Aims

To perform an incremental cost-utility analysis and assess the impact of differential costs and case volume on the cost-effectiveness of robotic arm-assisted unicompartmental knee arthroplasty (rUKA) compared to manual (mUKA).

Artificial intelligence (AI) is, in essence, the concept of ‘computer thinking’, encompassing methods that train computers to perform and learn from executing certain tasks, called machine learning, and methods to build intricate computer models that both learn and adapt, called complex neural networks. Computer vision is a function of AI by which machine learning and complex neural networks can be applied to enable computers to capture, analyze, and interpret information from clinical images and visual inputs. This annotation summarizes key considerations and future perspectives concerning computer vision, questioning the need for this technology (the ‘why’), the current applications (the ‘what’), and the approach to unlocking its full potential (the ‘how’).

Cite this article: Bone Joint J 2022;104-B(8):911–914.

Aims

Pelvic incidence (PI) is a position-independent spinopelvic parameter traditionally used by spinal surgeons to determine spinal alignment. Its relevance to the arthroplasty surgeon in assessing patient risk for total hip arthroplasty (THA) instability preoperatively is unclear. This study was undertaken to investigate the significance of PI relative to other spinopelvic parameter risk factors for instability to help guide its clinical application.

Aims

The aim of this study was to assess whether it is possible to predict the mortality, and the extent and time of neurological recovery from the time of the onset of symptoms and MRI grade, in patients with the cerebral fat embolism syndrome (CFES). This has not previously been investigated.

Aims

In this study, we aimed to explore surgical variations in the Femoral Neck System (FNS) used for stable fixation of Pauwels type III femoral neck fractures.

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?

Aims. Mechanical impingement of the iliopsoas (IP) tendon accounts for 2% to 6% of persistent postoperative pain after total hip arthroplasty (THA). The most common initiator is anterior acetabular component protrusion, where the anterior margin is not covered by anterior acetabular wall. A CT scan can be used to identify and measure this overhang; however, no threshold exists for determining symptomatic anterior IP impingement due to overhang. A case-control study was conducted in which CT scan measurements were used to define a threshold that differentiates patients with IP impingement from asymptomatic patients after THA. Methods. We analyzed the CT scans of 622 patients (758 THAs) between May 2011 and May 2020. From this population, we identified 136 patients with symptoms suggestive of IP impingement. Among them, six were subsequently excluded: three because the diagnosis was refuted intraoperatively, and three because they had another obvious cause of impingement, leaving 130 hips (130 patients) in the study (impingement) group. They were matched to a control group of 138 asymptomatic hips (138 patients) after THA. The anterior acetabular component overhang was measured on an axial CT slice based on anatomical landmarks (orthogonal to the pelvic axis). Results. The impingement group had a median overhang of 8 mm (interquartile range (IQR) 5 to 11) versus 0 mm (IQR 0 to 4) for the control group (p < 0.001). Using receiver operating characteristic curves, an overhang threshold of 4 mm was best correlated with a diagnosis of impingement (sensitivity 79%, specificity 85%; positive predictive value 75%, negative predictive value 85%). Conclusion. Pain after THA related to IP impingement can be reasonably linked to acetabular overhang if it exceeds 4 mm on a CT scan. Below this threshold, it seems logical to look for another cause of IP irritation or another reason for the pain after THA before concluding that impingement is present. Cite this article: Bone Joint J 2024;106-B(3 Supple A):97–103

Aims. The Wrightington classification system of fracture-dislocations of the elbow divides these injuries into six subtypes depending on the involvement of the coronoid and the radial head. The aim of this study was to assess the reliability and reproducibility of this classification system. Methods. This was a blinded study using radiographs and CT scans of 48 consecutive patients managed according to the Wrightington classification system between 2010 and 2018. Four trauma and orthopaedic consultants, two post CCT fellows, and one speciality registrar based in the UK classified the injuries. The seven observers reviewed preoperative radiographs and CT scans twice, with a minimum four-week interval. Radiographs and CT scans were reviewed separately. Inter- and intraobserver reliability were calculated using Fleiss and Cohen kappa coefficients. The Landis and Koch criteria were used to interpret the strength of the kappa values. Validity was assessed by calculating the percentage agreement against intraoperative findings. Results. Of the 48 patients, three (6%) had type A injury, 11 (23%) type B, 16 (33%) type B+, 16 (33%) Type C, two (4%) type D+, and none had a type D injury. All 48 patients had anteroposterior (AP) and lateral radiographs, 44 had 2D CT scans, and 39 had 3D reconstructions. The interobserver reliability kappa value was 0.52 for radiographs, 0.71 for 2D CT scans, and 0.73 for a combination of 2D and 3D reconstruction CT scans. The median intraobserver reliability was 0.75 (interquartile range (IQR) 0.62 to 0.79) for radiographs, 0.77 (IQR 0.73 to 0.94) for 2D CT scans, and 0.89 (IQR 0.77 to 0.93) for the combination of 2D and 3D reconstruction. Validity analysis showed that accuracy significantly improved when using CT scans (p = 0.018 and p = 0.028 respectively). Conclusion. The Wrightington classification system is a reliable and valid method of classifying fracture-dislocations of the elbow. CT scans are significantly more accurate than radiographs when identifying the pattern of injury, with good intra- and interobserver reproducibility. Cite this article: Bone Joint J 2020;102-B(8):1041–1047

