Aims. To investigate the optimal thresholds and diagnostic efficacy of commonly used serological and synovial fluid detection indexes for diagnosing periprosthetic joint infection (PJI) in patients who have rheumatoid arthritis (RA). Methods. The data from 348 patients who had RA or osteoarthritis (OA) and had previously undergone a total knee (TKA) and/or a total hip arthroplasty (THA) (including RA-PJI: 60 cases, RA-non-PJI: 80 cases; OA-PJI: 104 cases, OA-non-PJI: 104 cases) were retrospectively analyzed. A receiver operating characteristic curve was used to determine the optimal thresholds of the CRP, ESR, synovial fluid white blood cell count (WBC), and polymorphonuclear neutrophil percentage (PMN%) for diagnosing RA-PJI and OA-PJI. The diagnostic efficacy was evaluated by comparing the area under the curve (AUC) of each index and applying the results of the combined index diagnostic test. Results. For PJI prediction, the results of serological and synovial fluid indexes were different between the RA-PJI and OA-PJI groups. The optimal cutoff value of CRP for diagnosing RA-PJI was 12.5 mg/l, ESR was 39 mm/hour, synovial fluid WBC was 3,654/μl, and PMN% was 65.9%; and those of OA-PJI were 8.2 mg/l, 31 mm/hour, 2,673/μl, and 62.0%, respectively. In the RA-PJI group, the specificity (94.4%), positive predictive value (97.1%), and AUC (0.916) of synovial fluid WBC were higher than those of the other indexes. The optimal cutoff values of synovial fluid WBC and PMN% for diagnosing RA-PJI after THA were significantly higher than those of TKA. The specificity and positive predictive value of the combined index were 100%. Conclusion. Serum inflammatory and synovial fluid indexes can be used for diagnosing RA-PJI, for which synovial fluid WBC is the best detection index. Combining multiple detection indexes can provide a reference basis for the early and accurate diagnosis of RA-PJI. Cite this article: Bone Joint Res 2023;12(9):559–570

Adolescent idiopathic scoliosis (AIS), defined by an age at presentation of 11 to 18 years, has a prevalence of 0.47% and accounts for approximately 90% of all cases of idiopathic scoliosis. Despite decades of research, the exact aetiology of AIS remains unknown. It is becoming evident that it is the result of a complex interplay of genetic, internal, and environmental factors. It has been hypothesized that genetic variants act as the initial trigger that allow epigenetic factors to propagate AIS, which could also explain the wide phenotypic variation in the presentation of the disorder. A better understanding of the underlying aetiological mechanisms could help to establish the diagnosis earlier and allow a more accurate prediction of deformity progression. This, in turn, would prompt imaging and therapeutic intervention at the appropriate time, thereby achieving the best clinical outcome for this group of patients. Cite this article: Bone Joint J 2022;104-B(8):915–921

Artificial intelligence and machine-learning analytics have gained extensive popularity in recent years due to their clinically relevant applications. A wide range of proof-of-concept studies have demonstrated the ability of these analyses to personalize risk prediction, detect implant specifics from imaging, and monitor and assess patient movement and recovery. Though these applications are exciting and could potentially influence practice, it is imperative to understand when these analyses are indicated and where the data are derived from, prior to investing resources and confidence into the results and conclusions. In this article, we review the current benefits and potential limitations of machine-learning for the orthopaedic surgeon with a specific emphasis on data quality

