Related Experiment Video
Updated: Apr 30, 2026
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
1.3K
Enhancing radiology workflows through collaborative AI-assisted chest X-ray reporting using large vision-language
Chantal Pellegrini1,2, Ege Özsoy3,4, Florian T Gassert5
1School of Computation, Information and Technology, Technical University of Munich, Munich, Germany. chantal.pellegrini@gmail.com.
Insights Into Imaging
|April 28, 2026
Summary
Artificial intelligence (AI) assistance in radiology reporting significantly reduced chest X-ray interpretation writing time, especially for complex cases. This AI tool improved efficiency and radiologist satisfaction without compromising report quality, suggesting practical integration into clinical workflows.
Area of Science:
- Radiology
- Artificial Intelligence
- Medical Informatics
Background:
- Radiology reporting is time-consuming and crucial for patient care.
- Efficiency and satisfaction in radiology workflows are areas for improvement.
- Artificial intelligence (AI) tools show promise in assisting medical professionals.
Purpose of the Study:
- To evaluate AI-assisted reporting for chest X-rays.
- To assess improvements in reporting efficiency and radiologist satisfaction.
- To determine if AI assistance compromises report quality.
Main Methods:
- Retrospective study with three radiologists analyzing 50 chest X-rays.
- Comparison of reporting with and without AI assistance (large vision-language model - LVLM).
- Evaluation of writing time, suggestion acceptance, report quality, and user satisfaction via Likert scale.
Main Results:
- AI assistance reduced mean writing time by 7.80% (significant for complex cases: 18.34%).
- Efficiency gains correlated with suggestion acceptance, user-dependent (up to 27.24%).
- Report quality and length remained stable; radiologists rated usability highly (4.33) and desired regular use (4).
Conclusions:
- Collaborative AI assistance can enhance radiology reporting efficiency, particularly for complex cases.
- AI tools are well-received by radiologists and can be integrated into workflows without quality compromise.
- Further prospective validation is warranted for clinical implementation of AI-assisted reporting.