Explore the vast range of titles available.

Most artificial intelligence (AI) research and innovations have concentrated in high-income countries, where imaging data, IT infrastructures and clinical expertise are plentiful. However, slower progress has been made in limited-resource environments where medical imaging is needed. For example, in Sub-Saharan Africa, the rate of perinatal mortality is very high due to limited access to antenatal screening. In these countries, AI models could be implemented to help clinicians acquire fetal ultrasound planes for the diagnosis of fetal abnormalities. So far, deep learning models have been proposed to identify standard fetal planes, but there is no evidence of their ability to generalise in centres with low resources, i.e. with limited access to high-end ultrasound equipment and ultrasound data. This work investigates for the first time different strategies to reduce the domain-shift effect arising from a fetal plane classification model trained on one clinical centre with high-resource settings and transferred to a new centre with low-resource settings. To that end, a classifier trained with 1792 patients from Spain is first evaluated on a new centre in Denmark in optimal conditions with 1008 patients and is later optimised to reach the same performance in five African centres (Egypt, Algeria, Uganda, Ghana and Malawi) with 25 patients each. The results show that a transfer learning approach for domain adaptation can be a solution to integrate small-size African samples with existing large-scale databases in developed countries. In particular, the model can be re-aligned and optimised to boost the performance on African populations by increasing the recall to 0.92 ± 0.04 and at the same time maintaining a high precision across centres. This framework shows promise for building new AI models generalisable across clinical centres with limited data acquired in challenging and heterogeneous conditions and calls for further research to develop new solutions for the usability of AI in countries with fewer resources and, consequently, in higher need of clinical support.

We aimed to develop and evaluate Explainable Artificial Intelligence (XAI) for fetal ultrasound using actionable concepts as feedback to end-users, using a prospective cross-center, multi-level approach. We developed, implemented, and tested a deep-learning model for fetal growth scans using both retrospective and prospective data. We used a modified Progressive Concept Bottleneck Model with pre-established clinical concepts as explanations (feedback on image optimization and presence of anatomical landmarks) as well as segmentations (outlining anatomical landmarks). The model was evaluated prospectively by assessing the following: the model’s ability to assess standard plane quality, the correctness of explanations, the clinical usefulness of explanations, and the model’s ability to discriminate between different levels of expertise among clinicians. We used 9352 annotated images for model development and 100 videos for prospective evaluation. Overall classification accuracy was 96.3%. The model’s performance in assessing standard plane quality was on par with that of clinicians. Agreement between model segmentations and explanations provided by expert clinicians was found in 83.3% and 74.2% of cases, respectively. A panel of clinicians evaluated segmentations as useful in 72.4% of cases and explanations as useful in 75.0% of cases. Finally, the model reliably discriminated between the performances of clinicians with different levels of experience (p- values < 0.01 for all measures) Our study has successfully developed an Explainable AI model for real-time feedback to clinicians performing fetal growth scans. This work contributes to the existing literature by addressing the gap in the clinical validation of Explainable AI models within fetal medicine, emphasizing the importance of multi-level, cross-institutional, and prospective evaluation with clinician end-users. The prospective clinical validation uncovered challenges and opportunities that could not have been anticipated if we had only focused on retrospective development and validation, such as leveraging AI to gauge operator competence in fetal ultrasound.

Clinicians are increasingly using point-of-care ultrasonography for bedside examinations of patients. However, proper training is needed in this technique, and it is unknown whether the skills learned from focused Ultrasonography courses are being transferred to diagnostic performance on patients. Thirty-one physicians were randomized to participate in a focused Ultrasonography course or control circumstances before they examined 4 patients with different abdominal conditions by ultrasonography. Performance scores and diagnostic accuracy were compared using independent samples t test and binary logistic regression, respectively. There was a significant difference in the performance score between the intervention group (27.4%) and the control group (18.0%, P = .004) and the diagnostic accuracy between the intervention group (65%) and the control group (39%, P = .014). Clinicians could successfully transfer learning from an Ultrasonography course to improve diagnostic performance on patients. However, our results also indicate a need for more training when new technologies such as point-of-care ultrasonography are introduced. •Successful transfer from ultrasonographic training to diagnostic performance on patients.•Clinicians can use point-of-care ultrasonography to increase diagnostic accuracy.•Formalized courses should be an integrated part of the initial ultrasonographic training.•Courses longer than 4 hours are needed when ultrasonography is introduced.

