现代医药卫生2025，Vol.41Issue(12)：2737-2741,2749,6.DOI:10.3969/j.issn.1009-5519.2025.12.001
AI辅助半自动分割软件在TACE术后肝脏恶性肿瘤病灶分割中的应用研究
Application of AI-assisted semi-automatic segmentation software in liver malignant tumor segmentation after transarterial chemoembolization
摘要
Abstract
Objective To explore the application effect of artificial intelligence(AI)assisted semi-auto-matic segmentation software in the segmentation of liver malignant tumor lesions after transarterial chemo-embolization(TACE)surgery.Methods A total of 37 patients with liver malignant tumors after TACE sur-gery at Hubei Provincial Hospital of Traditional Chinese Medicine from March to December 2022 were collect-ed.The manual and semi-automatic segmentationMethods were used to repeatedly measure the patients'plain scan and enhanced CT image data.The intraclass correlation coefficient(ICC)and 95%confidence inter-val(CI)were used to evaluate the consistency and variability of the longest diameter and volume measure-ments of tumor lesions within and between observers.Results The longest diameters of 37 tumor lesions were mainly distributed below 50 mm(81.08%),and the volumes were mostly below 50 mm3(72.97%).The distribution of tumor lesions in the left and right lobes of the liver were distributed relatively evenly.The 95%CI of manual and semi-automatic measurement of the longest diameter of tumor lesions were-1.12 to 1.13 and-1.43 to 0.89,respectively,with ICC values of 0.921 and 0.986,respectively.While the 95%CI between observers was-1.05 to 1.21 and-1.33 to 0.80,with ICC values of 0.801 and 0.883.The 95%CI of manual and semi-automatic measurement of tumor lesion volume were-40.34 to 30.20 and-8.03 to 8.93,respec-tively,with ICC values of 0.914 and 0.963,respectively.While the 95%CI of the inter-observer measurements were-56.22 to 54.66 and-28.08 to 31.27,with ICC values of 0.843 and 0.905 respectively.The manual measurement time of single tumor lesion was(26.77±5.04)seconds,while the semi-automatic measurement time was(6.62±2.95)seconds,and the difference was statistically significant(t=-24.88,P<0.01).Conclu-sion The AI-assisted semi-automatic segmentation software demonstrated excellent intra-observer and inter-observer consistency in the segmentation of malignant tumor lesions in the liver after TACE,and significantly reduced the operation time.关键词
人工智能/肝脏/经肝动脉化疗栓塞术/恶性肿瘤Key words
Artificial intelligence/Liver/Transarterial chemoembolization surgery/Malignant tumor分类
医药卫生引用本文复制引用
肖诗银,肖兰,卢翼,杨磊,王思琪,李锦巍,陈友三..AI辅助半自动分割软件在TACE术后肝脏恶性肿瘤病灶分割中的应用研究[J].现代医药卫生,2025,41(12):2737-2741,2749,6.基金项目
湖北省中医院院内科研项目(2021YJKT-19). （2021YJKT-19）