Aims. Bacterial infection activates neutrophils to release neutrophil extracellular traps (NETs) in bacterial biofilms of periprosthetic joint infections (PJIs). The aim of this study was to evaluate the increase in NET activation and release (NETosis) and haemostasis markers in the plasma of patients with PJI, to evaluate whether such plasma induces the activation of neutrophils, to ascertain whether increased NETosis is also mediated by reduced DNaseI activity, to explore novel therapeutic interventions for NETosis in PJI in vitro, and to evaluate the potential diagnostic use of these markers. Methods. We prospectively recruited 107 patients in the preoperative period of prosthetic surgery, 71 with a suspicion of PJI and 36 who underwent arthroplasty for non-septic indications as controls, and obtained citrated plasma. PJI was confirmed in 50 patients. We measured NET markers, inflammation markers, DNaseI activity, haemostatic markers, and the thrombin generation test (TGT). We analyzed the ability of plasma from confirmed PJI and controls to induce NETosis and to degrade in vitro-generated NETs, and explored the therapeutic restoration of the impairment to degrade NETs of PJI plasma with recombinant human DNaseI. Finally, we assessed the contribution of these markers to the diagnosis of PJI. Results. Patients with confirmed PJI had significantly increased levels of NET markers (cfDNA (p < 0.001), calprotectin (p < 0.001), and neutrophil elastase (p = 0.022)) and inflammation markers (IL-6; p < 0.001) in plasma. Moreover, the plasma of patients with PJI induced significantly more neutrophil activation than the plasma of the controls (p < 0.001) independently of tumour necrosis factor alpha. Patients with PJI also had a reduced DNaseI activity in plasma (p < 0.001), leading to a significantly impaired degradation of NETs (p < 0.001). This could be therapeutically restored with recombinant human DNaseI to the level in the controls. We developed a model to improve the diagnosis of PJI with cfDNA, calprotectin, and the start tail of TGT as predictors, though cfDNA alone achieved a good prediction and is simpler to measure. Conclusion. We confirmed that patients with PJI have an increased level of NETosis in plasma. Their plasma both induced NET release and had an impaired ability to degrade NETs mediated by a reduced DNaseI activity. This can be therapeutically restored in vitro with the approved Dornase alfa, Pulmozyme, which may allow novel methods of treatment. A combination of NETs and haemostatic biomarkers could improve the diagnosis of PJI, especially those patients in whom this diagnosis is uncertain. Cite this article: Bone Joint J 2024;106-B(9):1021–1030

Aims. Surgical site infection (SSI) after soft-tissue sarcoma (STS) resection is a serious complication. The purpose of this retrospective study was to investigate the risk factors for SSI after STS resection, and to develop a nomogram that allows patient-specific risk assessment. Methods. A total of 547 patients with STS who underwent tumour resection between 2005 and 2021 were divided into a development cohort and a validation cohort. In the development cohort of 402 patients, the least absolute shrinkage and selection operator (LASSO) regression model was used to screen possible risk factors of SSI. To select risk factors and construct the prediction nomogram, multivariate logistic regression was used. The predictive power of the nomogram was evaluated by receiver operating curve (ROC) analysis in the validation cohort of 145 patients. Results. LASSO regression analysis selected possible risk factors for SSI, including age, diabetes, operating time, skin graft or flap, resected tumour size, smoking, and radiation therapy. Multivariate analysis revealed that age, diabetes, smoking during the previous year, operating time, and radiation therapy were independent risk factors for SSI. A nomogram was developed based on the results of multivariate logistic regression analysis. In the development cohort, the incidence of SSI was 4.5% in the low-risk group (risk score < 6.89) and 26.6% in the high-risk group (risk score ≥ 6.89; p < 0.001). In the validation cohort, the incidence of SSI was 2.0% in the low-risk group and 15.9% in the high-risk group (p = 0.004). Conclusion. Our nomogram will enable surgeons to assess the risk of SSI in patients with STS. In patients with high risk of SSI, frequent monitoring and aggressive interventions should be considered to prevent this. Cite this article: Bone Joint J 2024;106-B(5):492–500