Introduction Using pre-procedure analgesia with the risk of apnoea may complicate the Less Invasive Surfactant Administration (LISA) procedure or reduce the effect of LISA. Methods The NONA-LISA trial (ClinicalTrials.gov, NCT05609877) is a multicentre, blinded, randomised controlled trial aiming at including 324 infants born before 30 gestational weeks, meeting the criteria for surfactant treatment by LISA. Infants will be randomised to LISA after administration of fentanyl 0.5–1 mcg/kg intravenously (fentanyl group) or isotonic saline solution intravenously (saline group). All infants will receive standardised non-pharmacological comfort care before and during the LISA procedure. Additional analgesics will be provided at the clinician’s discretion. The primary outcome is the need for invasive ventilation, meaning mechanical or manual ventilation via an endotracheal tube, for at least 30 min (cumulated) within 24 h of the procedure. Secondary outcomes include the modified COMFORTneo score during the procedure, bronchopulmonary dysplasia at 36 weeks, and mortality at 36 weeks. Discussion The NONA-LISA trial has the potential to provide evidence for a standardised approach to relief from discomfort in preterm infants during LISA and to reduce invasive ventilation. The results may affect future clinical practice. Impact Pre-procedure analgesia is associated with apnoea and may complicate procedures that rely on regular spontaneous breathing, such as Less Invasive Surfactant Administration (LISA). This randomised controlled trial addresses the effect of analgesic premedication in LISA by comparing fentanyl with a placebo (isotonic saline) in infants undergoing the LISA procedure. All infants will receive standardised non-pharmacological comfort. The NONA-LISA trial has the potential to provide evidence for a standardised approach to relief from discomfort or pain in preterm infants during LISA and to reduce invasive ventilation. The results may affect future clinical practice regarding analgesic treatment associated with the LISA procedure.

The placenta is crucial to fetal well-being and it plays a significant role in the pathogenesis of hypertensive pregnancy disorders. Moreover, a timely diagnosis of placenta previa may save lives. Ultrasound is the primary imaging modality in pregnancy, but high-quality imaging depends on the access to equipment and staff, which is not possible in all settings. Convolutional neural networks may help standardize the acquisition of images for fetal diagnostics. Our aim was to develop a deep learning based model for classification and segmentation of the placenta in ultrasound images. We trained a model based on manual annotations of 7,500 ultrasound images to identify and segment the placenta. The model's performance was compared to annotations made by 25 clinicians (experts, trainees, midwives). The overall image classification accuracy was 81%. The average intersection over union score (IoU) reached 0.78. The model’s accuracy was lower than experts’ and trainees’, but it outperformed all clinicians at delineating the placenta, IoU = 0.75 vs 0.69, 0.66, 0.59. The model was cross validated on 100 2nd trimester images from Barcelona, yielding an accuracy of 76%, IoU 0.68. In conclusion, we developed a model for automatic classification and segmentation of the placenta with consistent performance across different patient populations. It may be used for automated detection of placenta previa and enable future deep learning research in placental dysfunction.

Ultrasound has become a core skill in many specialties. We evaluated the learning and assessment of ultrasound skills in Obstetrics-Gynaecology in a series of eight studies. In the clinical setting, we found that trainees as well as experienced clinicians struggle with technical aspects of performance such as image optimization. We examined how to improve these aspects of performance in the simulated setting by determining mastery learning levels and exploring learning curves for novices. We then examined how to improve the efficiency of training as well as transfer of learning through the use of dyad practice as compared with single practice. We found that the use of simulation-based training focusing on technical aspects of performance in addition to clinical training led to sustained improvements in performance after two months of clinical training in all aspects of performance. In addition, we found an interaction effect between initial simulation-based training and subsequent clinical training on trainees’ need for supervision. These findings suggest that simulation-based training can work as preparation for future learning rather than merely as added learning. Finally, we found that the use of simulation-based initial training led to a large decrease in patients’ discomfort, improvements in their perceived safety and confidence in their ultrasound operator. However, simulation-based training comes at a cost and in the final study we developed a model for conducting cost-effectiveness studies and provided data from an example study on how to link training costs with quality of care.

