Large language models (LLMs) can provide more accurate and useful clinical histories for imaging orders than referring clinicians.
Key Details
- 1Study involved more than 28,000 deidentified UCSF patient records from 2012 to 2024.
- 2LLMs were prompted to generate clinical histories for imaging orders from patient notes.
- 320 radiologists (2-25 years experience) assessed the comprehensiveness, factuality, and conciseness of the indications.
- 4LLM-generated histories outperformed those written by referring providers in many cases, offering more accurate patient histories.
- 5Improved clinical context aids radiologists in protocol selection and diagnostic focus.
Why It Matters
Incomplete clinical context in imaging orders can compromise diagnostic accuracy and workflow efficiency. Integrating LLMs to enrich requisitions could enhance radiologists' ability to select optimal imaging protocols and deliver better patient care.
Source
Radiology Business
Related News
•Radiology Business
Rad Partners Wins $1.29M FDA Grant for AI Radiology Report Evaluation Study
Cognita Imaging, part of Rad Partners, has received a $1.29M FDA grant to develop a new framework using large language models (LLMs) to evaluate AI-generated radiology reports.
•Radiology Business
Real-World Study: Radiology AI Best in Emergency and Inpatient Settings
A commercial AI tool for intracranial aneurysm detection outperformed in inpatient and emergency settings but yielded limited benefits for outpatients in a major health system study.
•Radiology Business
New Rubric Enhances Safety of AI-Generated Radiology Summaries
Researchers developed a five-factor rubric to assess the safety and quality of AI-generated, patient-friendly radiology report summaries.