Aims. Advances in treatment have extended the life expectancy of patients with metastatic bone disease (MBD). Patients could experience more skeletal-related events (SREs) as a result of this progress. Those who have already experienced a SRE could encounter another local management for a subsequent SRE, which is not part of the treatment for the initial SRE. However, there is a noted gap in research on the rate and characteristics of subsequent SREs requiring further localized treatment, obligating clinicians to extrapolate from experiences with initial SREs when confronting subsequent ones. This study aimed to investigate the proportion of MBD patients developing subsequent SREs requiring local treatment, examine if there are prognostic differences at the initial treatment between those with single versus subsequent SREs, and determine if clinical, oncological, and prognostic features differ between initial and subsequent SRE treatments. Methods. This retrospective study included 3,814 adult patients who received local treatment – surgery and/or radiotherapy – for bone metastasis between 1 January 2010 and 31 December 2019. All included patients had at least one SRE requiring local treatment. A subsequent SRE was defined as a second SRE requiring local treatment. Clinical, oncological, and prognostic features were compared between single SREs and subsequent SREs using Mann-Whitney U test, Fisher’s exact test, and Kaplan–Meier curve. Results. Of the 3,814 patients with SREs, 3,159 (83%) patients had a single SRE and 655 (17%) patients developed a subsequent SRE. Patients who developed subsequent SREs generally had characteristics that favoured longer survival, such as higher BMI, higher albumin levels, fewer comorbidities, or lower neutrophil count. Once the patient got to the point of subsequent SRE, their clinical and oncological characteristics and one-year survival (28%) were not as good as those with only a single SRE (35%; p < 0.001), indicating that clinicians’ experiences when treating the initial SRE are not similar when treating a subsequent SRE. Conclusion. This study found that 17% of patients required treatments for a second, subsequent SRE, and the current clinical guideline did not provide a specific approach to this clinical condition. We observed that referencing the initial treatment, patients in the subsequent SRE group had longer six-week, 90-day, and one-year median survival than patients in the single SRE group. Once patients develop a subsequent SRE, they have a worse one-year survival rate than those who receive treatment for a single SRE. Future research should identify prognostic factors and assess the applicability of existing survival prediction models for better management of subsequent SREs. Cite this article: Bone Joint Res 2024;13(9):497–506

Aims. We assessed the value of the Clinical Frailty Scale (CFS) in the prediction of adverse outcome after hip fracture. Methods. Of 1,577 consecutive patients aged > 65 years with a fragility hip fracture admitted to one institution, for whom there were complete data, 1,255 (72%) were studied. Clinicians assigned CFS scores on admission. Audit personnel routinely prospectively completed the Standardised Audit of Hip Fracture in Europe form, including the following outcomes: 30-day survival; in-hospital complications; length of acute hospital stay; and new institutionalization. The relationship between the CFS scores and outcomes was examined graphically and the visual interpretations were tested statistically. The predictive values of the CFS and Nottingham Hip Fracture Score (NHFS) to predict 30-day mortality were compared using receiver operating characteristic area under the curve (AUC) analysis. Results. Significant non-linear associations between CFS and outcomes were observed. Risk of death within 30 days rose linearly for CFS 1 to 5, but plateaued for CFS > 5. The incidence of complications and length of stay rose linearly for CFS 1 to 4, but plateaued for CFS > 4. In contrast, the risk of new institutionalization rose linearly for CFS 1 to 8. The AUCs for 30-day mortality for the CFS and NHFS were very similar: CFS AUC 0.63 (95% CI 0.57 to 0.69) and NHFS AUC 0.63 (95% CI 0.57 to 0.69). Conclusion. Use of the CFS may provide useful information on outcomes for fitter patients presenting with hip fracture, but completion of the CFS by the admitting orthopaedic team does not appear successful in distinguishing between higher CFS categories, which define patients with frailty. This makes a strong case for the role of the orthogeriatrician in the early assessment of these patients. Further work is needed to understand why patients assessed as being of mild, moderate, and severe frailty do not result in different outcomes. Cite this article: Bone Joint J 2022;104-B(8):980–986

