Search research articles

关于 JoVE

概览领导团队博客 JoVE 帮助中心

作者

出版流程编辑委员会范围与政策同行评审常见问题投稿

图书馆员

用户评价订阅访问资源图书馆顾问委员会常见问题

研究

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments 存档

教育

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual 教师资源中心教师网站

使用条款与条件

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序

Same author

Pearls and pitfalls in the imaging of blunt bowel and mesenteric injury.

Abdominal radiology (New York)·2026

Same author

A primer on the imaging approach to acute gastrointestinal bleeding.

Abdominal radiology (New York)·2026

Same author

Zero-shot Thoracic Oncologic History Generation for Radiologists Using Retrieval-augmented Large Language Model Pipeline.

Radiology·2026

Same author

Association Between Electronic Health Record-Based Nursing Workload and Turnover: Retrospective Cohort Study.

JMIR nursing·2026

Same author

Frame-wise multi-echo distortion correction for superior functional MRI.

Imaging neuroscience (Cambridge, Mass.)·2026

Same author

Mesoscale CISS imaging for the detection of dural defects in spinal CSF leaks: A retrospective case series.

Headache·2026

Search research articles

相关实验视频

Updated: Jan 11, 2026

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

机器学习模型的大规模评估在放射学报告中识别后续建议.

Pan Xiao¹, Xiaobing Yu¹, Sung Min Ha¹

¹Mallinckrodt Institute of Radiology, Washington University School of Medicine, 4525 Scott Ave, MSC 8225-0082-03, St Louis, MO 63110.

|November 11, 2025

概括

机器学习模型有效地识别了放射学报告的后续建议. GPT-4和LSTM模型显示出高性能,改善了患者护理和降低风险.

更多相关视频

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Published on: April 18, 2025

相关实验视频

Last Updated: Jan 11, 2026

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Published on: April 18, 2025

科学领域:

放射学和医学信息学
医疗保健中的人工智能
自然语言处理 (Natural Language Processing) 是一种自然语言处理.

背景情况:

放射学报告包含关键的随访建议,用于患者护理和降低风险.
目前在各种报告和模式中识别这些建议的方法有限.
开源大型语言模型 (LLM) 提供了自动推识别的潜力.

研究的目的:

评估机器学习 (ML) 模型,包括LLAMA3和GPT-4,用于识别放射学报告中的后续建议.
为了比较各种文本分类方法在放射学报告的不同部分的性能.
评估ML模型在外部和时间数据集上的概括能力.

主要方法:

追溯分析了49769份来自多种成像方式和注释方法的放射学报告.
评估了32种关于"发现"和"印象"部分的文本分类方法.
使用MIMIC-CXR数据库和机构CT报告测试模型概括.

主要成果:

一个混合生成-歧视模型 (Hybrid-google) 在"发现"部分获得了最高的F1分 (0.835分).
一个基于注意力的双向LSTM (AttBiLSTM-random) 在"印象"部分获得了最高的F1分数 (0.979分).
带有前提示的GPT-4显示出强烈的泛化,F1得分为0.969 (MIMIC-CXR) 和0.973 (机构CT).

结论:

ML模型显示了在放射学报告中对随访建议的分类自动化的巨大潜力.
不同的ML架构在识别不同报告部分的建议方面表现出色.
像GPT-4这样的先进模型为外部和时间数据概括提供了强大的性能.