Aligning large language models with radiologists by reinforcement learning from AI feedback for chest CT reports.

Journal: European journal of radiology
PMID: 39954322

Abstract

BACKGROUND: Large language models (LLMs) often struggle to fully capture the nuanced preferences and clinical judgement of radiologists in medical report summarization even when fine-tuned on massive medical reports. This could lead to the generated radiology reports lacking the professionalism and sufficient quality required in critical diagnostic decision-making.

Authors

  • Lingrui Yang
    Department of Radiology, Guang'an men Hospital, China Academy of Chinese Medical Sciences, Beijing 100053, China. Electronic address: syxxylr@163.com.
  • Yuxing Zhou
    Central Research Institute, United Imaging Healthcare, Co., Ltd., Beijing 100094, China. Electronic address: yuxing.zhou@cri-united-imaging.com.
  • Jun Qi
    Plastic Surgery Hospital, Chinese Academy of Medical Science, Peking Union Medical College, Beijing, China.
  • Xiantong Zhen
    The University of Western Ontario, London, ON, Canada; Digital Image Group (DIG), London, ON, Canada.
  • Li Sun
    Icahn School of Medicine at Mount Sinai, New York, NY, USA.
  • Shan Shi
    Research Group of Integrated Metallic Nanomaterials Systems, Hamburg University of Technology, Hamburg, Germany; Institute of Materials Mechanics, Helmholtz-Zentrum Hereon, Geesthacht, Germany.
  • Qinghua Su
    Beijing Wuzi University, College of information, Beijing 101149, China. Electronic address: qinghuasu@126.com.
  • XueDong Yang

Keywords

External Resources

Popular Topics
Recent Journals