文献速递细胞病理与分子病理

AI辅助数字甲状腺FNA细胞学:提升高风险Bethesda分类一致性与敏感性并优化筛查效率

Artificial intelligence-assisted digital thyroid FNA cytology: Improved agreement and sensitivity for higher-risk Bethesda categories with enhanced screening efficiency.
PubMed 文献 · Cancer Cytopathol · 2026-01-01

研究评估了疾病特异性深度学习算法AIxTHY在数字细胞学平台辅助甲状腺细针穿刺(FNA)诊断的临床可行性。两名细胞病理学家和一名细胞技师用显微镜、AI辅助单层WSI(S-WSI)和AI辅助七层Z-stack WSI(7-WSI)三种方式独立复阅100张ThinPrep FNA玻片。AI辅助提高了与专家共识诊断的一致性,尤其在高风险分类(TBS III+),将二分类风险分层的敏感性从60%提升至78%-79%,准确率从61%提升至约71%,但特异性因假阳性增加而下降;AI辅助将玻片复阅时间显著减少40%-65%(p<0.01)。

要点

  • AI辅助可提高甲状腺FNA高风险Bethesda分类(TBS III+)的诊断敏感性和与共识的一致性。
  • AI辅助将玻片复阅时间减少40%-65%,阳性病例效率提升最明显。
  • 特异性因假阳性增加而下降,尤其在不明确病例中,需进一步优化。

核心结论

  • AIxTHY是改善数字甲状腺细胞病理检测和工作流程效率的可行辅助工具,但需进一步优化不明确病变的特异性。

研究设计与样本

  • 回顾性研究,纳入100张档案ThinPrep FNA玻片,由两名细胞病理学家和一名细胞技师独立复阅。
  • 采用三种阅片方式:显微镜、AI辅助单层WSI(S-WSI)和AI辅助七层Z-stack WSI(7-WSI),洗脱期2周。
  • 以专家细胞学共识诊断为金标准,比较TBS分类并记录复阅时间。

主要发现

  • AI辅助阅片提高了与共识诊断的总体一致性,尤其在高风险分类(TBS III+:意义不明的异型性、滤泡性肿瘤和恶性),并减少了与显微镜相比的降级。
  • 对于二分类风险分层(TBS III+ vs TBS II),AI辅助将敏感性从60%提高到78%至79%,准确率从61%提高到约71%。
  • 特异性因假阳性分类增加而下降,尤其在不明确病例中。
  • AI辅助阅片使各诊断类别的玻片复阅时间显著减少40%至65%(p < 0.01),阳性病例效率提升最大。
  • 与S-WSI相比,7-WSI未进一步改善诊断一致性或二分类性能,但额外减少了复阅时间。

临床怎么用

  • AIxTHY可作为数字甲状腺细胞病理的辅助工具,用于提高高风险分类的检出和 workflow 效率。
  • 在不明确病变中需注意特异性下降和假阳性增加的风险。

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: Accurate cytologic classification of thyroid nodules is essential for clinical management, but interobserver variability and indeterminate interpretations remain persistent challenges. The clinical feasibility of AIxTHY, a disease-specific deep-learning algorithm integrated into a digital cytology platform, was evaluated for assisting thyroid fine-needle aspiration (FNA) diagnosis using whole-slide imaging (WSI). METHODS: Two cytopathologists and one cytologist independently reviewed 100 archival ThinPrep FNA slides using three modalities: microscopy, artificial intelligence (AI)-assisted single-layer WSI (S-WSI), and AI-assisted seven-layer Z-stack WSI (7-WSI), with 2-week washout intervals. Reviewers assigned The Bethesda System (TBS) categories, and review times were recorded. Diagnoses were compared with expert cytologic consensus as ground truth. RESULTS: AI-assisted review modalities improved overall agreement with consensus diagnoses, particularly for higher-risk categories (TBS III+: atypia of undetermined significance, follicular neoplasm, and malignant), and reduced downgrading compared to microscopy. For binary risk stratification (TBS III+ vs TBS II), AI assistance increased sensitivity from 60% to 78% to 79% and accuracy from 61% to approximately 71%, whereas specificity decreased because of increased false-positive classifications, particularly among indeterminate cases. AI-assisted review significantly reduced slide review time by 40% to 65% across diagnostic categories (p < 0.01), with the greatest efficiency gains in positive cases. Compared with S-WSI, 7-WSI did not further improve diagnostic concordance or binary classification performance but provided additional reductions in review time. CONCLUSIONS: AIxTHY is a feasible adjunct for improving detection and workflow efficiency in digital thyroid cytopathology, although further refinement is needed to improve specificity in indeterminate lesions.

来源:Cancer Cytopathol,查看原文。版权归原作者所有。

引用

  1. Satturwar S, Li Z, Yang CS, et al. Artificial intelligence-assisted digital thyroid FNA cytology: Improved agreement and sensitivity for higher-risk Bethesda categories with enhanced screening efficiency. Cancer Cytopathol. 2026. PMID: 42779199. doi:10.1002/cncy.70149

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。