基于临床与CT影像特征列线图鉴别结节性甲状腺肿与甲状腺乳头状癌  

A nomogram based on Clinical and CT imaging features for the differential diagnosis of nodular goiter and papillary thyroid carcinoma

作　　者：张海明 李恒国[2] 李琼华 马立恒 ZHANG Haiming;LI Hengguo;LI Qionghua;MA Liheng(Department of Medical Imaging,The First Affiliated Hospital of Guangdong Pharmaceutical University,Guangzhou 518100,Guangdong,China;Medical Imaging Center,The First Affiliated Hospital of Jinan University,Guangzhou 510620,Guangdong,China)

机构地区：[1]广东药科大学附属第一医院医学影像科,广东广州518100 [2]暨南大学附属第一医院医学影像中心,广东广州510620

出　　处：《右江民族医学院学报》2024年第5期707-713,共7页Journal of Youjiang Medical University for Nationalities

摘　　要：目的探讨基于临床与CT影像征象列线图鉴别结节性甲状腺肿(nodular goiter,NG)与甲状腺乳头状癌(papillary thyroid carcinoma,PTC)的临床应用价值。方法收集经手术病理证实的NG和PTC共333例,其中NG 160例,PTC 173例。所有患者按7∶3比例随机分为训练组(233例)和验证组(100例)。采用多因素Logistic回归分析筛选出鉴别NG与PTC的独立预测因素并构建列线图模型。结果多因素Logistic回归分析显示,FT 4、形态、咬饼征、增强后模糊或范围缩小、囊变是鉴别NG与PTC的独立预测因素(P均<0.01)。基于独立预测因素构建的列线图模型在训练组(AUC为0.939,灵敏度为0.944,特异度为0.810)和验证组(AUC为0.969,灵敏度为0.887,特异度为1.00)均取得良好的预测效能。校准曲线显示列线图模型拟合良好,决策曲线分析的高净收益也体现了模型的临床实用性。结论基于临床与CT影像征象列线图模型在鉴别NG与PTC方面具有较高的临床应用价值,为临床提供一种非创、简单、有效的可视化工具。Objective To explore the clinical application value of nomogram based on clinical and CT imaging features in differentiating nodular goiter(NG)from papillary thyroid carcinoma(PTC).Methods A total of 333 cases of surgically and pathologically confirmed NG(160 cases)and PTC(173 cases)were collected,and were randomly divided into training group(233 cases)and validation group(100 cases)with a ratio of 7∶3.Multivariate logistic regression analysis was conducted to screen independent predictors for differentiating NG from PTC and to construct a nomogram model.Results Multivariate Logistic regression analysis showed that FT 4,morphology,cookie bite sign,blurring or reduction in size after enhancement,and cystic degeneration were independent predictors for differentiating NG from PTC(all P<0.01).The nomogram model based on these independent predictors achieved good predictive performance in both the training group(AUC=0.939,sensitivity=0.944,specificity=0.810)and the validation group(AUC=0.969,sensitivity=0.887,specificity=1.00).The calibration curve showed that the nomogram model fitted well,and the high net benefit of decision curve analysis also reflected the clinical practicability of the model.Conclusion The nomogram model based on clinical and CT imaging features demonstrates significant clinical utility in the differential diagnosis of NG and PTC,offering a non-invasive,straightforward,and efficient visualization tool for clinical application.

关 键 词：列线图 甲状腺肿 结节性 甲状腺癌 乳头状 体层摄影术 X线计算机 

分 类 号：R581.3[医药卫生—内分泌] R736.1[医药卫生—内科学]