Search

An image-quality CT phantom was scanned with three different 3D X-ray imaging guidance devices in the operating theatre: O-Arm, Loop-X, and Airo TruCT. Default acquisition and reconstruction parameters for lumbar spine procedures were used on each device. The tube current was set to a dose level of around 27 mGy. A task-based image quality assessment was performed by calculating the noise power spectrum (NPS) and task transfer function (TTF). A detectability index (d') was calculated for three simulated bone lesions. The noise magnitude of the O-Arm was higher than the Airo TruCT, and the Loop-X had higher noise than the Airo TruCT. The highest average NPS frequency was for the O-Arm images, and the lowest was for the Loop-X. The TTFs at 50% values were similar for the Airo TruCT and Loop-X devices. Compared to Airo TruCT, the TTF at 50% value increased with the O-Arm by 53.12% and 41.20% for the Teflon and Delrin inserts, respectively. Compared to Airo TruCT, the d' value was lower with Loop-X by -26.73%, -27.02%, and -23.95% for lytic lesions, sclerotic lesions, and high-density bone, respectively. Each 3D-imaging spine surgery guidance device has its own strengths and weaknesses in terms of image quality. Cone-beam CT systems apparently offer the best compromise between noise and spatial resolution for spine surgery.

Rationale and Objectives: Interpreting radiographs in emergency settings is stressful and a burden for radiologists. The main objective was to assess the performance of three commercially available artificial intelligence (AI) algorithms for detecting acute peripheral fractures on radiographs in daily emergency practice., Materials and Methods: Radiographs were collected from consecutive patients admitted for skeletal trauma at our emergency department over a period of 2 months. Three AI algorithms-SmartUrgence, Rayvolve, and BoneView-were used to analyze 13 body regions. Four musculoskeletal radiologists determined the ground truth from radiographs. The diagnostic performance of the three AI algorithms was calculated at the level of the radiography set. Accuracies, sensitivities, and specificities for each algorithm and two-by-two comparisons between algorithms were obtained. Analyses were performed for the whole population and for subgroups of interest (sex, age, body region)., Results: A total of 1210 patients were included (mean age 41.3 ± 18.5 years; 742 [61.3%] men), corresponding to 1500 radiography sets. The fracture prevalence among the radiography sets was 23.7% (356/1500). Accuracy was 90.1%, 71.0%, and 88.8% for SmartUrgence, Rayvolve, and BoneView, respectively; sensitivity 90.2%, 92.6%, and 91.3%, with specificity 92.5%, 70.4%, and 90.5%. Accuracy and specificity were significantly higher for SmartUrgence and BoneView than Rayvolve for the whole population (P < .0001) and for subgroups. The three algorithms did not differ in sensitivity (P = .27). For SmartUrgence, subgroups did not significantly differ in accuracy, specificity, or sensitivity. For Rayvolve, accuracy and specificity were significantly higher with age 27-36 than ≥53 years (P = .0029 and P = .0019). Specificity was higher for the subgroup knee than foot (P = .0149). For BoneView, accuracy was significantly higher for the subgroups knee than foot (P = .0006) and knee than wrist/hand (P = .0228). Specificity was significantly higher for the subgroups knee than foot (P = .0003) and ankle than foot (P = .0195)., Conclusion: The performance of AI detection of acute peripheral fractures in daily radiological practice in an emergency department was good to high and was related to the AI algorithm, patient age, and body region examined., Competing Interests: Declaration of Competing Interest Valérie Bousson is a paid consultant for Milvue starting July 1, 2022. Grégoire Attané was paid by Gleamer between March and August 2019 to label radiographs. The remaining authors declare none. The three companies, Milvue, AZmed, and Gleamer, provided the AI algorithms to our radiology department for free. None of the three companies had access to the study methodology or data during the course of the study., (Copyright © 2023 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.)

The interest of researchers, clinicians and radiologists, in artificial intelligence (AI) continues to grow. Deep learning is a subset of machine learning, in which the computer algorithm itself can determine the optimal imaging features to answer a clinical question. Convolutional neural networks are the most common architecture for performing deep learning on medical images. The various musculoskeletal applications of deep learning are the detection of abnormalities on X-rays or cross-sectional images (CT, MRI), for example the detection of fractures, meniscal tears, anterior cruciate ligament tears, degenerative lesions of the spine, bone metastases, classification of e.g., dural sac stenosis, degeneration of intervertebral discs, assessment of skeletal age, and segmentation, for example of cartilage. Software developments are already impacting the daily practice of orthopedic imaging by automatically detecting fractures on radiographs. Improving image acquisition protocols, improving the quality of low-dose CT images, reducing acquisition times in MRI, or improving MR image resolution is possible through deep learning. Deep learning offers an automated way to offload time-consuming manual processes and improve practitioner performance. This article reviews the current state of AI in musculoskeletal imaging., (Copyright © 2022 Société française de rhumatologie. Published by Elsevier Masson SAS. All rights reserved.)