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

Aims. The aim of this study was to compare any differences in the primary outcome (biphasic flexion knee moment during gait) of robotic arm-assisted bi-unicompartmental knee arthroplasty (bi-UKA) with conventional mechanically aligned total knee arthroplasty (TKA) at one year post-surgery. Methods. A total of 76 patients (34 bi-UKA and 42 TKA patients) were analyzed in a prospective, single-centre, randomized controlled trial. Flat ground shod gait analysis was performed preoperatively and one year postoperatively. Knee flexion moment was calculated from motion capture markers and force plates. The same setup determined proprioception outcomes during a joint position sense test and one-leg standing. Surgery allocation, surgeon, and secondary outcomes were analyzed for prediction of the primary outcome from a binary regression model. Results. Both interventions were shown to be effective treatment options, with no significant differences shown between interventions for the primary outcome of this study (18/35 (51.4%) biphasic TKA patients vs 20/31 (64.5%) biphasic bi-UKA patients; p = 0.558). All outcomes were compared to an age-matched, healthy cohort that outperformed both groups, indicating residual deficits exists following surgery. Logistic regression analysis of primary outcome with secondary outcomes indicated that the most significant predictor of postoperative biphasic knee moments was preoperative knee moment profile and trochlear degradation (Outerbridge) (R. 2. = 0.381; p = 0.002, p = 0.046). A separate regression of alignment against primary outcome indicated significant bi-UKA femoral and tibial axial alignment (R. 2. = 0.352; p = 0.029), and TKA femoral sagittal alignment (R. 2. = 0.252; p = 0.016). The bi-UKA group showed a significant increased ability in the proprioceptive joint position test, but no difference was found in more dynamic testing of proprioception. Conclusion. Robotic arm-assisted bi-UKA demonstrated equivalence to TKA in achieving a biphasic gait pattern after surgery for osteoarthritis of the knee. Both treatments are successful at improving gait, but both leave the patients with a functional limitation that is not present in healthy age-matched controls. Cite this article: Bone Joint J 2022;103-B(4):433–443

Aims. The aim of this study was to assess whether supine flexibility predicts the likelihood of curve progression in patients with adolescent idiopathic scoliosis (AIS) undergoing brace treatment. Methods. This was a retrospective analysis of patients with AIS prescribed with an underarm brace between September 2008 to April 2013 and followed up until 18 years of age or required surgery. Patients with structural proximal curves that preclude underarm bracing, those who were lost to follow-up, and those who had poor compliance to bracing (<16 hours a day) were excluded. The major curve Cobb angle, curve type, and location were measured on the pre-brace standing posteroanterior (PA) radiograph, supine whole spine radiograph, initial in-brace standing PA radiograph, and the post-brace weaning standing PA radiograph. Validation of the previous in-brace Cobb angle regression model was performed. The outcome of curve progression post-bracing was tested using a logistic regression model. The supine flexibility cut-off for curve progression was analyzed with receiver operating characteristic curve. Results. A total of 586 patients with mean age of 12.6 years (SD 1.2) remained for analysis after exclusion. The baseline Cobb angle was similar for thoracic major curves (31.6° (SD 3.8°)) and lumbar major curves (30.3° (SD 3.7°)). Curve progression was more common in the thoracic curves than lumbar curves with mean final Cobb angles of 40.5° (SD 12.5°) and 31.8° (SD 9.8°) respectively. This dataset matched the prediction model for in-brace Cobb angle with less mean absolute error in thoracic curves (0.61) as compared to lumbar curves (1.04). Reduced age and Risser stage, thoracic curves, increased pre-brace Cobb angle, and reduced correction and flexibility rates predicted increased likelihood of curve progression. Flexibility rate of more than 28% has likelihood of preventing curve progression with bracing. Conclusion. Supine radiographs provide satisfactory prediction for in-brace correction and post-bracing curve magnitude. The flexibility of the curve is a guide to determine the likelihood for brace success. Cite this article: Bone Joint J 2020;102-B(2):254–260

Objectives. In this prospective cohort study, we investigated whether patient-specific finite element (FE) models can identify patients at risk of a pathological femoral fracture resulting from metastatic bone disease, and compared these FE predictions with clinical assessments by experienced clinicians. Methods. A total of 39 patients with non-fractured femoral metastatic lesions who were irradiated for pain were included from three radiotherapy institutes. During follow-up, nine pathological fractures occurred in seven patients. Quantitative CT-based FE models were generated for all patients. Femoral failure load was calculated and compared between the fractured and non-fractured femurs. Due to inter-scanner differences, patients were analyzed separately for the three institutes. In addition, the FE-based predictions were compared with fracture risk assessments by experienced clinicians. Results. In institute 1, median failure load was significantly lower for patients who sustained a fracture than for patients with no fractures. In institutes 2 and 3, the number of patients with a fracture was too low to make a clear distinction. Fracture locations were well predicted by the FE model when compared with post-fracture radiographs. The FE model was more accurate in identifying patients with a high fracture risk compared with experienced clinicians, with a sensitivity of 89% versus 0% to 33% for clinical assessments. Specificity was 79% for the FE models versus 84% to 95% for clinical assessments. Conclusion. FE models can be a valuable tool to improve clinical fracture risk predictions in metastatic bone disease. Future work in a larger patient population should confirm the higher predictive power of FE models compared with current clinical guidelines. Cite this article: F. Eggermont, L. C. Derikx, N. Verdonschot, I. C. M. van der Geest, M. A. A. de Jong, A. Snyers, Y. M. van der Linden, E. Tanck. Can patient-specific finite element models better predict fractures in metastatic bone disease than experienced clinicians? Towards computational modelling in daily clinical practice. Bone Joint Res 2018;7:430–439. DOI: 10.1302/2046-3758.76.BJR-2017-0325.R2

