| Objectives:To explore the value of using the imaging morphology and dual-energy CT multi parameters of papillary thyroid carcinoma(PTC)to predict central group lymph node metastasis(CLNM).Methods:The data of 340 patients were retrospectively analyzed,340 patients with single PTC confirmed by postoperative pathology and received central lymph node dissection were selected,including 178 cases with central lymph node metastasis and 162 cases without metastasis.All of the patients underwent dual-energy CT enhanced scan before operation.The relationship of imaging morphology of PTC foci with CLNM was studied by chi-square test and Spearman correlation analysis,including the maximum diameter of PTC foci,the contact range with thyroid capsule,the marginal morphology,calcification of PTC foci,location of PTC foci and the completeness of enhanced ring of PTC foci.The CT value,the iodine concentration(IC),Overlay value of the PTC foci in the non-contrast,arterial and venous phases of cervical enhanced CT scan were measured,normalized CT value(NCT),normalized iodine concentration(NIC)and the spectrum curve slope(SCP)of the PTC foci in arterial and venous phases were calculated.Independent sample t-test was used to compare the differences of data between CLNM group and non-metastasis group,the receiver-operating characteristic(ROC)curve was drew to select critical value with best sensitivity and s specificity,Binary Logistic regression analysis combined with ROC curve was used to get sensitivity and specificity in predicting CLNM with multiple CT values or dual-energy CT multiple parameters.Results:1.Imaging morphology characteristics of PTC foci:The distribution differences of the maximum diameter of PTC foci,the contact range with thyroid gland capsule,the marginal morphology,and the completeness of enhanced ring of PTC foci were statistically significant between the CLNM group and non-metastasis group(P=0.002,0.005),while the distribution differences of calcification and the location of PTC foci between the CLNM group and non-metastasis group were not statistically significant(P=0.942,0.283).Correlation analysis showed that CLNM was positively correlated with the maximum diameter of the lesions and the contact range with capsule of thyroid gland(r=0.170,P=0.002;r=0.146,P=0.007),CLNM was negatively correlated with marginal morphology and the completeness enhanced ring of the foci(r=-0.140,P=0.010;r=-0.134,P=0.013),There was no correlation between CLNM with calcification and the location of foci(P=0.942,P=0.083).The results of ROC curve showed that the diagnosis efficiency of CLNM was low with the maximum diameter of the primary lesion,contact range with thyroid gland capsule,the marginal morphology and the completeness of the enhanced ring of lesions.The AUC were 0.585,0.584,0.596 and 0.565.2.Various CT values quantitative parameters of PTC foci:311(91.5%)PTC foci showed obvious enhancement in the arterial phase and decreased in the venous phase compared with arterial phase.The arterial phase CT value of PTC foci in CLNM group were higher than non-metastasis group,but the difference was not statistically significant(P>0.05).In CLNM group,the CT value of PTC foci was(112.62±14.11)HU,net increased CT value was(58.95±12.09)HU,and the NCT was 0.67±0.25.In the non-metastasis group,the CT value of PTC lesions was(103.59±9.68)HU,net increased CT value was(50.59±13.81)HU,and the NCT was 0.52±0.23.The difference of CT value,net increased CT value and the NCT of PTC foci in venous phases was statistically significant between the CLNM group and non-metastasis group,the thresholds were 107.5 HU,53 HU,0.76,the sensitivity and specificity were 71.4%?81.8%;76.2%?59.1%;71.4%?72.7%.Combined multiple CT values in venous phase to predict CLNM,ROC curve was draw,AUC was 0.786,the sensitivity and specificity were 85.7%?73.6%,the sensitivity was higher than before.3.Dual-energy quantitative parameters of PTC foci:There were no statistically significant differences in IC,NIC,Overlay value and SCP in arterial phase of PTC foci between CLNM group and non-metastasis group(P>0.05).There were statistically significant differences in IC,NIC,Overlay value and SCP in venous phase of PTC foci between the two groups(t=3.306,P<0.001;t=2.309,P<0.05;t=2.123,P<0.05;t=2.128,P<0.05),the IC of PTC foci of CLNM group and non-metastasis group in venous phase were(2.83±0.81)mg/ml and(2.28±0.67)mg/ml;the NIC were 0.72±0.15 and 0.63±0.12;the Overlay value were(57.81±18.81)HU and(52.75±18.26HU);the SCP were 3.54±0.95 and 2.99±0.68,the threshold of IC,NIC,Overlay value and SCP for predicting CLNM were 2.54 mg/ml,0.71,56.5HU and 3.31,the sensitivity and specificity were72.7%,81.4%;76.1%,85.9%;84.7%,62.8%;65.9%,88.6%.Binary Logistic regression was used to obtain the combined coefficient of predicting CLNM by multiple dual energy quantitative parameters of the PTC foci in venous phase,and the ROC curve was drawn.The AUC was 0.798,the sensitivity and specificity were 81.8%and 87.0%.4.Imaging morphology characteristics of PTC foci combined with CT value parameters and dual-energy parameters:The maximum diameter,the contact range with thyroid capsule,the irregular marginal morphology and the incompleteness of enhanced ring of PTC foci combined with three kinds of CT values and four kinds of dual-energy parameters in venous phase to predict CLNM,the AUC of ROC curve was 0.911,the sensitivity and specificity were 90.6%and 79.8%.Conclusion:1.For imaging morphology characteristics of PTC foci,The maximum diameter of PTC foci,the contact range with thyroid gland capsule,the marginal morphology and the completeness of enhanced ring of PTC foci were correlated with CLNM,but the correlation degree was weak and the prediction efficiency was low,so it is not recommended to be used as the prediction indicator for CLNM alone.2.Three kinds of CT value parameters and multiple dual energy quantitative parameters can be used as prediction indicators for CLNM.The prediction specificity of NIC and SCP is high.The efficiency of combined dual-energy quantitative parameters in predicting CLNM was more balanceable,sensitivity and specificity were both>80%.The combination of imaging morphology characteristics and multiple quantitative parameters can increase the sensitivity of prediction. |
PDF Full Text Request