Aims. The aim of this study was to assess the reliability of using MRI scans to calculate the Spinal Instability Neoplastic Score (SINS) in patients with metastatic spinal cord compression (MSCC). Methods. A total of 100 patients were retrospectively included in the study. The SINS score was calculated from each patient’s MRI and CT scans by two consultant musculoskeletal radiologists (reviewers 1 and 2) and one consultant spinal surgeon (reviewer 3). In order to avoid potential bias in the assessment, MRI scans were reviewed first. Bland-Altman analysis was used to identify the limits of agreement between the SINS scores from the MRI and CT scans for the three reviewers. Results. The limit of agreement between the SINS score from the MRI and CT scans for the reviewers was -0.11 for reviewer 1 (95% CI 0.82 to -1.04), -0.12 for reviewer 2 (95% CI 1.24 to -1.48), and -0.37 for reviewer 3 (95% CI 2.35 to -3.09). The use of MRI tended to increase the score when compared with that using the CT scan. No patient having their score calculated from MRI scans would have been classified as stable rather than intermediate or unstable when calculated from CT scans, potentially leading to suboptimal care. Conclusion. We found that MRI scans can be used to calculate the SINS score reliably, compared with the score from CT scans. The main difference between the scores derived from MRI and CT was in defining the type of bony lesion. This could be made easier by knowing the site of the primary tumour when calculating the score, or by using inverted T1-volumetric interpolated breath-hold examination MRI to assess the bone more reliably, similar to using CT. Cite this article: Bone Joint J 2021;103-B(5):971–975

Aims. Cross-table lateral (CTL) radiographs are commonly used to measure acetabular component anteversion after total hip arthroplasty (THA). The CTL measurements may differ by > 10° from CT scan measurements but the reasons for this discrepancy are poorly understood. Anteversion measurements from CTL radiographs and CT scans are compared to identify spinopelvic parameters predictive of inaccuracy. Methods. THA patients (n = 47; 27 males, 20 females; mean age 62.9 years (SD 6.95)) with preoperative spinopelvic mobility, radiological analysis, and postoperative CT scans were retrospectively reviewed. Acetabular component anteversion was measured on postoperative CTL radiographs and CT scans using 3D reconstructions of the pelvis. Two cohorts were identified based on a CTL-CT error of ≥ 10° (n = 11) or < 10° (n = 36). Spinopelvic mobility parameters were compared using independent-samples t-tests. Correlation between error and mobility parameters were assessed with Pearson’s coefficient. Results. Patients with CTL error > 10° (10° to 14°) had stiffer lumbar spines with less mean lumbar flexion (38.9°(SD 11.6°) vs 47.4° (SD 13.1°); p = 0.030), different sagittal balance measured by pelvic incidence-lumbar lordosis mismatch (5.9° (SD 18.8°) vs -1.7° (SD 9.8°); p = 0.042), more pelvic extension when seated (pelvic tilt -9.7° (SD 14.1°) vs -2.2° (SD 13.2°); p = 0.050), and greater change in pelvic tilt between supine and seated positions (12.6° (SD 12.1°) vs 4.7° (SD 12.5°); p = 0.036). The CTL measurement error showed a positive correlation with increased CTL anteversion (r = 0.5; p = 0.001), standing lordosis (r = 0.23; p = 0.050), seated lordosis (r = 0.4; p = 0.009), and pelvic tilt change between supine and step-up positions (r = 0.34; p = 0.010). Conclusion. Differences in spinopelvic mobility may explain the variability of acetabular anteversion measurements made on CTL radiographs. Patients with stiff spines and increased compensatory pelvic movement have less accurate measurements on CTL radiographs. Flexion of the contralateral hip is required to obtain clear CTL radiographs. In patients with lumbar stiffness, this movement may extend the pelvis and increase anteversion of the acetabulum on CTL views. Reliable analysis of acetabular component anteversion in this patient population may require advanced imaging with a CT scan. Cite this article: Bone Joint J 2021;103-B(7 Supple B):59–65