Aims. Intra-articular (IA) tumours around the knee are treated with extra-articular (EA) resection, which is associated with poor functional outcomes. We aim to evaluate the accuracy of MRI in predicting IA involvement around the knee. Methods. We identified 63 cases of high-grade sarcomas in or around the distal femur that underwent an EA resection from a prospectively maintained database (January 1996 to April 2020). Suspicion of IA disease was noted in 52 cases, six had IA pathological fracture, two had an effusion, two had prior surgical intervention (curettage/IA intervention), and one had an osseous metastasis in the proximal tibia. To ascertain validity, two musculoskeletal radiologists (R1, R2) reviewed the preoperative imaging (MRI) of 63 consecutive cases on two occasions six weeks apart. The radiological criteria for IA disease comprised evidence of tumour extension within the suprapatellar pouch, intercondylar notch, extension along medial/lateral retinaculum, and presence of IA fracture. The radiological predictions were then confirmed with the final histopathology of the resected specimens. Results. The resection histology revealed 23 cases (36.5%) showing IA disease involvement compared with 40 cases without (62%). The intraobserver variability of R1 was 0.85 (p < 0.001) compared to R2 with κ = 0.21 (p = 0.007). The interobserver variability was κ = 0.264 (p = 0.003). Knee effusion was found to be the most sensitive indicator of IA involvement, with a sensitivity of 91.3% but specificity of only 35%. However, when combined with a pathological fracture, this rose to 97.5% and 100% when disease was visible in Hoffa’s fat pad. Conclusion. MRI imaging can sometimes overestimate IA joint involvement and needs to be correlated with clinical signs. In the light of our findings, we would recommend EA resections when imaging shows effusion combined with either disease in Hoffa’s fat pad or retinaculum, or pathological fractures. Cite this article: Bone Joint J 2023;105-B(6):696–701

Aims. The lateral centre-edge angle (LCEA) is a plain radiological measure of superolateral cover of the femoral head. This study aims to establish the correlation between 2D radiological and 3D CT measurements of acetabular morphology, and to describe the relationship between LCEA and femoral head cover (FHC). Methods. This retrospective study included 353 periacetabular osteotomies (PAOs) performed between January 2014 and December 2017. Overall, 97 hips in 75 patients had 3D analysis by Clinical Graphics, giving measurements for LCEA, acetabular index (AI), and FHC. Roentgenographical LCEA, AI, posterior wall index (PWI), and anterior wall index (AWI) were measured from supine AP pelvis radiographs. The correlation between CT and roentgenographical measurements was calculated. Sequential multiple linear regression was performed to determine the relationship between roentgenographical measurements and CT FHC. Results. CT-measured LCEA and AI correlated strongly with roentgenographical LCEA (r = 0.92; p < 0.001) and AI (r = 0.83; p < 0.001). Radiological LCEA correlated very strongly with CT FHC (r = 0.92; p < 0.001). The sum of AWI and PWI also correlated strongly with CTFHC (r = 0.73; p < 0.001). CT measurements of LCEA and AI were 3.4° less and 2.3° greater than radiological LCEA and AI measures. There was a linear relation between radiological LCEA and CT FHC. The linear regression model statistically significantly predicted FHC from LCEA, F(1,96) = 545.1 (p < 0.001), adjusted R. 2. = 85.0%, with the prediction equation: CT FHC(%) = 42.1 + 0.77(XRLCEA). Conclusion. CT and roentgenographical measurement of acetabular parameters are comparable. Currently, a radiological LCEA greater than 25° is considered normal. This study demonstrates that those with hip pain and normal radiological acetabular parameters may still have deficiencies in FHC. More sophisticated imaging techniques such as 3D CT should be considered for those with hip pain to identify deficiencies in FHC. Cite this article: Bone Jt Open 2022;3(1):12–19

