<oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
  <dc:creator>Arthaud, Jérémie</dc:creator>
  <dc:creator>Thoeny, Harriet C.</dc:creator>
  <dc:creator>Grek, Vincent</dc:creator>
  <dc:creator>Widmer, Lucien</dc:creator>
  <dc:date>2026</dc:date>
  <dc:description xmlns:ns0="xml" ns0:lang="en">Background Up to 30% of imaging examinations are deemed inappropriate, leading to unnecessary radiation exposure, 
emergency department (ED) overcrowding, and rising healthcare costs. While clinical decision support systems (CDSSs)
such as ESR iGuide aim to improve imaging appropriateness, their clinical impact and user adoption remain limited. Large
language models (LLMs) may offer a faster, more accessible alternative for radiologic decision-making.
Purpose To evaluate the performance of OpenAccessGPT in reducing inappropriate radiological examinations requested
from an ED by comparing its recommendations with ESR iGuide and assessing appropriateness and time efficiency.

Materials and methods This retrospective, single-center, semi-quantitative cohort study was conducted at a tertiary hospital
in Switzerland. A total of 201 consecutive ED radiology requests (March 2023–April 2024) met inclusion criteria (50 CT,
50 US, 50 X-rays, 51 MRI). Four hypotheses were tested: (1) OpenAccessGPT’s ability to reduce imaging requests; (2)
concordance with ESR iGuide; (3) time efficiency; and (4) the rate of inappropriate examinations in clinical practice when
confronted to ESR iGuide. Statistical analyses included proportion tests and paired t-tests (p &lt; .05).

Results Mean patient age was 55.3 years (SD = 20.9); 45% were female. Eleven cases were excluded due to unavailable
ESR iGuide scenarios. OpenAccessGPT advised against imaging in 6% of cases (12/201), with 33.3% agreement with ESR
iGuide (5/12). Overall concordance was 78.9% (150/190), below the 90% threshold (χ² = 25.79, p &lt; .001). GPT was significantly
faster (21.6 s vs. 101 s; Δ = 79.6 s; t[189] = 12.8; p &lt; .001). Institutional practice aligned with ESR iGuide in 82.6% of
cases (χ² = 11.46, p &lt; .001).

Conclusion OpenAccessGPT showed only limited ability to reduce inappropriate imaging. Despite substantial faster
response times, its clinical reliability remained insufficient for standalone use.</dc:description>
  <dc:format>application/pdf</dc:format>
  <dc:identifier>https://folia.unifr.ch/unifr/documents/337361</dc:identifier>
  <dc:identifier>https://folia.unifr.ch/documents/337361/files/s44326-026-00114-0.pdf</dc:identifier>
  <dc:language>eng</dc:language>
  <dc:relation>https://link.springer.com/article/10.1007/s44326-026-00114-0</dc:relation>
  <dc:relation>info:eu-repo/semantics/altIdentifier/doi/10.1007/s44326-026-00114-0</dc:relation>
  <dc:relation>info:eu-repo/semantics/altIdentifier/issn/3004-8613</dc:relation>
  <dc:rights>info:eu-repo/semantics/openAccess</dc:rights>
  <dc:rights>CC BY</dc:rights>
  <dc:source>Journal of Medical Imaging and Interventional Radiology. - Springer Science and Business Media LLC. - 2026, vol. 13, no. 1</dc:source>
  <dc:subject xmlns:ns1="xml" ns1:lang="en">Clinical decision support systems</dc:subject>
  <dc:subject xmlns:ns2="xml" ns2:lang="en">Large language models</dc:subject>
  <dc:subject xmlns:ns3="xml" ns3:lang="en">Generative pre-trained transformer</dc:subject>
  <dc:subject xmlns:ns4="xml" ns4:lang="en">ESR iGuide</dc:subject>
  <dc:subject xmlns:ns5="xml" ns5:lang="en">Emergency department</dc:subject>
  <dc:subject>info:eu-repo/classification/udc/616</dc:subject>
  <dc:title xmlns:ns6="xml" ns6:lang="en">AI as a clinical decision support system: accuracy of an open GPT-model in emergency practice</dc:title>
  <dc:type>http://purl.org/coar/resource_type/c_6501</dc:type>
</oai_dc:dc>