Aims. The aim of this study was to examine the implant accuracy of custom-made partial pelvis replacements (PPRs) in revision total hip arthroplasty (rTHA). Custom-made implants offer an option to achieve a reconstruction in cases with severe acetabular bone loss. By analyzing implant deviation in CT and radiograph imaging and correlating early clinical complications, we aimed to optimize the usage of custom-made implants. Methods. A consecutive series of 45 (2014 to 2019) PPRs for Paprosky III defects at rTHA were analyzed comparing the preoperative planning CT scans used to manufacture the implants with postoperative CT scans and radiographs. The anteversion (AV), inclination (IC), deviation from the preoperatively planned implant position, and deviation of the centre of rotation (COR) were explored. Early postoperative complications were recorded, and factors for malpositioning were sought. The mean follow-up was 30 months (SD 19; 6 to 74), with four patients lost to follow-up. Results. Mean CT defined discrepancy (Δ) between planned and achieved AV and IC was 4.5° (SD 3°; 0° to 12°) and 4° (SD 3.5°; 1° to 12°), respectively. Malpositioning (Δ > 10°) occurred in five hips (10.6%). Native COR reconstruction was planned in 42 cases (93%), and the mean 3D deviation vector was 15.5 mm (SD 8.5; 4 to 35). There was no significant influence in malpositioning found for femoral stem retention, surgical approach, or fixation method. Conclusion. At short-term follow-up, we found that PPR offers a viable solution for rTHA in cases with massive acetabular bone loss, as highly accurate positioning can be accomplished with meticulous planning, achieving anatomical reconstruction. Accuracy of achieved placement contributed to reduced complications with no injury to vital structures by screw fixation. Cite this article: Bone Joint J 2022;104-B(10):1110–1117

Aims. The aim of this study was to assess arthritis of the basal joint of the thumb quantitatively using bone single-photon emission CT/CT (SPECT/CT) and evaluate its relationship with patients’ pain and function. Methods. We retrospectively reviewed 30 patients (53 hands) with symptomatic basal joint arthritis of the thumb between April 2019 and March 2020. Visual analogue scale (VAS) scores for pain, grip strength, and pinch power of both hands and Patient-Rated Wrist/Hand Evaluation (PRWHE) scores were recorded for all patients. Basal joint arthritis was classified according to the modified Eaton-Glickel stage using routine radiographs and the CT scans of SPECT/CT, respectively. The maximum standardized uptake value (SUVmax) from SPECT/CT was measured in the four peritrapezial joints and the highest uptake was used for analysis. Results. According to Eaton-Glickel classification, 11, 17, 17, and eight hands were stage 0 to I, II, III, and IV, respectively. The interobserver reliability for determining the stage of arthritis was moderate for radiographs (k = 0.41) and substantial for CT scans (k = 0.67). In a binary categorical analysis using SUVmax, pain (p < 0.001) and PRWHE scores (p = 0.004) were significantly higher in hands with higher SUVmax. Using multivariate linear regression to estimate the pain VAS, only SUVmax (B 0.172 (95% confidence interval (CI) 0.065 to 0.279; p = 0.002) showed a significant association. Estimating the variation of PRWHE scores using the same model, only SUVmax (B 1.378 (95% CI, 0.082 to 2.674); p = 0.038) showed a significant association. Conclusion. The CT scans of SPECT/CT provided better interobserver reliability than routine radiographs for evaluating the severity of arthritis. A higher SUVmax in SPECT/CT was associated with more pain and functional disabilities of basal joint arthritis of the thumb. This approach could be used to complement radiographs for the evaluation of patients with this condition. Cite this article: Bone Joint J 2021;103-B(8):1380–1385

Aims. Glenoid bone loss is a significant problem in the management of shoulder instability. The threshold at which the bone loss is considered “critical” requiring bony reconstruction has steadily dropped and is now approximately 15%. This necessitates accurate measurement in order that the correct operation is performed. CT scanning is the most commonly used modality and there are a number of techniques described to measure the bone loss however few have been validated. The aim of this study was to assess the accuracy of the most commonly used techniques for measuring glenoid bone loss on CT. Methods. Anatomically accurate models with known glenoid diameter and degree of bone loss were used to determine the mathematical and statistical accuracy of six of the most commonly described techniques (relative diameter, linear ipsilateral circle of best fit (COBF), linear contralateral COBF, Pico, Sugaya, and circle line methods). The models were prepared at 13.8%, 17.6%, and 22.9% bone loss. Sequential CT scans were taken and randomized. Blinded reviewers made repeated measurements using the different techniques with a threshold for theoretical bone grafting set at 15%. Results. At 13.8%, only the Pico technique measured under the threshold. At 17.6% and 22.9% bone loss all techniques measured above the threshold. The Pico technique was 97.1% accurate, but had a high false-negative rate and poor sensitivity underestimating the need for grafting. The Sugaya technique had 100% specificity but 25% of the measurements were incorrectly above the threshold. A contralateral COBF underestimates the area by 16% and the diameter by 5 to 7%. Conclusion. No one method stands out as being truly accurate and clinicians need to be aware of the limitations of their chosen technique. They are not interchangeable, and caution must be used when reading the literature as comparisons are not reliable. Cite this article: Bone Jt Open 2023;4(7):478–489

