Oct 2026· Journal of the American College of Emergency Physicians Open· Vol 7 5, pp.
100491
· 0 citations· 35 references
Medicine
TL;DR
To improve integration, stakeholders should begin any new POCUS AI development project by first examining the different domains where POCUS AI applications are most needed, including education, clinical practice, workflow, research, and administration.
Abstract
The use of artificial intelligence (AI) within medicine has increased dramatically in the past few years, and applications for point-of-care ultrasound (POCUS) have followed a similar trend. Although physicians believe that POCUS AI applications have the potential to improve clinical practice, adoption of current applications remains limited. A lack of outcomes-based evidence, bias within datasets, poor assimilation within current workflows, and regulatory uncertainty are some of the major barriers that lead to poor adoption rates. To improve integration, stakeholders (physician leaders, researchers, health care executives, and industry partners) should begin any new POCUS AI development project by first examining the different domains where POCUS AI applications are most needed, including education, clinical practice, workflow, research, and administration. A gap analysis with clearly defined outcomes should come next, followed by an examination of how the new POCUS AI application will integrate into existing workflows. Training data sets that are reflective of real-world scenarios, including limitations encountered by end users, are essential. POCUS AI applications that integrate with existing workflows, have explainable outputs, and have been codeveloped with end users will improve adoption. Although regulatory pathways are evolving, engaging regulators early in the process and identifying viable reimbursement pathways are key strategies that will improve both the development and adoption of POCUS AI applications in the future.
Deep vein thrombosis (DVT) is a significant global health burden. Point-of-care ultrasound (POCUS) enhances diagnostic access but is dependent on the operator. Artificial intelligence (AI) has been proposed to guide novices and standardize image acquisition, often within a hybrid model incorporating remote clinician review. This systematic review aimed to synthesize the diagnostic accuracy, clinical feasibility, and health economic impact of AI-guided POCUS for DVT diagnosis within AI-human hybrid models. This systematic review was conducted in accordance with PRISMA guidelines, searching databases (PubMed/Medline, Scopus, Web of Science, IEEE, and Google Scholar) for studies published between 2021 and 2025. Four studies met the eligibility criteria, assessing AI-guided POCUS systems for DVT diagnosis. These studies demonstrated that the hybrid AI-human model achieved high diagnostic performance for ruling out proximal DVT, with a sensitivity of 90-100% and a negative predictive value (NVP) of 87.5-100%. Reviewer expertise significantly impacted accuracy; emergency medicine POCUS-trained physicians outperformed general radiologists. The model was feasible for non-expert operators (e.g., nurses) with minimal training and showed potential to reduce unnecessary duplex ultrasound referrals by 29–58%, with associated workflow efficiency. AI-guided POCUS combined with mandatory clinician review forms an effective hybrid model that enhances access, standardizes quality, and safely rules out DVT. Successful real-world implementation requires targeted reviewer training, workflow integration, and further health economic and independent pragmatic evaluation.
Syed Zulfiqar Ali Shah, Muhammad Kaleem Akhtar, Waqar Ahmad et al.· Veredas do Direito· 0 citations
Abstract Artificial intelligence (AI) is entering clinical practice faster than the evidence base supporting it. Clinicians, who remain the licensed decision-makers at the bedside, increasingly find themselves as end-users of tools whose strengths, failure modes, and external validity they have had no opportunity to assess. This article offers a practical framework that does not require coding or mathematical literacy. We outline how AI is built, validated, deployed, and monitored, and where each phase typically goes wrong. We propose seven questions that clinicians can run through to evaluate any clinical AI tool in the time it takes to read an abstract, alongside a traffic-light schema for matching oversight to risk and a short list of demands clinicians should make of vendors and institutions. We then examine the deeper questions of equity, accountability, and the therapeutic relationship that AI is now forcing into view. AI literacy belongs alongside biostatistics and evidence-based medicine as a core clinical competency.
Alaa Abdelqader, M. Alkhateeb, Abdullah Al-Marrawi et al.· Avicenna Journal of Medicine· 0 citations
Artificial Intelligence (AI) for detecting clinically significant prostate cancer (csPCa) on MRI has achieved diagnostic performance comparable to that of radiologists. By autonomously interpreting examinations, AI could improve workflow efficiency and help address increasing imaging demands and radiologist shortages. Despite this promise, autonomous AI has not been implemented in clinical practice. This narrative review explores remaining technical and societal barriers to deploying autonomous csPCa detection. We focus on three key domains: limitations in the current evidence base, safety issues and mitigation strategies, and the perspectives of patients and radiologists. Our findings highlight the need for evidence from large, multicenter, prospective trials and evaluation frameworks that reflect the consequences of clinical decision-making, as well as further exploration of safeguards to monitor and address mismatches between training data and incoming scans during deployment. Moreover, patients and radiologists show limited acceptance of autonomous AI, although this may improve with greater transparency, targeted education, and clearer guidelines on medico-legal responsibilities. Addressing these challenges is essential to the responsible deployment of autonomous AI and to realizing its efficiency gains in clinical practice.
Lisa D Koopmans, Fernando Vega Lara, Christian Roest et al.· Abdominal Radiology· 0 citations
Abstract This is a narrative review of recent peer-reviewed literature on artificial intelligence (AI) in emergency imaging spanning patient scheduling, image optimization, detection of critical findings, communication, and regulatory oversight. We review the current AI applications in emergency radiology and summarize clinically meaningful performance metrics. There are multiple instances of high performance in identifying imaging findings and prioritizing worklists, but limited prospective data demonstrating patient outcome improvements. We discuss the challenges in going from AI model performance to better outcomes, including the impact of false positives on workflow, ethical and regulatory issues, and implementation challenges. Practical tips for the evaluation, adoption, and monitoring of AI tools are outlined. Regulatory frameworks are in evolution and lag this rapidly changing landscape. Radiologist oversight remains critical at all steps to ensure safe and effective use of AI tools.
Anjali Agrawal, Fausto Labruto, Melissa A. Davis· Indian Journal of Radiology...· 0 citations
A new method, called CW-Net, translates the reasoning process of an autonomous vehicle’s AI system into understandable concepts that explain its behavior.
