Page 126 - 南京医科大学自然版
P. 126
第46卷第6期
·908 · 南 京 医 科 大 学 学 报 2026年6月
A B C D
E F G H
A-D:A 29⁃year⁃old female patient with papillary thyroid carcinoma in the right lobe,pathologically confirmed with CLNM. E-H:A 25⁃year⁃old fe⁃
male patient with papillary thyroid carcinoma in the left lobe,pathologically confirmed without CLNM. Figures A and E are 70 keV images from the arte⁃
rial phase;figures B and F are Zeff images from the arterial phase;figures C and G are iodine maps from the arterial phase;figures D and H are 70 keV
images from the non⁃contrast phase. Quantitative measurements of the primary lesions in both patients revealed lower values in the CLNM group com⁃
pared to the non⁃CLNM group:arterial phase 70 keV CT value(93.4 HU vs. 116.3 HU),arterial phase Zeff(8.07 vs. 8.78),arterial phase IC(1.31 mg/mL vs.
2.93 mg/mL),and non⁃contrast phase 70 keV CT value(53.6 HU vs. 92.1 HU).
图1 CLNM和非CLNM组甲状腺乳头状癌光谱CT图像
Figure 1 Spectral CT images of papillary thyroid carcinoma in the CLNM group and non⁃CLNM group
表4 多因素逻辑回归及预测模型
Table 4 Multivariate logistic regression and predictive models
Model Coefficient SD OR(95%CI) P Tolerance VIF
Clinical model
Cystic 002.621 1.107 13.750(1.569-120.502) 0.018 1.000 1.000
Quantitative parameter model
000.034 0.013 01.035(1.009-1.062) 0.009 0.262 3.822
NC 40 keVHU
0-0.158 0.044 00.854(0.783-0.930) 0.000 0.269 3.715
NC 70 keVHU
NC Zeff -15.273 6.854 00.000(0.000-0.159) 0.026 0.920 1.087
Combined model
000.039 0.016 01.040(1.007,1.073) 0.015 0.255 3.926
NC 40 keVHU
0-0.203 0.055 00.816(0.733-0.909) 0.000 0.267 3.741
NC 70 keVHU
-21.632 7.634 00.000(0.000,0.001) 0.005 0.915 1.093
NC Zeff
Gender 001.845 0.847 06.326(1.202-33.284) 0.029 0.918 1.090
Cystic 003.676 1.677 39.479(1.477-1055.527) 0.028 0.891 1.122
表5 不同模型对cN0 PTC患者CLNM的预测效能
Table 5 Predictive performance of different models for CLNM in cN0 PTC patients
Model AUC Sensitivity(%) Specificity(%) PPV(%) NPV(%)
Clinical model 0.591 20.0 98.2 50.3 92.5
Quantitative Parameter model 0.835 73.3 85.7 83.7 76.2
Combined model 0.896 86.7 80.4 81.6 85.8
分析常规特征以及 CT 定量参数,而应将两者结合 研究,可能存在偏倚,且样本量不大;②采用手动勾
分析。 画 ROI 的方法测定量参数。综上,本研究结果显示
本研究主要存在以下不足:①单中心、回顾性 光 谱 CT 定 量 参 数 能 有 效 提 高 对 cN0 PTC 患 者

