切换至 "中华医学电子期刊资源库"

中华医学超声杂志(电子版) ›› 2018, Vol. 15 ›› Issue (10) : 773 -778. doi: 10.3877/cma.j.issn.1672-6448.2018.10.009

所属专题: 文献

浅表器官超声影像学

超声检查结合数学模型对TI-RADS 3~4类甲状腺结节良恶性的诊断
李璐1, 周娟1, 邓红艳1, 马雯婷1, 杭菁1, 叶新华1,(), 李勇2   
  1. 1. 210029 南京,南京医科大学第一附属医院超声医学科
    2. 210014 南京,江苏省农业科学院农产品质量安全与营养研究所
  • 出版日期:2018-10-01
  • 通信作者: 叶新华

Diagnosis of thyroid nodules by using ultrasound imaging analysis combined with mathematical models

Lu Li1, Juan Zhou1, Hongyan Deng1, Wenting Ma1, Jing Hang1, Xinhua Ye1,(), Yong Li2   

  1. 1. Department of Ultrasound, the First Affiliated Hospital, Nanjing Medical University, Nanjing 210029, China
    2. Institute of Food Quality and Safety, Jiangsu Academy of Agricultural Sciences, Nanjing 210014, China
  • Published:2018-10-01
  • Corresponding author: Xinhua Ye
  • About author:
    Corresponding author: Ye Xinhua, Email:
引用本文:

李璐, 周娟, 邓红艳, 马雯婷, 杭菁, 叶新华, 李勇. 超声检查结合数学模型对TI-RADS 3~4类甲状腺结节良恶性的诊断[J]. 中华医学超声杂志(电子版), 2018, 15(10): 773-778.

Lu Li, Juan Zhou, Hongyan Deng, Wenting Ma, Jing Hang, Xinhua Ye, Yong Li. Diagnosis of thyroid nodules by using ultrasound imaging analysis combined with mathematical models[J]. Chinese Journal of Medical Ultrasound (Electronic Edition), 2018, 15(10): 773-778.

目的

探讨超声影像结合数学模型在甲状腺结节良恶性判别诊断中的应用。

方法

回顾性分析2016年8月~12月在南京医科大学第一附属医院行超声检查发现的甲状腺结节,并均有手术或穿刺活检病理证实的患者共128例,纳入分析结节共170个。整理分析结节的超声图像特征信息,利用偏最小二乘-判别分析法(PLS-DA)和Logistic回归分析方法建立甲状腺结节良恶性风险预估模型并对其预测分析,然后采用逐步回归分析方法筛选出有统计学意义的特征变量,再次建模预测分析。

结果

在未删减自变量前,PLS-DA真阳性预测值(96.95%)、真阴性预测值(97.73%)较Logistic(89.86%、93.12%)高(P<0.05)。利用Stepwise方法筛选出重要变量包括甲状腺内部整体回声、形态、边缘、内部结构、强回声、纵横比和血管模式,再次建模后,PLS-DA真阳性预测率(98.12%)、真阴性预测率(98.49%)和Logistic真阳性预测率(95.09%)、真阴性预测率(95.31%)较筛选变量之前都有所提高(P<0.05),而且,PLS-DA真阳性预测率和真阴性预测率均明显高于Logistic(P<0.05)。

结论

PLS-DA和Logistic方法均可以构建甲状腺癌诊断模型,基于Stepwise筛选变量可以使诊断模型更加稳健,PLS-DA模型准确率结果要优于Logistic。

Objectives

In the present study, two mathematical models were constructed based on the characteristics of US image to discriminate the benign and malignant thyroid nodules.

Methods

A retrospective study was conducted in 128 patients with thyroid nodules from 2016/8 to 2016/12 in the First Affiliated Hospital of Nanjing Medical University. There were totally 170 pathologically-confirmed thyroid nodules. The gray scale image and color Doppler flow imaging (CDFI) sonograms of each thyroid nodule was reviewed. The data set was analyzed by the partial least squares-discriminant analysis (PLS-DA) and logistic regression (Logistic). Then the two methods were used after selecting statistically significant variables by stepwise regression analysis.

Result

The true positive and negative rates of PLS-DA were 96.95% and 97.73%, respectively, which were significantly higher than the true positive rate (89.86%) and true negative rate (93.12%) of Logistic (P<0.05). After stepwise regression analysis, seven significant variables were selected including the echogenicity of thyroid, shape, margin, internal content, calcification, orientation and vascularity. Based on the selected variables, the true positive and negative rates of PLS-DA were 98.12% and 98.49%, while the true positive and negative rates of Logistic were 95.09% and 95.31%, respectively. Compared to the values before variable selection the true rates of both methods were improved (P<0.05). Moreover, the result of PLS-DA was better than that of Logistic (P<0.05).