Aims. Excessive posterior pelvic tilt (PT) may increase the risk of anterior instability after total hip arthroplasty (THA). The aim of this study was to investigate the changes in PT occurring from the preoperative supine to postoperative standing position following THA, and identify factors associated with significant changes in PT. Methods. Supine PT was measured on preoperative CT scans and standing PT was measured on preoperative and one-year postoperative standing lateral radiographs in 933 patients who underwent primary THA. Negative values indicate posterior PT. Patients with > 13° of posterior PT from preoperative supine to postoperative standing (ΔPT ≤ -13°) radiographs, which corresponds to approximately a 10° increase in functional anteversion of the acetabular component, were compared with patients with less change (ΔPT > -13°). Logistic regression analysis was used to assess preoperative demographic and spinopelvic parameters predictive of PT changes of ≤ -13°. The area under receiver operating characteristic curve (AUC) determined the diagnostic accuracy of the predictive factors. Results. PT changed from a mean of 3.8° (SD 6.0°)) preoperatively to -3.5° (SD 6.9°) postoperatively, a mean change of -7.4 (SD 4.5°; p < 0.001). A total of 95 patients (10.2%) had ≤ -13° change in PT from preoperative supine to postoperative standing. The strongest predictive preoperative factors of large changes in PT (≤ -13°) from preoperative supine to postoperative standing were a large posterior change in PT from supine to standing, increased supine PT, and decreased standing PT (p < 0.001). Flexed-seated PT (p = 0.006) and female sex (p = 0.045) were weaker significant predictive factors. When including all predictive factors, the accuracy of the AUC prediction was 84.9%, with 83.5% sensitivity and 71.2% specificity. Conclusion. A total of 10% of patients had > 13° of posterior PT postoperatively compared with their supine pelvic position, resulting in an increased functional anteversion of > 10°. The strongest predictive factors of changes in postoperative PT were the preoperative supine-to-standing differences, the anterior supine PT, and the posterior standing PT. Surgeons who introduce the acetabular component with the patient supine using an anterior approach should be aware of the potentially large increase in functional anteversion occurring in these patients. Cite this article: Bone Joint J 2024;106-B(3 Supple A):74–80

Aims. To devise a method to quantify and optimize tightness when inserting cortical screws, based on bone characterization and screw geometry. Methods. Cortical human cadaveric diaphyseal tibiae screw holes (n = 20) underwent destructive testing to firstly establish the relationship between cortical thickness and experimental stripping torque (T. str. ), and secondly to calibrate an equation to predict T. str. Using the equation’s predictions, 3.5 mm screws were inserted (n = 66) to targeted torques representing 40% to 100% of T. str. , with recording of compression generated during tightening. Once the target torque had been achieved, immediate pullout testing was performed. Results. Cortical thickness predicted T. str. (R. 2. = 0.862; p < 0.001) as did an equation based on tensile yield stress, bone-screw friction coefficient, and screw geometries (R. 2. = 0.894; p < 0.001). Compression increased with screw tightness up to 80% of the maximum (R. 2. = 0.495; p < 0.001). Beyond 80%, further tightening generated no increase in compression. Pullout force did not change with variations in submaximal tightness beyond 40% of T. str. (R. 2. = 0.014; p = 0.175). Conclusion. Screw tightening between 70% and 80% of the predicted maximum generated optimum compression and pullout forces. Further tightening did not considerably increase compression, made no difference to pullout, and increased the risk of the screw holes being stripped. While further work is needed for development of intraoperative methods for accurate and reliable prediction of the maximum tightness for a screw, this work justifies insertion torque being considerably below the maximum. Cite this article: Bone Joint Res 2020;9(8):493–500