Ultrasound assessment of fetal size and growth is the mainstay of monitoring fetal well-being during pregnancy, as being small for gestational age (SGA) or large for gestational age (LGA) poses significant risks for both the fetus and the mother. This study aimed to enhance the prediction accuracy of abnormal fetal growth. We developed a deep learning model, trained on a dataset of 433,096 ultrasound images derived from 94,538 examinations conducted on 65,752 patients. The deep learning model performed significantly better in detecting both SGA (58% vs 70%) and LGA compared with the current clinical standard, the Hadlock formula (41% vs 55%), p < 0.001. Additionally, the model estimates were significantly less biased across all demographic and technical variables compared to the Hadlock formula. Incorporating key anatomical features such as cortical structures, liver texture, and skin thickness was likely to be responsible for the improved prediction accuracy observed.

The use of medical simulators for training technical and diagnostic skills has rapidly increased over the past decade. Yet, most available medical simulators have not been developed based on a structured evaluation of their intended uses but rather out of expected commercial value. Moreover, educators often struggle to access simulators because of cost or because no simulators have been developed for a particular procedure. In this report, we introduce “the V-model” as a conceptual framework to illustrate how simulator development can be guided by the intended uses in an iterative fashion. Applying a needs-based conceptual framework when developing simulators is important to increase the accessibility and sustainability of simulation-based medical education. It will minimize the developmental barriers and costs, while at the same time improving educational outcomes. Two new simulators for invasive ultrasound-guided procedures are used as examples, the chorionic villus sampling model and the ultrasound-guided aspiration trainer. Our conceptual framework and the use cases can serve as a template for future simulator development and documentation hereof.

Background Assessment of laypersons’ Paediatric Basic Life Support (PBLS) skills is important to ensure acquisition of effective PBLS competencies. However limited evidence exists on which PBLS skills are essential for laypersons. The same challenges exist with respect to the assessment of foreign body airway obstruction management (FBAOM) skills. We aimed to establish international consensus on how to assess laypersons’ PBLS and FBAOM skills. Methods A Delphi consensus survey was conducted. Out of a total of 84 invited experts, 28 agreed to participate. During the first Delphi round experts suggested items to assess laypersons’ PBLS and FBAOM skills. In the second round, the suggested items received comments from and were rated by 26 experts (93%) on a 5-point scale (1 = not relevant to 5 = essential). Revised items were anonymously presented in a third round for comments and 23 (82%) experts completed a re-rating. Items with a score above 3 by more than 80% of the experts in the third round were included in an assessment instrument. Results In the first round, 19 and 15 items were identified to assess PBLS and FBAOM skills, respectively. The ratings and comments from the last two rounds resulted in nine and eight essential assessment items for PBLS and FBAOM skills, respectively. The PBLS items included: “Responsiveness”,” Call for help”, “Open airway”,” Check breathing”, “Rescue breaths”, “Compressions”, “Ventilations“, “Time factor” and “Use of AED”. The FBAOM items included: “Identify different stages of foreign body airway obstruction”, “Identify consciousness”, “Call for help”, “Back blows“, “Chest thrusts/abdominal thrusts according to age”, “Identify loss of consciousness and change to CPR”, “Assessment of breathing” and “Ventilation”. Discussion For assessment of laypersons some PBLS and FBAOM skills described in guidelines are more important than others. Four out of nine of PBLS skills focus on airway and breathing skills, supporting the major importance of these skills for laypersons’ resuscitation attempts. Conclusions International consensus on how to assess laypersons’ paediatric basic life support and foreign body airway obstruction management skills was established. The assessment of these skills may help to determine when laypersons have acquired competencies. Trial registration Not relevant.