Conclusion

PLS-DA and Logistic based on the ultrasonic image are useful in the diagnosis of thyroid nodules. Based on the variables selected by stepwise regression analysis, the diagnosis models were built and the accuracy rate of PLS-DA and Logistic could be improved. Moreover, PLS-DA seems to be more powerful than Logistic.

表1 赋值变量
图1 PLS-DA模型训练集及预测集的散点图
表2 蒙特卡罗模拟运行1000次PLS-DA模型分析结果
图2 不同变量的VIP值
图3 Logistic回归模型训练集及预测集的散点图
表3 蒙特卡罗模拟运行1000次Logistic模型分析结果
表4 筛选变量后PLS-DA模型分析结果
表5 筛选变量后Logistic模型分析结果
1
Oetting A, Yen PM. New insights into thyroid hormone action [J]. Best Pract Res Clin Endocrinol Metab, 2007, 21(2): 193-208.
2
Paschou SA, Vryonidou A, Goulis DG. Thyroid nodules: A guide to assessment, treatment and follow-up [J]. Maturitas, 2017, 96: 1-9.
3
Haugen BR, Alexander EK, Bible KC, et al. 2015 American thyroid association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer: The American thyroid association guidelines task force on thyroid nodules and differentiated thyroid cancer [J]. Thyroid, 2016, 26(1): 1-133.
4
Yi KH. The revised 2016 Korean thyroid association guidelines for thyroid nodules and cancers: Differences from the 2015 American thyroid association guidelines [J]. Endocrinol Metab (Seoul), 2016, 31(3): 373-378.
5
Brown RE, Harave S. Diagnostic imaging of benign and malignant neck masses in children-a pictorial review [J]. Quant Imaging Med Surg, 2016, 6(5): 591-604.
6
Pemayun TG. Current diagnosis and management of thyroid nodules [J]. Acta Med Indones, 2016, 48(3): 247-257.
7
Hee SJ, Hwan BJ, Jin C, et al. Ultrasonography diagnosis and imaging-based management of thyroid nodules: Revised Korean society of thyroid radiology consensus statement and recommendations [J].Korean J Radiol, 2016, 17(3): 370-95.
8
Tessler FN, Middleton WD, Grant EG, et al. ACR Thyroid imaging, reporting and data system (TI-RADS): White paper of the ACR TI-RADS committee [J]. J Am Coll Radiol, 2017, 14(5): 587-595.
9
Djukovic D, Zhang J, Raftery D. Colorectal cancer detection using targeted LC-MS metabolic profiling [J]. Methods Mol Biol, 2018, 1765: 229-240.
10
田刚,周明术,宋敏, 等. 基于胸腔积液肿瘤标志物的PLS-DA和ANN-MPL模型对肺癌的诊断价值分析 [J]. 成都医学院学报,2013, 8(5): 521-524.
11
韩放,吴晶辰,徐江峰, 等. 利用PLS-VIP方法筛选差异表达基因 [J]. 北京大学学报(自然科学版), 2009, 45(1): 1-5.
12
Musumarra G, Barresi V, Condorelli DF, et al. Genome-based identification of diagnostic molecular markers for human lung carcinomas by PLS-DA [J]. Comput Biol Chem, 2005, 29(3):183-95.