Aims. The number of convolutional neural networks (CNN) available for fracture detection and classification is rapidly increasing. External validation of a CNN on a temporally separate (separated by time) or geographically separate (separated by location) dataset is crucial to assess generalizability of the CNN before application to clinical practice in other institutions. We aimed to answer the following questions: are current CNNs for fracture recognition externally valid?; which methods are applied for external validation (EV)?; and, what are reported performances of the EV sets compared to the internal validation (IV) sets of these CNNs?. Methods. The PubMed and Embase databases were systematically searched from January 2010 to October 2020 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. The type of EV, characteristics of the external dataset, and diagnostic performance characteristics on the IV and EV datasets were collected and compared. Quality assessment was conducted using a seven-item checklist based on a modified Methodologic Index for NOn-Randomized Studies instrument (MINORS). Results. Out of 1,349 studies, 36 reported development of a CNN for fracture detection and/or classification. Of these, only four (11%) reported a form of EV. One study used temporal EV, one conducted both temporal and geographical EV, and two used geographical EV. When comparing the CNN’s performance on the IV set versus the EV set, the following were found: AUCs of 0.967 (IV) versus 0.975 (EV), 0.976 (IV) versus 0.985 to 0.992 (EV), 0.93 to 0.96 (IV) versus 0.80 to 0.89 (EV), and F1-scores of 0.856 to 0.863 (IV) versus 0.757 to 0.840 (EV). Conclusion. The number of externally validated CNNs in orthopaedic trauma for fracture recognition is still scarce. This greatly limits the potential for transfer of these CNNs from the developing institute to another hospital to achieve similar diagnostic performance. We recommend the use of geographical EV and statements such as the Consolidated Standards of Reporting Trials–Artificial Intelligence (CONSORT-AI), the Standard Protocol Items: Recommendations for Interventional Trials–Artificial Intelligence (SPIRIT-AI) and the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis–Machine Learning (TRIPOD-ML) to critically appraise performance of CNNs and improve methodological rigor, quality of future models, and facilitate eventual implementation in clinical practice. Cite this article: Bone Jt Open 2021;2(10):879–885

Aims. The use of technology to assess balance and alignment during total knee surgery can provide an overload of numerical data to the surgeon. Meanwhile, this quantification holds the potential to clarify and guide the surgeon through the surgical decision process when selecting the appropriate bone recut or soft tissue adjustment when balancing a total knee. Therefore, this paper evaluates the potential of deploying supervised machine learning (ML) models to select a surgical correction based on patient-specific intra-operative assessments. Methods. Based on a clinical series of 479 primary total knees and 1,305 associated surgical decisions, various ML models were developed. These models identified the indicated surgical decision based on available, intra-operative alignment, and tibiofemoral load data. Results. With an associated area under the receiver-operator curve ranging between 0.75 and 0.98, the optimized ML models resulted in good to excellent predictions. The best performing model used a random forest approach while considering both alignment and intra-articular load readings. Conclusion. The presented model has the potential to make experience available to surgeons adopting new technology, bringing expert opinion in their operating theatre, but also provides insight in the surgical decision process. More specifically, these promising outcomes indicated the relevance of considering the overall limb alignment in the coronal and sagittal plane to identify the appropriate surgical decision

Aims. Accurate estimations of the risk of fracture due to metastatic bone disease in the femur is essential in order to avoid both under-treatment and over-treatment of patients with an impending pathological fracture. The purpose of the current retrospective in vivo study was to use CT-based finite element analyses (CTFEA) to identify a clear quantitative differentiating factor between patients who are at imminent risk of fracturing their femur and those who are not, and to identify the exact location of maximal weakness where the fracture is most likely to occur. Methods. Data were collected on 82 patients with femoral metastatic bone disease, 41 of whom did not undergo prophylactic fixation. A total of 15 had a pathological fracture within six months following the CT scan, and 26 were fracture-free during the five months following the scan. The Mirels score and strain fold ratio (SFR) based on CTFEA was computed for all patients. A SFR value of 1.48 was used as the threshold for a pathological fracture. The sensitivity, specificity, positive, and negative predicted values for Mirels score and SFR predictions were computed for nine patients who fractured and 24 who did not, as well as a comparison of areas under the receiver operating characteristic curves (AUC of the ROC curves). Results. The sensitivity of SFR was 100% compared with 88% for the Mirels score, and the specificity of SFR was 67% compared with 38% for the Mirels score. The AUC was 0.905 for SFR compared with 0.578 for the Mirels score (p = 0.008). Conclusion. All the patients who sustained a pathological fracture of the femur had an SFR of > 1.48. CTFEA was far better at predicting the risk of fracture and its location accurately compared with the Mirels score. CTFEA is quick and automated and can be incorporated into the protocol of CT scanners. Cite this article: Bone Joint J 2020;102-B(5):638–645

