机器人外科学杂志（中英文）2025，Vol.6Issue(6)：899-903,5.DOI:10.12180/j.issn.2096-7721.2025.06.004
二维超声联合人工智能对甲状腺结节良恶性的预测价值
Predictive value of two-dimensional ultrasound combined with artificial intelligence in differentiating benign and malignant thyroid nodules
摘要
Abstract
Objective:To investigate the predictive value of two-dimensional(2D)ultrasound combined with artificial intelligence(AI)for differentiating benign and malignant thyroid nodules.Methods:A retrospective analysis was conducted on 212 thyroid nodule patients who were treated at the Sunsimiao Hospital,Beijing University of Chinese Medicine from October 2021 to September 2024.Using surgical pathology as the gold standard,the diagnostic value,consistency with the gold standard,and accuracy of 2D ultrasound,AI,and their combination in differentiating benign and malignant thyroid nodules were evaluated.Results:Among 260 thyroid nodules,surgical pathology confirmed 105 malignant and 155 benign nodules.ROC curve analysis showed that the AUC values for 2D ultrasound and AI in diagnosing thyroid nodule benignity and malignancy were 0.830 and 0.745,with sensitivities of 86.67%and 83.81%,and specificities of 81.94%and 69.03%,respectively.The combined approach achieved an AUC of 0.872,sensitivity of 95.24%,and specificity of 81.94%.Using surgical pathology as the gold standard,the Kappa values for 2D ultrasound and AI were 0.672 and 0.511,indicating moderate agreement,while the combination yielded a Kappa value of 0.765,indicating good agreement.The combined method showed higher sensitivity and negative predictive value than either 2D ultrasound or AI alone;higher specificity,accuracy,and positive predictive value than AI alone;lower missed detection rate than both individual methods;and lower misdiagnosis rate than AI(P<0.05).Conclusion:Both 2D ultrasound and AI have diagnostic value for differentiating benign and malignant thyroid nodules,and their combination has superior predictive performance.关键词
甲状腺结节/良恶性/二维超声/人工智能/预测价值Key words
Thyroid Nodule/Benign and Malignant/Two-dimensional Ultrasound/Artificial Intelligence/Predictive Value分类
医药卫生引用本文复制引用
杨燕,丁治国,祁烁,王琪琪,张肖梅,张武平..二维超声联合人工智能对甲状腺结节良恶性的预测价值[J].机器人外科学杂志（中英文）,2025,6(6):899-903,5.基金项目
铜川市甲状腺病防治中心2022年度中医药科研项目(TJF-2022-09) （TJF-2022-09）
铜川市甲状腺病防治中心2023年度科研专项(TJF-QN-2023-07)Traditional Chinese Medicine Research Project of Tongchuan City Thyroid Disease Prevention and Treatment Center in 2022(TJF-2022-09) （TJF-QN-2023-07）
Special Research Project of Tongchuan City Thyroid Disease Prevention and Treatment Center in 2023(TJF-QN-2023-07) （TJF-QN-2023-07）
