This did however improve with model training, with mean scores (range) of 0.39 (0.31–0.47), 0.57 (0.48–0.63) and 0.32 (0.28–0.39) across the 5 folds of cross-validation. Performance deteriorated on external validation across all three groups but remained highest for the radiology report cohort.

Even with further training on a large cohort of routinely collected free-text data a clinical LLM fails to adequately perform clinical inference in NLP tasks regarding identification of those selected to undergo THA. This likely relates to the complexity and heterogeneity of free-text information and the way that patients are determined to be surgical candidates.

Aims

Periacetabular osteotomy (PAO) is the preferred treatment for symptomatic acetabular dysplasia in adolescents and young adults. There remains a lack of consensus regarding whether intra-articular procedures such as labral repair or improvement of femoral offset should be performed at the time of PAO or addressed subsequent to PAO if symptoms warrant. The purpose was to determine the rate of subsequent hip arthroscopy (HA) in a contemporary cohort of patients, who underwent PAO in isolation without any intra-articular procedures.

Aims

The aim of this study was to produce clinical consensus recommendations about the non-surgical treatment of children with Perthes’ disease. The recommendations are intended to support clinical practice in a condition for which there is no robust evidence to guide optimal care.

Aims

The mean age of patients undergoing total knee arthroplasty (TKA) has reduced with time. Younger patients have increased expectations following TKA. Aseptic loosening of the tibial component is the most common cause of failure of TKA in the UK. Interest in cementless TKA has re-emerged due to its encouraging results in the younger patient population. We review a large series of tantalum trabecular metal cementless implants in patients who are at the highest risk of revision surgery.

Aims

Impaired fracture repair in patients with type 2 diabetes mellitus (T2DM) is not fully understood. In this study, we aimed to characterize the local changes in gene expression (GE) associated with diabetic fracture. We used an unbiased approach to compare GE in the fracture callus of Zucker diabetic fatty (ZDF) rats relative to wild-type (WT) littermates at three weeks following femoral osteotomy.

Aims

The aim of this study was to investigate the agreement in interpretation of the quality of the paediatric hip ultrasound examination, the reliability of geometric and morphological assessment, and the relationship between these measurements.

Abstract

Abstract

Periacetabular osteotomy (PAO) is the preferred treatment for symptomatic acetabular dysplasia in adolescents and young adults. There remains a lack of consensus regarding whether intra-articular work such as labral repair or improvement of femoral offset should be performed at the time of PAO or addressed subsequent to PAO if symptoms warrant. The purpose was to determine the rate of subsequent hip arthroscopy (HA) in a contemporary PAO cohort with no intra-articular work performed at the time of PAO.

From June 2012 to March 2022, 368 rectus sparing PAOs were performed and followed for a minimum of one year (mean 5.9 years). The average age was 24 (range 14–46) and 89% were female. Patients were evaluated at last follow-up for patient-reported outcomes (PROMs). Clinical records were reviewed for complications or subsequent surgery. Radiographs were reviewed for the following acetabular parameters: LCEA, ACEA, AI, and the alpha-angle (AA). Patients were cross-referenced from the two largest hospital systems in our area to determine if subsequent HA was performed. Descriptive statistics were used to analyze risk factors for HA.

16 hips (15 patients) (4.4%) underwent subsequent HA with labral repair and femoral osteochondroplasty most common. For those with a minimum of two years of follow-up, 5.3% underwent subsequent HA. No hips underwent THA; one revision PAO was performed. 14 hips experienced a complication and 99 underwent hardware removal. All PROMs improved significantly post-operatively. Radiographically 80% of hips were in goal for acetabular correction parameters with no significant differences between those who underwent subsequent HA and those who did not.

Rectus sparing PAO is associated with a low rate of subsequent HA for intra-articular pathology at 5-year follow-up. Acetabular correction alone may be sufficient as the primary intervention for the majority of patients with symptomatic acetabular dysplasia.

Intertrochanteric fractures are common, accounting for nearly 30% of all fracture related admissions. Some have suggested that these fractures should be treated in community hospitals so as not to tax the resources of Level One trauma centers. Since many factors predictive of fixation failure are related to technical aspects of the surgery, the purpose of this study was to compare radiographic parameters after fixation comparing trauma fellowship trained surgeons to non-fellowship trained community surgeons to see if these fractures can be treated successfully in either setting.

Using our hospital system's trauma database, we identified 100 consecutive patients treated with cephalomedullary nails by traumatologists, and 100 consecutive patients treated by community surgeons. Quality of reduction, neck shaft angle (NSA), tip-to apex distance (TAD) were compared.

The mean TAD for the trauma group was 10mm compared to 21mm for the community group (p<0.001). The mean postoperative NSA for the trauma group was 133 degrees compared to 127 degrees for the community group (p<0.001). The mean difference in the NSA of the fractured side compared to the normal hip was 2.5 degrees of valgus in the trauma group compared to 5 degrees of varus for the community group (p<0.001). There were 93 good reductions in the trauma group compared to 19 in the community group (p<0.001). There were no poor reductions in the trauma group and 49 poor reductions in the community group (p<0.001).

Fellowship trained traumatologists achieved significantly more accurate reductions and implant placement during cephalomedullary nailing of intertrochanteric hip fractures.

Dual Mobility (DM) Total Hip Replacements (THRs) were introduced to reduce dislocation risk, which is the most common cause of early revision. The in-vivo mechanics of these implants is not well understood, despite their increased use in both elective and trauma settings. Therefore, the aim of this study was to comprehensively assess retrieved DM polyethylene liners for signs of damage using visual inspection and semi-quantitative geometric assessment techniques.

