圣盖伦国际乳腺癌共识性临床决策验证:深度大语言模型输出与全球专家小组建议之间的一致性评估
Yi Pan1, Chenglong Duan1, Jinsui Du1
1Department of Breast Surgery, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, Shaanxi Province, China.
Annals of surgical oncology
|February 10, 2026
概括
DeepSeek的大型语言模型 (LLM) 显示了与乳腺癌专家共识的中度一致. 与Gemini 2.0 Pro和ChatGPT-4o相比,它显示出更高的响应稳定性,这表明临床决策支持的巨大潜力.
科学领域:
- 在瘤学中使用人工智能
- 医疗信息学 医疗信息学
- 计算生物学 计算生物学
背景情况:
- 新兴的大型语言模型 (LLM) DeepSeek 显示出对医疗应用的希望.
- 对DeepSeek在乳腺癌临床决策中的表现进行系统评估,特别是对专家共识和领先的LLM进行系统评估,是有限的.
研究的目的:
- 评估DeepSeek LLM与国际乳腺癌专家共识的一致性.
- 在乳腺癌临床场景中比较DeepSeek与Gemini 2.0 Pro和ChatGPT-4o的性能.
主要方法:
- 分析了来自第19届圣加伦国际乳腺癌会议 (SG-BCC) 的139个共识问题.
- 在DeepSeek,Gemini 2.0 Pro和ChatGPT-4o模型中,每个问题都回答了五次.
- 使用一致率,答案稳定性和统计措施 (皮尔森相关性,绝对比例差异) 来进行比较.
主要成果:
- 在专家的共识下,DeepSeek-V3 实现了 63.31% 的整体一致率.
- DeepSeek-V3显示了高的平均答案稳定性 (86.69%),并且在稳定性方面表现优于Gemini 2.0 Pro和ChatGPT-4o (p < 0.001).
- 对于频繁的答案,DeepSeek-V3的一致率与双子座2.0 Pro和ChatGPT-4o (p = 0.849) 相比.
结论:
- 深度搜索模型与乳腺癌专家小组的共识有中等一致性.
- DeepSeek在答案强度方面显示出显著的优势,这表明乳腺癌临床决策的潜力很高.
- 对DeepSeek在瘤学中的应用进行进一步研究是有必要的.
相关概念视频
01:20Kendall's Coefficient of Concordance
1.0K
Kendall's Coefficient of Concordance (W), also known as Kendall's W, is a non-parametric statistical measure used to assess the agreement or concordance between multiple raters or judges when they rank a set of items. It is often used when you have ordinal data (ranks) and you want to see if there is consistency or consensus among the raters. It is widely applied in research areas such as psychology, medicine, and social sciences, where multiple judges are asked to rank or rate subjects...
1.0K
01:50Global Climate Change
29.0K
Throughout its ~4.5 billion year history, the Earth has experienced periods of warming and cooling. However, the current drastic increase in global temperatures is well outside of the Earth’s cyclic norms, and evidence for human-caused global climate change is compelling. Paleoclimatology, the study of ancient climate conditions, provides ample evidence for human-caused global climate change by comparing recent conditions with those in the past.
29.0K
01:18Wood Panel Products
411
Wood panel products are essential materials used in construction for applications such as flooring, siding, and roofing, typically available in standard dimensions of 4 feet by 8 feet, with thicknesses varying from one-quarter of an inch to one and one-eighth inches. Among the most common types of wood panels is plywood, which is produced by gluing multiple layers of thin wood veneers under pressure. The grain of the outer veneers runs lengthwise, while the grains of the interior layers run...
411
01:16Language
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
01:29Reliability and Validity
14.1K
Reliability and validity are two important considerations that must be made with any type of data collection. Reliability refers to the ability to consistently produce a given result. In the context of psychological research, this would mean that any instruments or tools used to collect data do so in consistent, reproducible ways.
14.1K
02:00Internal Energy
36.9K
The total of all possible kinds of energy present in a substance is called the internal energy (U), sometimes symbolized as E. Suppose a system with initial internal energy, Uinitial, undergoes a change in energy (transfer of work or heat), and the final internal energy of the system is Ufinal. Change in internal energy equals the difference between Ufinal and Uinitial.
36.9K