文献速递甲状腺疾病

人工智能在甲状腺癌诊断中的应用:2026年更新

Artificial Intelligence Applications in Thyroid Cancer Diagnosis: 2026 Update.
PubMed 文献 · Thyroid · 2026-01-07

本更新综述讨论人工智能在甲状腺结节、颈部淋巴结及细胞学或组织学标本评估中的作用,并为希望安装AI系统的临床实践提供指导。迄今已有六个评估甲状腺结节超声图像的AI平台获美国食品药品监督管理局批准,性能总体良好,尤其与经验较少的医生相比。多模态大语言模型也已在此背景下测试,但迄今结果不太令人印象深刻。

要点

  • 六个甲状腺结节超声AI平台已获FDA批准,性能总体良好。
  • 多模态大语言模型在甲状腺结节评估中结果尚不理想。
  • 评估和安装AI系统需考虑信息技术、供应商支持、有效性、可用性和财务五方面。

核心结论

  • 用于甲状腺结节评估的AI工具持续增多,评估淋巴结和玻片标本的商业软件可能很快面世。

研究设计与样本

  • 本更新综述讨论AI在甲状腺结节、颈部淋巴结及细胞学或组织学标本评估中的作用,并为希望安装AI系统的实践提供指导。

主要发现

  • 迄今已有六个评估甲状腺结节超声图像的AI平台获美国食品药品监督管理局批准,性能总体良好,尤其与经验较少的医生相比。
  • 多模态大语言模型也已在此背景下测试,但迄今结果不太令人印象深刻。
  • 评估淋巴结和活检或手术标本的软件显示出前景,但尚无系统进入市场。

临床怎么用

  • 评估和安装AI系统需考虑信息技术、供应商支持、有效性、可用性和财务五个方面。
  • 建议从产品选择开始,让所有这些领域具有主题专业知识的人员参与。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: In this updated review, we will discuss the role of artificial intelligence (AI) in the assessment of thyroid nodules, cervical lymph nodes, and cytology or histology specimens, followed by guidance for practices wishing to install AI systems. SUMMARY: To date, six AI platforms for assessing sonograms of thyroid nodules have been cleared by the United States Food and Drug Administration, with generally good performance, especially when compared with that of less-experienced physicians. Multimodality large language models have also been tested in this context, but with less impressive results so far. Software to evaluate lymph nodes and biopsy or surgical specimens shows promise, though no systems have yet reached the marketplace. The process of evaluating and installing an AI system in a practice requires consideration in five areas: information technology, vendor support, effectiveness, usability, and finance. Engagement of people with subject matter expertise in all these domains is recommended, beginning with product selection. CONCLUSIONS: AI tools for the evaluation of thyroid nodules continue to proliferate, and commercial software to assess lymph nodes and slide specimens may soon become available. Attention to detail is critical for successful implementation and operation.

来源:Thyroid,查看原文。版权归原作者所有。

引用

  1. Thomas J, Tessler FN. Artificial Intelligence Applications in Thyroid Cancer Diagnosis: 2026 Update. Thyroid. 2026. PMID: 41791884. doi:10.1177/10507256251412316

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。