Retrieved DM liners (n=20) were visually inspected for the presence of seven established modes of polyethylene damage. If embedded debris was identified on the external surface, its material composition was characterised using energy-dispersive x-ray analysis (EDX). Additionally, each liner was geometrically assessed for signs of wear/deformation using a validated methodology.

Visual inspection of the liners revealed that scratching and pitting were the most common damage modes on either surface. Burnishing was observed on 50% and 15% of the internal and external surfaces, respectively. In addition, embedded debris was identified on 25% of the internal and 65% of the external surfaces. EDX analysis of the debris identified several materials including iron, titanium, cobalt-chrome, and tantalum. Geometric analysis demonstrated highly variable damage patterns across the liners.

The results of this study provide insight into the in-vivo mechanics of DM bearings. For example, the results suggest that the internal bearing (i.e., between the head and liner) acts as the primary articulation site for DM-THRs as evidenced by a higher incidence of burnishing and larger, more concentrated regions of penetration across the liners’ internal surfaces. Furthermore, circumferential, and crescent-shaped damage patterns were identified on the articulating surfaces of the liners thus providing evidence that these components can rotate within the acetabular shell with varying degrees of mobility. The mechanics of DM bearings are complex and may be influenced by several factors (e.g., soft tissue fibrosis, patient activities) and thus further investigation is warranted.

Finally, the results of this study suggest that DM liners may be susceptible to ex-vivo surface damage and thus caution is advised when handling and/or assessing these types of components.

Joint tissues release extracellular vesicles (EVs) that potentially sustain joint homeostasis and contribute to osteoarthritis (OA) pathogenesis. EVs are putative novel therapeutics for OA, and transport biologically active molecules (including small non-coding RNAs (SNCRNAs)) between cells. This study identified altering SNCRNA cargo in EVs in OA which may act as early diagnostic markers and treatment targets.

OA was surgically induced in four skeletally mature Standardbred horses using an osteochondral fragment model in the left middle carpal joint. The right joint underwent sham surgery. Synovial fluid (SF) and plasma were obtained weekly throughout the 70-day study. EVs were isolated using size exclusion chromatography and characterised using nanoparticle tracking (Nanosight), and exosome fluorescence detection and tetraspanin phenotyping (Exoview). RNA was extracted from EVs derived from SF (sham and OA joints) and plasma collected at days 10, 35, 42, 49, 56, 63, and subjected to small RNA sequencing on a NovaSeq SP100 flow cell (Illumina).

Nanosight-derived EV characteristics of size and concentration were not significantly different following disease induction. The diameter of the temporal population of plasma and SF-derived exosomes changed significantly for CD9 and CD81 following OA induction with significant temporal, and disease-related changes in CD63 and CD81 protein expressin in plasma and SF.

In SF and plasma-derived EVs snoRNAs, snRNAs, tRNAs, lncRNA, y-RNA, piRNAs and scRNA were found. Following pairwise analysis of all-time points we identified 27 miRs DE in plasma and 45 DE miRs in SF. Seven were DE in plasma and SF; miR-451, miR-25, miR-215, miR-92a, miR-let-7c, miR-486-5p, miR-23a. In plasma and SF 35 and 21 snoRNAs were DE with four DE in plasma and SF; U3, snord15, snord46, snord58.

This work has identified alterations to OA EV sncRNAs in plasma and SF providing a greater understanding of the role of EVs in early OA.

Dual mobility (DM) total hip replacements (THRs) were introduced to reduce dislocation risk, which is the most common cause of early revision. Although DM THRs have shown good overall survivorship and low dislocation rates, the mechanisms which describe how these bearings function in-vivo are not fully understood. Therefore, the study aim was to comprehensively assess retrieved DM polyethylene liners for signs of damage using visual inspection and semi-quantitative geometric assessment methods.

Retrieved DM liners (n=18) were visually inspected for the presence of surface damage, whereby the internal and external surfaces were independently assigned a score of one (present) or zero (not present) for seven damage modes. The severity of damage was not assessed. The material composition of embedded debris was characterised using energy-dispersive x-ray analysis (EDX). Additionally, each liner was geometrically assessed for signs of wear/deformation [1].

Scratching and pitting were the most common damage modes on either surface. Additionally, burnishing was observed on 50% of the internal surfaces and embedded debris was identified on 67% of the external surfaces. EDX analysis of the debris identified several materials including titanium, cobalt-chrome, iron, and tantalum. Geometric analysis demonstrated highly variable damage patterns across the liners.

The incidence of burnishing was three times greater for the internal surfaces, suggesting that this acts as the primary articulation site. The external surfaces sustained more observable damage as evidenced by a higher incidence of embedded debris, abrasion, delamination, and deformation. In conjunction with the highly variable damage patterns observed, these results suggest that DM kinematics are complex and may be influenced by several factors (e.g., soft tissue fibrosis, patient activities) and thus further investigation is warranted.

Aims

Osteoporosis can determine surgical strategy for total hip arthroplasty (THA), and perioperative fracture risk. The aims of this study were to use hip CT to measure femoral bone mineral density (BMD) using CT X-ray absorptiometry (CTXA), determine if systematic evaluation of preoperative femoral BMD with CTXA would improve identification of osteopenia and osteoporosis compared with available preoperative dual-energy X-ray absorptiometry (DXA) analysis, and determine if improved recognition of low BMD would affect the use of cemented stem fixation.