Background: Bone fractures are common in acute care, and point-of-care ultrasound (POCUS) is an emerging diagnostic tool that can be complementary to or even in some cases an alternative to X-ray imaging. With the rise of artificial intelligence (AI) model incorporation in diagnostic interpretations, this systematic review and meta-analysis aimed to evaluate the diagnostic performance of deep learning models for detecting bone fractures on ultrasound images, using radiographic imaging with expert interpretation as the reference standard. Methods: Comprehensive literature searches were conducted from inception to 15 August 2025, in databases including MEDLINE (Ovid interface), Embase (Ovid interface), CINAHL Plus with Full Text (EBSCOhost interface), Web of Science, ACM Digital Library, Scopus, and Google Scholar. We included studies evaluating deep learning models applied to ultrasound images or sweeps for fracture classification. The included studies predominantly evaluated pediatric cohorts presenting to emergency departments with suspected upper extremity fractures. Sensitivity and specificity were pooled using a bivariate random-effects model on the logit scale, and summary receiver operating characteristic (sROC) curve was constructed. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool. Results: A total of 580 papers were identified in the preliminary literature search, and 333 studies were screened after duplicate removal. Screening 27 full-text articles resulted in eight studies that were included in the review. Across 21 reported model evaluations from five studies with extractable diagnostic data, pooled sensitivity was 0.77 (95% CI: 0.69-0.83) and pooled specificity was 0.90 (95% CI: 0.86-0.94), with an overall sROC AUC of 0.91. Patient-level evaluations yielded higher pooled sensitivity (0.89; 95% CI: 0.81-0.94) and specificity (0.94; 95% CI: 0.81-0.98) compared to video-level analyses. Conclusions: AI-assisted ultrasound demonstrates promising diagnostic performance for fracture detection with generally high specificity and variable sensitivity across architectures and anatomic targets. Future work should prioritize prospective, externally validated studies with standardized acquisition protocols and clinically meaningful reporting at the patient level.

Background: Access to mentorship and interview preparation resources remains inconsistent across Canadian medical schools, potentially disadvantaging students applying to diagnostic radiology. The Canadian Association of Radiologists’ Medical Student Network (CAR MSN) hosts an annual virtual mock interview program pairing fourth-year and third-year medical students with diagnostic radiology residents for structured practice interviews and individualized feedback. Methods: This study used a pre- and post-survey design to evaluate the program’s impact on self-reported interview preparedness across 5 annual cycles (2022-2026). A total of 315 registrants from all 17 Canadian medical schools completed pre-event surveys, and 157 participated in the mock interviews. Paired pre- and post-event data were available from 54 participants and their level of preparedness for the Canadian Resident Matching Service (CaRMS) interviews was analyzed using the Wilcoxon signed-rank test. Results: Self-reported preparedness scores improved significantly, with median scores increasing from 3 (IQR 2-3) to 4 (IQR 4-4) on a 5-point Likert scale ( r = .85, P < .001). However, a 65.6% attrition rate was observed between surveys. Participants rated the mock interview as highly helpful, and 96.3% felt the questions were representative of CaRMS interviews. General perceived stressors related to the unfamiliarity with the programs and the interview process decreased after participation. Conclusion: Despite the attrition rate, these findings suggest that nationally accessible, resident-led mock interview programs can meaningfully improve interview preparedness and confidence for radiology applicants. Future implementation of mentorship-based initiatives may promote equitable access to radiology-specific career guidance.

It is often desirable to generalize medical imaging AI models trained with dense annotations to data acquired from different ultrasound scanners or clinical sites; however, retraining these models with new annotations is often difficult and costly. We examine this challenge in pediatric wrist fracture assessment using point-of-care ultrasound (POCUS), where fractures are common and can be effectively triaged via ultrasound. AI has shown radiologist-level performance for fracture detection, often aided by high-quality bony structure segmentation. However, due to significant domain shifts, models perform poorly on data from other centers or probes, and obtaining segmentation labels across devices is impractical due to manual annotation effort and data privacy concerns. To address this, we propose a target-informed self-supervised pretraining and model-ensemble strategy. Specifically, our approach combines masked image modeling (MIM) and contrastive learning to learn target-domain structural representations without labels, and introduces a confidence-aware infusion head to adaptively integrate predictions. The source dataset, collected with a Philips Lumify probe, contained dense labels, while the target dataset, acquired with a TeleMED portable probe, was unlabeled. The datasets were kept strictly separate throughout the entire process. Our method used labeled source data for supervised training and leveraged target-domain pretraining to improve generalization. On 318 images from 62 pediatric POCUS videos, this approach significantly improved cross-device performance, achieving over 6% Dice improvement on the target domain versus the baseline. These results demonstrate a label-efficient and privacy-preserving approach for cross-device-robust ultrasound AI, offering a framework that can be extended to multi-center studies or federated learning setups. GitHub Repository:https://github.com/yuyue2uofa/CrossDomainPOCUS.

Background/Objective: Recent North American guidelines suggest that CT is indicated for further evaluation where ultrasound (US) is negative, although the negative predictive value (NPV) of ultrasound in adult patients when the appendix is not seen remains unclear. To assess the negative predictive value (NPV) of ultrasound in adult patients when the appendix is not seen. Methods: A scoping review of MEDLINE and EMBASE was performed from inception to 13 May 2025 using PRISMA-ScR guidelines to identify studies evaluating the outcome of adult patients where the appendix is not seen on ultrasound, with preference for studies where there were no secondary signs of acute appendicitis (right lower quadrant free fluid, abscess, ileus, echogenic fat or regional lymphadenopathy). Original studies with at least 10 patients were included in the review. The reference standard included a combination of clinical follow-up, CT and/or pathology. Data synthesis was provided as a qualitative review of the existing literature. Results: Six studies were included in the review. The number of included patients range from 12 to 179 with a mean age of 29–38 years. Few studies reported the patient BMI. NPVs ranged from 80 to 90% for all indeterminate ultrasounds and 83 to 95% for studies where secondary signs of appendicitis were excluded (90 to 95% when non-surgical reference standards were included). Two studies reported NPVs of 96–100% when the pre-test probability was low. Conclusions: The NPV of indeterminate ultrasound for adult patients with right lower quadrant pain and no secondary signs of appendicitis is likely ≥90%. When combined with a low clinical suspicion, the NPV is likely >95%. The appropriateness of a subsequent CT indication when the appendix is not visualized on ultrasound should be determined on an individualized basis.