Objectives: A national retrospective survey on patient doses was performed by the French Society of Medical physicists to assess reference levels (RLs) in interventional radiology as required by the European Directive 2013/59/Euratom., Methods: Fifteen interventional procedures in neuroradiology, vascular radiology and osteoarticular procedures were analysed. Kerma area product (KAP), fluoroscopy time (FT), reference air kerma and number of images were recorded for 10 to 30 patients per procedure. RLs were calculated as the 3rd quartiles of the distributions., Results: Results on 4600 procedures from 36 departments confirmed the large variability in patient dose for the same procedure. RLs were proposed for the four dosimetric estimators and the 15 procedures. RLs in terms of KAP and FT were 90 Gm.cm 2 and 11 mins for cerebral angiography, 35 Gy.cm 2 and 16 mins for biliary drainage, 75 Gy.cm 2 and 6 mins for lower limbs arteriography and 70 Gy.cm 2 and 11 mins for vertebroplasty. For these four procedures, RLs were defined according to the complexity of the procedure. For all the procedures, the results were lower than most of those already published., Conclusions: This study reports RLs in interventional radiology based on a national survey. Continual evolution of practices and technologies requires regular updates of RLs., Key Points: • Delivered dose in interventional radiology depends on procedure, practice and patient. • National RLs are proposed for 15 interventional procedures. • Reference levels (RLs) are useful to benchmark practices and optimize protocols. • RLs are proposed for kerma area product, air kerma, fluoroscopy time and number of images. • RLs should be adapted to the procedure complexity and updated regularly.

The French regulations concerning the involvement of medical physicists in medical imaging procedures are relatively vague. In May 2013, the ASN and the SFPM issued recommendations regarding Medical Physics Personnel for Medical Imaging: Requirements, Conditions of Involvement and Staffing Levels. In these recommendations, the various areas of activity of medical physicists in radiology and nuclear medicine have been identified and described, and the time required to perform each task has been evaluated. Criteria for defining medical physics staffing levels are thus proposed. These criteria are defined according to the technical platform, the procedures and techniques practised on it, the number of patients treated and the number of persons in the medical and paramedical teams requiring periodic training. The result of this work is an aid available to each medical establishment to determine their own needs in terms of medical physics., (© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.)

Aims: To evaluate the effects of combined antiretroviral drugs (HAART) on liver CYP3A4 activity using the [(14)C-N-methyl]-erythromycin breath test (ERMBT)., Methods: HIV-infected patients (31 women, 30 men) with mean (+/- SD) age of 38 +/- 9 years were enrolled and underwent complete clinical and laboratory evaluation. Patients were divided into five groups and were treated with two nucleoside analogues (NAs) and one of the following: nelfinavir alone (n = 13), any ritonavir-boosted protease inhibitor with (n = 8) or without (n = 13) nevirapine, nevirapine alone (n = 15), or a third NA (n = 12). Three or four ERMBTs were performed 7 days prior to (D-7) and at the beginning of treatment (D0), D14 (only for patients taking nevirapine) and on D28., Results: Mean baseline liver CYP3A4 activity displayed high interindividual variability (47%) but low intraindividual variability (15%). Women had 30% higher ERMBT values than men [2.7 +/- 1.3 vs. 1.9 +/- 0.7; 95% confidence interval (CI) 20.5, 49.5; P = 0.003]. The ERMBT data correlated with body weight, alpha- and beta-globulins and alanin aminotransferases (0.10 < r(s) < 0.20; P < 0.01). Whereas nevirapine had no effect on liver CYP3A4 activity, nelfinavir-based and ritonavir-boosted drug regimens inhibited it by 69% (95% CI 64.7, 72.9; P = 0.005) and by 95% (95% CI 93.3, 96.7; P = 0.001), respectively., Conclusion: Evaluation of the effect of HAART on liver CYP3A4 activity may aid in preventing inappropriate treatment regimens in HIV-infected patients.

The [14C-N-methyl]-erythromycin breath test (14C-ERMBT) is one of the most valuable probes for liver cytochrome P450-3A4 activity in humans. In order to extend the use of this test in France, we herein provide safety data regarding either patient dosimetry or worker exposure to [14C-N-methyl]-erythromycin. In order to determine the maximum radiation exposure for patient and nuclear medicine technician following one intravenous 14C-ERMBT [111 kiloBequerel (kBq)], we have used the dosimetric data gathered in animal studies and extrapolated to humans using a weight-based method, approximate data provided by the French Society of Radioprotection and erythromycin pharmacokinetics in humans, considering always the worst conditions for the patient and worker exposure determination. The radioactivity administered to a patient after one 14C-ERMBT was equal to 108.8 kBq (i.e. 98% of the total radioactivity in the 14C-erythromycin vial) leading to a patient effective dose of 20 microsievert (microSv) and a maximum effective dose after 14CO2 inhalation by the exposed worker of 16 microSv compared with a mean individual annual effective dose from natural and artificial radioactivity exposure of 3500 microSv in France. The 14C-ERMBT is safe and complies with the European regulations regarding the administration of 14C-labelled compounds in humans. It can therefore be used in clinical research in France without any particular safety requirement.