13
唐熙,王丰. 乳腺实性肿块超声良恶性鉴别诊断的Logistic回归模型 [J/CD]. 中华医学超声杂志(电子版), 2010, 7(6): 1023-1028.
14
Truett J, Cornfield J, Kannel W. A multivariate analysis of the risk of coronary heart disease in Framingham [J]. J Chronic Dis, 1967,20(7): 511-24.
15
Biondo S, Ramos E, Deiros M, et al. Prognostic factors for mortality in left colonic peritonitis: a new scoring system [J]. J Am Coll Surg, 2000, 191(6): 635-42.
16
张国龙,陈景武. 医学研究中Logistic回归与其他方法的结合应用 [J]. 数理医药学杂志, 2007, 20(6): 763-764.
17
潘群艳,马苏亚,薛尧, 等. Logistic回归模型评价剪切波弹性成像技术鉴别乳腺病灶良恶性的价值 [J/CD]. 中华医学超声杂志(电子版), 2013, 10(8): 669-673.
18
蒋红卫,夏结来. 偏最小二乘回归及其应用 [J]. 医学争鸣, 2003, 24(3): 280-283.
[1] 丁雷, 罗文, 杨晓, 庞丽娜, 张佩蒂, 刘海静, 袁佳妮, 刘瑾. 高帧频超声造影在评价C-TIRADS 4-5类甲状腺结节成像特征中的应用[J]. 中华医学超声杂志(电子版), 2023, 20(09): 887-894.
[2] 张茜, 陈佳慧, 高雪萌, 赵傲雪, 黄瑛. 基于高帧频超声造影的影像组学特征鉴别诊断甲状腺结节良恶性的价值[J]. 中华医学超声杂志(电子版), 2023, 20(09): 895-903.
[3] 李卫民, 陈军民, 黄艳丽, 范晓芳, 韩文, 贾磊, 张俊超, 瞿辰. 基于中国甲状腺超声报告与数据系统分析超声在不同大小甲状腺结节中的诊断价值[J]. 中华医学超声杂志(电子版), 2023, 20(07): 743-748.
[4] 付泽辉, 王思齐, 卢叶君, 张剑, 贺烨, 陈卉. 超声对易误诊的等回声、高回声甲状腺结节良恶性的鉴别[J]. 中华医学超声杂志(电子版), 2023, 20(05): 517-523.
[5] 郭云云, 解翔, 彭梅, 姜凡, 毕玉, 何年安, 胡蕾, 杨杨, 王涛, 石玉洁, 陈冬冬. ACR-TIRADS与C-TIRADS分类分别联合二维剪切波弹性成像对甲状腺结节分类的诊断效能——多中心回顾性研究[J]. 中华医学超声杂志(电子版), 2023, 20(05): 511-516.
[6] 郭东, 姚春, 庞海苏, 单悦. 超声微血流成像联合弹性成像评分对甲状腺TI-RADS 4类结节良恶性的鉴别诊断价值[J]. 中华医学超声杂志(电子版), 2022, 19(10): 1098-1102.
[7] 付泽辉, 卢叶君, 张剑, 莫晓民, 贺烨, 张晓青, 陶楚楚, 陈卉. 常规超声特征联合内部粗大钙化特征鉴别甲状腺结节良恶性的价值[J]. 中华医学超声杂志(电子版), 2022, 19(08): 767-773.
[8] 李盈盈, 李欣洋, 阎琳, 肖静, 张明博, 罗渝昆. S-detect技术辅助住院医师诊断甲状腺影像报告和数据系统4类≤1 cm甲状腺结节的应用价值[J]. 中华医学超声杂志(电子版), 2022, 19(07): 682-687.
[9] 康亚宁, 付超, 司彩凤, 郭艳飞, 李静, 崔可飞. 不同TI-RADS分类系统对甲状腺结节的诊断效能对比研究[J]. 中华医学超声杂志(电子版), 2022, 19(06): 561-566.
[10] 刘如玉, 姜玉新, 赵瑞娜, 王莹, 赖兴建, 朱沈玲, 杨筱, 张晓燕, 高璐滢, 席雪华, 高琼, 张波. 三维超声与二维超声对甲状腺结节恶性风险分级评估的对比研究[J]. 中华医学超声杂志(电子版), 2022, 19(04): 317-324.
[11] 广旸, 何文, 吴佳俊, 赵明昌, 张雨康, 万芳. 基于深度学习的甲状腺结节超声图像分割的临床应用[J]. 中华医学超声杂志(电子版), 2022, 19(03): 206-211.
[12] 李俊, 彭健韵, 邱婉冰, 窦倩怡, 潘福顺, 梁瑾瑜. 甲状腺结节恶性风险分层(指南):ACR TI-RADS与C-TIRADS诊断效能及不同医师使用指南一致性的多中心回顾性比较研究[J]. 中华普通外科学文献(电子版), 2023, 17(06): 401-407.
[13] 柴吉鑫, 张雪, 何时知, 齐艳涛, 王婧婧, 敖亚洲, 陈泳. 不同穿刺方法对甲状腺结节细胞学检查的应用[J]. 中华普外科手术学杂志(电子版), 2023, 17(03): 315-318.
[14] 李华娟, 唐英俊, 王赛妮, 徐旺, 林玲, 李羲, 黄华萍. 肺结节临床与CT影像学特征分析及良恶性预测模型构建[J]. 中华肺部疾病杂志(电子版), 2023, 16(03): 318-323.
[15] 徐婷婷, 赵国政, 刘淑萍, 龚凤玲, 贾利花, 李利. 咽食管憩室的超声诊断学特征分析[J]. 中华诊断学电子杂志, 2022, 10(02): 73-76.
阅读次数
全文


摘要