Related Experiment Video
Updated: Apr 3, 2026
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
1.3K
Multidimensional evaluation of large language models in radiology report readability.
Yunhai Mao1, Chunyan Wang1, Yuxin Li1
1Department of Radiology, the Third Hospital of Jilin University, Changchun, China.
NPJ Digital Medicine
|April 1, 2026
Summary
Large language models (LLMs) enhance radiology report readability for patients, with comprehension influenced by education and age. LLMs should assist radiologists, requiring personalized approaches for diverse patient demographics.
Area of Science:
- Medical Informatics
- Radiology Communication
- Artificial Intelligence in Healthcare
Background:
- Patient comprehension of radiology reports is often limited.
- Effective communication of medical information is crucial for patient outcomes.
- Large language models (LLMs) offer potential solutions for simplifying complex medical texts.
Purpose of the Study:
- To evaluate the impact of LLMs on radiology report readability.
- To assess how demographic factors influence patient comprehension of these reports.
- To compare the performance of different LLMs in generating patient-centered reports.
Main Methods:
- A sequential two-stage study design.
- Retrospective evaluation of 320 radiology reports.
- Clinical validation with 800 patients using simplified reports.
Main Results:
- All evaluated LLMs significantly improved report readability.
- Higher education and older age correlated with better patient comprehension.
- Simplified reports enhanced subjective/objective comprehension and reduced medical anxiety.
Conclusions:
- LLMs can improve patient understanding of radiology reports.
- Demographic factors necessitate tailored communication strategies.
- LLMs should augment, not replace, radiologists' communication efforts.