Aims. The study objective was to prospectively assess clinical outcomes for a pilot cohort of tibial shaft fractures treated with a new tibial nailing system that produces controlled axial interfragmentary micromotion. The hypothesis was that axial micromotion enhances fracture healing compared to static interlocking. Methods. Patients were treated in a single level I trauma centre over a 2.5-year period. Group allocation was not randomized; both the micromotion nail and standard-of-care static locking nails (control group) were commercially available and selected at the discretion of the treating surgeons. Injury risk levels were quantified using the Nonunion Risk Determination (NURD) score. Radiological healing was assessed until 24 weeks or clinical union. Low-dose CT scans were acquired at 12 weeks and virtual mechanical testing was performed to objectively assess structural bone healing. Results. A total of 37 micromotion patients and 46 control patients were evaluated. There were no significant differences between groups in terms of age, sex, the proportion of open fractures, or NURD score. There were no nonunions (0%) in the micromotion group versus five (11%) in the control group. The proportion of fractures united was significantly higher in the micromotion group compared to control at 12 weeks (54% vs 30% united; p = 0.043), 18 weeks (81% vs 59%; p = 0.034), and 24 weeks (97% vs 74%; p = 0.005). Structural bone healing scores as assessed by CT scans tended to be higher with micromotion compared to control and this difference reached significance in patients who had biological comorbidities such as smoking. Conclusion. In this pilot study, micromotion fixation was associated with improved healing compared to standard tibial nailing. Further prospective clinical studies will be needed to assess the strength and generalizability of any potential benefits of micromotion fixation. Cite this article: Bone Jt Open 2021;2(10):825–833

Aims. To benchmark the radiation dose to patients during the course of treatment for a spinal deformity. Methods. Our radiation dose database identified 25,745 exposures of 6,017 children (under 18 years of age) and adults treated for a spinal deformity between 1 January 2008 and 31 December 2016. Patients were divided into surgical (974 patients) and non-surgical (5,043 patients) cohorts. We documented the number and doses of ionizing radiation imaging events (radiographs, CT scans, or intraoperative fluoroscopy) for each patient. All the doses for plain radiographs, CT scans, and intraoperative fluoroscopy were combined into a single effective dose by a medical physicist (milliSivert (mSv)). Results. There were more ionizing radiation-based imaging events and higher radiation dose exposures in the surgical group than in the non-surgical group (p < 0.001). The difference in effective dose for children between the surgical and non-surgical groups was statistically significant, the surgical group being significantly higher (p < 0.001). This led to a higher estimated risk of cancer induction for the surgical group (1:222 surgical vs 1:1,418 non-surgical). However, the dose difference for adults was not statistically different between the surgical and non-surgical groups. In all cases the effective dose received by all cohorts was significantly higher than that from exposure to natural background radiation. Conclusion. The treatment of spinal deformity is radiation-heavy. The dose exposure is several times higher when surgical treatment is undertaken. Clinicians should be aware of this and review their practices in order to reduce the radiation dose where possible. Cite this article: Bone Joint J 2021;103-B(4):1–7

Aims. To report the development of the technique for minimally invasive lumbar decompression using robotic-assisted navigation. Methods. Robotic planning software was used to map out bone removal for a laminar decompression after registration of CT scan images of one cadaveric specimen. A specialized acorn-shaped bone removal robotic drill was used to complete a robotic lumbar laminectomy. Post-procedure advanced imaging was obtained to compare actual bony decompression to the surgical plan. After confirming accuracy of the technique, a minimally invasive robotic-assisted laminectomy was performed on one 72-year-old female patient with lumbar spinal stenosis. Postoperative advanced imaging was obtained to confirm the decompression. Results. A workflow for robotic-assisted lumbar laminectomy was successfully developed in a human cadaveric specimen, as excellent decompression was confirmed by postoperative CT imaging. Subsequently, the workflow was applied clinically in a patient with severe spinal stenosis. Excellent decompression was achieved intraoperatively and preservation of the dorsal midline structures was confirmed on postoperative MRI. The patient experienced improvement in symptoms postoperatively and was discharged within 24 hours. Conclusion. Minimally invasive robotic-assisted lumbar decompression utilizing a specialized robotic bone removal instrument was shown to be accurate and effective both in vitro and in vivo. The robotic bone removal technique has the potential for less invasive removal of laminar bone for spinal decompression, all the while preserving the spinous process and the posterior ligamentous complex. Spinal robotic surgery has previously been limited to the insertion of screws and, more recently, cages; however, recent innovations have expanded robotic capabilities to decompression of neurological structures. Cite this article: Bone Jt Open 2024;5(9):809–817