通过可视化大型语言模型的胸部X射线解释的局限性,双子座1.0,双子座1.5 Pro,GPT-4 Turbo和GPT-4o
Chih-Hsiung Chen1, Chang-Wei Chen2, Kuang-Yu Hsieh1
1Department of Critical Care Medicine, Mennonite Christian Hospital, Hualien 970472, Taiwan.
Diagnostics (Basel, Switzerland)
|February 13, 2026
概括
大型语言模型 (LLM) 在解释胸部X射线 (CXR) 中表现有前途,但在小病变和特定细节方面扎. 双子1.5 Pro在性能方面领先,尽管需要进一步的AI开发才能完成CXR分析.
科学领域:
- 医疗成像中的人工智能
- 计算机视觉用于诊断放射学.
背景情况:
- 准确的胸部X射线 (CXR) 解释对于诊断至关重要,需要识别病变,诊断,位置,大小和数量.
- 在没有临床背景的情况下,仅用于图像CXR解释的可视化大语言模型 (LLM) 的有效性尚未得到充分证实.
研究的目的:
- 评估领先的LLM在仅基于图像数据的CXR解释方面的表现.
- 评估LLM在初级诊断和识别各种肺部疾病中的关键成像特征中的准确性.
主要方法:
- 通过Gemini 1.0,Gemini 1.5 Pro,GPT-4 Turbo和GPT-4o对13个诊断类别的247个CXR进行了分析.
- 评估了LLM输出的主要诊断准确性 (完全正确,部分正确,不正确) 和关键特征识别.
- 像成像视图和设备这样的非诊断特征被单独分析.
主要成果:
- 在检测大型,双边和多重病变以及突出的设备 (例如肺,起器) 方面,LLM表现出更高的灵敏度.
- 双子 1.5 Pro实现了最高的总体检测率,其次是双子 1.0,GPT-4o和GPT-4 Turbo.
- 模型的性能在描述视图和标记等成像特征方面有所不同,在识别特定的医学术语方面有一些局限性.
结论:
- LLM可以识别某些CXR诊断和特征,特别是在显著异常的情况下.
- 在检测小病变,确定横向性和复杂的诊断推理方面仍然存在局限性.
- 对于LLM来说,在没有附带临床信息的情况下可靠地解释CXR,需要进一步的进展.
更多相关视频
06:12Inducing Acute Liver Injury in Rats via Carbon Tetrachloride CCl4 Exposure Through an Orogastric Tube
Published on: April 28, 2020
12.0K
07:36Eye Tracking During Visually Situated Language Comprehension: Flexibility and Limitations in Uncovering Visual Context Effects
Published on: November 30, 2018
16.5K
相关概念视频
Limiting Reactant
70.6K
The relative amounts of reactants and products represented in a balanced chemical equation are often referred to as stoichiometric amounts. However, in reality, the reactants are not always present in the stoichiometric amounts indicated by the balanced equation.
70.6K
X-ray Crystallography
26.3K
The size of the unit cell and the arrangement of atoms in a crystal may be determined from measurements of the diffraction of X-rays by the crystal, termed X-ray crystallography.
Diffraction
Diffraction is the change in the direction of travel experienced by an electromagnetic wave when it encounters a physical barrier whose dimensions are comparable to those of the wavelength of the light. X-rays are electromagnetic radiation with wavelengths about as long as the distance between neighboring...
Diffraction
Diffraction is the change in the direction of travel experienced by an electromagnetic wave when it encounters a physical barrier whose dimensions are comparable to those of the wavelength of the light. X-rays are electromagnetic radiation with wavelengths about as long as the distance between neighboring...
26.3K
Physiological Pharmacokinetic Models: Blood Flow-Limited Versus Diffusion-Limited Models
353
Physiological pharmacokinetic models, often called flow-limited or perfusion models, typically assume a swift drug distribution between tissue and venous blood, creating a rapid drug equilibrium. This premise is based on the idea that drug diffusion is extremely fast, and the cell membrane presents no barrier to drug permeation. In this scenario, where no drug binding occurs, the drug concentration in the tissue equals that of the venous blood leaving the tissue. This greatly simplifies the...
353
Vision
60.3K
Vision is the result of light being detected and transduced into neural signals by the retina of the eye. This information is then further analyzed and interpreted by the brain. First, light enters the front of the eye and is focused by the cornea and lens onto the retina—a thin sheet of neural tissue lining the back of the eye. Because of refraction through the convex lens of the eye, images are projected onto the retina upside-down and reversed.
60.3K
Language
921
Language is a unique communication system that uses words and systematic rules to organize and transmit information. Unlike other forms of communication, which may involve postures, movements, odors, or vocalizations, language relies on symbols and grammar. This makes human communication distinct from that of other species, who also communicate but do not use language in the same way humans do.
Corballis and Suddendorf (2007) and Tomasello and Rakoczy (2003) highlight the role of language in...
Corballis and Suddendorf (2007) and Tomasello and Rakoczy (2003) highlight the role of language in...
921
Flail Chest-I
771
Overview of Flail Chest
Flail chest is a severe and potentially life-threatening condition characterized by the fracture of three or more adjacent ribs in multiple places. It is most commonly caused by direct impacts and trauma, such as motor vehicle accidents or injuries from a steering wheel impact. It can also occur due to falls in elderly individuals with osteoporosis, or assaults involving sharp objects.
Pathophysiology
The pathophysiology of flail chest is complex, involving fractures of...
Flail chest is a severe and potentially life-threatening condition characterized by the fracture of three or more adjacent ribs in multiple places. It is most commonly caused by direct impacts and trauma, such as motor vehicle accidents or injuries from a steering wheel impact. It can also occur due to falls in elderly individuals with osteoporosis, or assaults involving sharp objects.
Pathophysiology
The pathophysiology of flail chest is complex, involving fractures of...
771
