大型语言模型能够将CT和MRI自由文本放射学报告翻译成多种语言
Aymen Meddeb1, Sophia Lüken1, Felix Busch1
1From the Departments of Neuroradiology (A.M., M.P.W.) and Radiology (S.L.), Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Berlin, Germany; Department of Neuroradiology, Hôpital Maison-Blanche, CHU Reims, Université Reims-Champagne-Ardenne, 45 Rue Cognacq-Jay, 51092 Reims, France (A.M.); Berlin Institute of Health at Charité-Universitätsmedizin Berlin, Berlin, Germany (A.M.); School of Medicine and Health, Department of Diagnostic and Interventional Radiology, Klinikum rechts der Isar, TUM University Hospital, Technical University of Munich, Munich, Germany (F.B., L.A., M.R.M., K.B.); Department of Advanced Biomedical Sciences, University of Naples Federico II, Naples, Italy (L.U.); Department of Radiology, Karolinska University Hospital, Stockholm, Sweden (E.K.); Department for Clinical Science, Intervention and Technology (CLINTEC), Division of Radiology, Karolinska Institute, Stockholm, Sweden (A.T.); Department of Radiology, National Institute Mongi Ben Hamida of Neurology, Tunis, Tunisia (S.J., I.D.); Department of Radiology, School of Medicine, University of Crete, Heraklion, Greece (M.E.K., M.T.); Computational Biomedicine Laboratory, Institute of Computer Science, Foundation for Research and Technology (FORTH), Heraklion, Greece (M.E.K.); Department of Radiology, University of Health Sciences, Basaksehir Cam and Sakura City Hospital, Basaksehir, Istanbul, Turkey (B.K.); Department of Radiology, Koc University Hospital, Istanbul, Turkey (S.Y.); Department of Radiology, Jinling Hospital, Affiliated Hospital of Medical School of Nanjing University, Nanjing, China (L.Z., B.H.); Laboratory for Digital Public Health Technologies, ITMO University, St Petersburg, Russian Federation (A.A., E.A.Y., T.L.); Department of Radiology, Chiang Mai University, Chiang Mai, Thailand (W.M., S.A.); Department of Medicine, Surgery, and Dentistry, University of Salerno, Baronissi, Italy (R.C.); and School of Medicine and Health, Institute for Cardiovascular Radiology and Nuclear Medicine, German Heart Center Munich, TUM University Hospital, Technical University of Munich, Munich, Germany (K.B.).
大型语言模型 (LLM) 在翻译放射学报告方面显示出高准确性,GPT-4在整体上表现最好. 虽然LLM提高了清晰度和可读性,但医学术语的准确性需要进一步发展.
科学领域:
- 医学成像和放射学医学成像和放射学
- 自然语言处理自然语言处理.
- 机器翻译 机器翻译
背景情况:
- 高质量的放射学报告翻译对于患者护理至关重要.
- 专家人类翻译人员的有限可用性需要探索替代解决方案.
- 大型语言模型 (LLM) 为这个领域的自动翻译提供了一个有希望的途径.
研究的目的:
- 为了评估各种LLM的准确性和质量,用于放射学报告翻译.
- 评估高资源和低资源语言的表现.
- 将LLM翻译与人类专家基准进行比较.
主要方法:
- 一个由100个合成放射学报告组成的数据集由18位放射学家翻译成9种语言.
- 十个LLM,包括GPT-4,Llama3和Mixtral,进行了自动翻译.
- 翻译质量使用BLEU,TER和chrF++指标进行评估,并进行放射科医生的定性审查.
主要成果:
- 总体来说,GPT-4的翻译质量优越,特别是在英语到德语,希腊语,泰语和土耳其语的翻译中.
- GPT-3.5在英语-法语方面表现出色,Qwen1.5在英语-中文方面表现出色,Mixtral 8x22B在意大利-英语方面表现出色.
- 在清晰度和可读性方面,LLM获得了高分,但在医学术语方面显示了中等的准确性.
结论:
- 在放射学报告翻译方面,LLM提供了高准确度和质量,尽管性能因模型和语言对而异.
- 需要进一步改进,以提高LLM翻译的医学术语准确性.
- 在放射学中,LLM是克服翻译障碍的宝贵工具.
更多相关视频
02:09Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function
Published on: April 12, 2024
13:12Translational Brain Mapping at the University of Rochester Medical Center: Preserving the Mind Through Personalized Brain Mapping
Published on: August 12, 2019
相关概念视频
Improving Translational Accuracy
Imaging Studies I: CT and MRI
Description of the Procedures
Computed Tomography (CT) scan:
Computed Tomography (CT) scans use X-ray technology to generate detailed images of bones, organs, and tissues. During the scan, the patient lies on a moving table...
Computed Tomography
The technique was invented in the 1970s and is based on the principle that as X-rays pass through the body, they are absorbed or reflected at different levels. In the technique, a patient lies on a motorized platform while a computerized axial tomography (CAT) scanner rotates...
Magnetic Resonance Imaging