Aims. In computer simulations, the shape of the range of motion (ROM) of a stem with a cylindrical neck design will be a perfect cone. However, many modern stems have rectangular/oval-shaped necks. We hypothesized that the rectangular/oval stem neck will affect the shape of the ROM and the prosthetic impingement. Methods. Total hip arthroplasty (THA) motion while standing and sitting was simulated using a MATLAB model (one stem with a cylindrical neck and one stem with a rectangular neck). The primary predictor was the geometry of the neck (cylindrical vs rectangular) and the main outcome was the shape of ROM based on the prosthetic impingement between the neck and the liner. The secondary outcome was the difference in the ROM provided by each neck geometry and the effect of the pelvic tilt on this ROM. Multiple regression was used to analyze the data. Results. The stem with a rectangular neck has increased internal and external rotation with a quatrefoil cross-section compared to a cone in a cylindrical neck. Modification of the cup orientation and pelvic tilt affected the direction of projection of the cone or quatrefoil shape. The mean increase in internal rotation with a rectangular neck was 3.4° (0° to 7.9°; p < 0.001); for external rotation, it was 2.8° (0.5° to 7.8°; p < 0.001). Conclusion. Our study shows the importance of attention to femoral implant design for the assessment of prosthetic impingement. Any universal mathematical model or computer simulation that ignores each stem’s unique neck geometry will provide inaccurate predictions of prosthetic impingement. Cite this article: Bone Joint Res 2021;10(12):780–789

Aims. A comprehensive classification for coronal lower limb alignment with predictive capabilities for knee balance would be beneficial in total knee arthroplasty (TKA). This paper describes the Coronal Plane Alignment of the Knee (CPAK) classification and examines its utility in preoperative soft tissue balance prediction, comparing kinematic alignment (KA) to mechanical alignment (MA). Methods. A radiological analysis of 500 healthy and 500 osteoarthritic (OA) knees was used to assess the applicability of the CPAK classification. CPAK comprises nine phenotypes based on the arithmetic HKA (aHKA) that estimates constitutional limb alignment and joint line obliquity (JLO). Intraoperative balance was compared within each phenotype in a cohort of 138 computer-assisted TKAs randomized to KA or MA. Primary outcomes included descriptive analyses of healthy and OA groups per CPAK type, and comparison of balance at 10° of flexion within each type. Secondary outcomes assessed balance at 45° and 90° and bone recuts required to achieve final knee balance within each CPAK type. Results. There was similar frequency distribution between healthy and arthritic groups across all CPAK types. The most common categories were Type II (39.2% healthy vs 32.2% OA), Type I (26.4% healthy vs 19.4% OA) and Type V (15.4% healthy vs 14.6% OA). CPAK Types VII, VIII, and IX were rare in both populations. Across all CPAK types, a greater proportion of KA TKAs achieved optimal balance compared to MA. This effect was largest, and statistically significant, in CPAK Types I (100% KA vs 15% MA; p < 0.001), Type II (78% KA vs 46% MA; p = 0.018). and Type IV (89% KA vs 0% MA; p < 0.001). Conclusion. CPAK is a pragmatic, comprehensive classification for coronal knee alignment, based on constitutional alignment and JLO, that can be used in healthy and arthritic knees. CPAK identifies which knee phenotypes may benefit most from KA when optimization of soft tissue balance is prioritized. Further, it will allow for consistency of reporting in future studies. Cite this article: Bone Joint J 2021;103-B(2):329–337