Artificial intelligence (AI) has transitioned from the lab to the bedside, and it is increasingly being used in healthcare. Radiology and Radiography are on the frontline of AI implementation, because of the use of big data for medical imaging and diagnosis for different patient groups. Safe and effective AI implementation requires that responsible and ethical practices are upheld by all key stakeholders, that there is harmonious collaboration between different professional groups, and customised educational provisions for all involved. This paper outlines key principles of ethical and responsible AI, highlights recent educational initiatives for clinical practitioners and discusses the synergies between all medical imaging professionals as they prepare for the digital future in Europe. Responsible and ethical AI is vital to enhance a culture of safety and trust for healthcare professionals and patients alike. Educational and training provisions for medical imaging professionals on AI is central to the understanding of basic AI principles and applications and there are many offerings currently in Europe. Education can facilitate the transparency of AI tools, but more formalised, university-led training is needed to ensure the academic scrutiny, appropriate pedagogy, multidisciplinarity and customisation to the learners’ unique needs are being adhered to. As radiographers and radiologists work together and with other professionals to understand and harness the benefits of AI in medical imaging, it becomes clear that they are faced with the same challenges and that they have the same needs. The digital future belongs to multidisciplinary teams that work seamlessly together, learn together, manage risk collectively and collaborate for the benefit of the patients they serve.

This study aimed to systematically review the literature to synthesise and summarise the evidence surrounding the efficacy of artificial intelligence (AI) in classifying paediatric pneumonia on chest radiographs (CXRs). Following the initial search of studies that matched the pre-set criteria, their data were extracted using a data extraction tool, and the included studies were assessed via critical appraisal tools and risk of bias. Results were accumulated, and outcome measures analysed included sensitivity, specificity, accuracy, and area under the curve (AUC). Five studies met the inclusion criteria. The highest sensitivity was by an ensemble AI algorithm (96.3%). DenseNet201 obtained the highest level of specificity and accuracy (94%, 95%). The most outstanding AUC value was achieved by the VGG16 algorithm (96.2%). Some of the AI models achieved close to 100% diagnostic accuracy. To assess the efficacy of AI in a clinical setting, these AI models should be compared to that of radiologists. The included and evaluated AI algorithms showed promising results. These algorithms can potentially ease and speed up diagnosis once the studies are replicated and their performances are assessed in clinical settings, potentially saving millions of lives.

Purpose: DTI characterizes tissue microstructure and provides proxy measures of nerve health. Echo‐planar imaging is a popular method of acquiring DTI but is susceptible to various artifacts (e.g., susceptibility, motion, and eddy currents), which may be ameliorated via preprocessing. There are many pipelines available but limited data comparing their performance, which provides the rationale for this study.Methods: DTI was acquired from the upper limb of heathy volunteers at 3T in blip‐up and blip‐down directions. Data were independently corrected using (i) FSL's TOPUP & eddy, (ii) FSL's TOPUP, (iii) DSI Studio, and (iv) TORTOISE. DTI metrics were extracted from the median, radial, and ulnar nerves and compared (between pipelines) using mixed‐effects linear regression. The geometric similarity of corrected b = 0 images and the slice matched T1‐weighted (T1w) images were computed using the Sörenson‐Dice coefficient.Results: Without preprocessing, the similarity coefficient of the blip‐up and blip‐down datasets to the T1w was 0·80 and 0·79, respectively. Preprocessing improved the geometric similarity by 1% with no difference between pipelines. Compared to TOPUP & eddy, DSI Studio and TORTOISE generated 2% and 6% lower estimates of fractional anisotropy, and 6% and 13% higher estimates of radial diffusivity, respectively. Estimates of anisotropy from TOPUP & eddy versus TOPUP were not different but TOPUP reduced radial diffusivity by 3%. The agreement of DTI metrics between pipelines was poor.Conclusions: Preprocessing DTI from the upper limb improves geometric similarity but the choice of the pipeline introduces clinically important variability in diffusion parameter estimates from peripheral nerves.

Introduction: To reduce the risks involved with ionising radiation exposure, typical values (TVs) and diagnostic reference levels (DRLs) have been established to help keep radiation doses 'as low as reasonably practicable. TVs/DRLs provide standardised radiation dose metrics that can be used for comparative purposes. However, for paediatrics, such values should consider the size of the child instead of their age. This study aimed to establish and compare paediatric TVs for chest, abdomen and pelvis radiography. Methods: Study methods followed processes for establishing paediatric DRLs as outlined by the Health Information and Quality Authority (HIQA). Kerma-area product (KAP) values, excluding rejected images, were retrospectively acquired from the study institution's Picture Archiving and Communications System (PACS). Paediatric patients were categorised into the following weight-based groupings (5 to <15 kg, 15 to <30 kg, 30 to <50 kg, 50 to 80 kg) and stratified based on the examination that was performed (chest, abdomen, and pelvis), and where it was performed (the different X-ray rooms). Anonymised data were inputted into Microsoft Excel for analysis. Median and 3rd quartile KAP values were reported together with graphical illustrations. Results: Data from 407 X-ray examinations were analysed. For the previously identified weight categories (5 to <15 kg, 15 to <30 kg, 30 to <50 kg, 50 to 80 kg), TVs for the chest were 0.10, 0.19, 0.37 and 0.53 dGy.cm(2), respectively. For the abdomen 0.39, 1.04, 3.51 and 4.05 dGy.cm(2) and for the pelvis 0.43, 0.87, 3.50 and 7.58 dGy.cm(2). Between X-ray rooms TVs varied against the institutional TVs by -60 to 119 % (chest), -50 to 103 % (abdomen) and -14 and 24 %% (pelvis). Conclusion: TVs in this study follow established trends with patient weight and examination type and are comparable with published literature. Variations do exist between individual examination rooms and reasons are multifactorial. Given that age and size do not perfectly correlate further work should be undertaken around weight-based TVs/DRLs in the paediatric setting.

Introduction: Previous literature reviews revealed that abdominal Xrays (AXR) performed for the accident and emergency department (A&E), had low sensitivity, high further imaging and non-alignment rate to the Royal College of Radiologists (RCR) guidelines. A study was performed to investigate the current practice with the aim of making recommendations to improve practice, which can reduce patients' radiation exposures, while can re-routing resources to other priorities. Methods: A study was performed in one of the UK's largest A&Es, in accordance with the RCR guidelines. All the AXR requests from A&E, regardless of the patient's age, within a 28 -day period, were retrospectively assessed. Non-A&E patients and abandoned examinations due to uncooperative patients were excluded. The total number of AXR requests received by the A&E imaging department was 169, with 28/169 falling into the exclusion criteria. Results: Of the 141 included requests, five unjustified requests were correctly rejected. The remaining 136 requests were accepted and performed, though only 115/136 (84.6%) of these were justified. The most common justified and unjustified indications were obstruction and renal stones, respectively. Only 4% of reported AXR had pathological abnormalities, while 45/136 patients had further imaging. Conclusions: The small proportion of significant findings echoed previous studies, suggesting an AXR overuse. Over 80% of noncompliant requests were performed, and awareness of the justification guidelines can be increased by clinical governance, posters, or an algorithm previously presented. The 32.4% further imaging rate recorded in this study, as opposed to the 73.7% reported in previous literature, merits attention. Implications to practice: Stopping the overuse of AXR can minimise the radiation dose received and relieve the mounting pressure in imaging and reporting, which can serve other patients who would benefit from the services otherwise.