人工智能辅助超声弹性成像在甲状腺结节良恶性鉴别中的应用

打开文本图片集

中图分类号 R445.1 R581文献标识码 A文章编号 2096-7721（2025）06-0910-06

Application AI-assisted ultrasound elastography in differentiating benign and malignant thyroid nodules

ZHOU Wei WANG Chuang ZHAN Wentao LIANG Runa

AbstractObjective：ToinvestigatetheapplicationA-assistedultrasoundelastographyindiferentiatingbenignandmaligant thyroidnodules.Methods：81patientswithpathologicallyconfirmedthyroidnodulesat fromJuly2019to February224underwentA-asistedultrasoundelastogaphy.Theresultserecomparedwithpathologicalfidings（thegoldstandard）， withthe Kappa testused toevaluateconsistencyROCcurves were plotedtoassessdagnosticefcacyResults：Among 96nodulesfrom 81patients，patologyidentifedaligantad3enignnodules.Aasistedelastogahyfiingmachedpathologalresultsin76 nodules （concordance rate： 79.16% ; missed diagnosis rate： 4.54% ）， showing strong consistency （Kappa value= 0.785， P<0.001 ）. Malignant noduleshowdistcohgd，ndeueauresrarilaisdeIilnigoulesade O～I.The sensitivity，specificity，and AUC AI-assisted elastography for differentiation were 93.42% 91.55% ，and 0.920（ 95% CI： 0.859～ 0.939），respectively， surpassing standalone AI or elastography P<0.05 ）.Conclusion： AI-assisted ultrasound elastography has significant advantage indiferentiatingbenignandmalignantthyroidndules.Itscoreclinicalvalueliesinleveragingdeplearingalgorits to acuratelynalelasograicauresablgientngaligantdentiatio.Thisppoachproidesliableostic references while reducing subjectivity and uncertainty.

Key wordsArtificial Intelligence; Thyroid Nodules; Elastography

甲状腺结节在临床中较为常见，随着高分辨率超声检查在体检和临床诊疗中的广泛应用，甲状腺结节检出率显著提高。（剩余10863字）