Study On Differentiation Of Benign And Malignant Thyroid Nodules And Prediction Of Cervical Lymph Node Metastasis By Quantitative Parameter Energy Spectrum CT Combined With Texture Analysis

The Differential Diagnostic Performance of Spectral CT Quantitative Parameters and Texture Analysis on Benign and Malignant Thyroid Nodules Objective:To explore the value of spectral CT quantitative parameters and texture analysis in the differential diagnosis of benign and malignant thyroid nodules.Methods: The spectral CT scan data of 77 patients with thyroid nodules confirmed by pathology were analyzed retrospectively.And these nodules were classified into benign nodules and malignant nodules based on the pathological results of surgery.The spectral parameters were measured and calculated in venous phase enhanced images,including the iodine concentration（IC）,normalized iodine concentration（NIC）and slope of energy spectral curve（λHU）.The area of interest of nodule was delineated layer by layer on the enhanced scanning venous iodine-based material image,and the whole nodule was extracted for 3D texture feature analysis,and the optimal texture parameters were screened by single factor analysis and multi-factor analysis.Receiver operating characteristic curves（ROC）were plotted for energy spectrum quantitative parameters,texture parameters and their combination to compare diagnostic efficacy.Results: 93 thyroid nodules were included,52 in the benign group and 41 in the malignant group.There were significant differences in IC,NIC and λHU between benign nodule groups and malignant nodule groups（P<0.05）.IC has the best diagnostic performance,and the AUC was 0.784,the sensitivity was 75.61%,and specificity was75.00%.Among the texture features,LLL_firstorder_Skewness has the best diagnostic efficiency,with an AUC value of 0.773,sensitivity and specificity was 90.24% and57.69%,respectively.Combined with the best spectral parameter（IC）and texture parameter（LLL_firstorder_Skewness）,the AUC,sensitivity and specificity can be improved to 0.836,90.20% and 67.30%,respectively.When the energy spectrum parameters（IC,NIC and λHU）and texture parameters（LLL_firstorder_Skewness,glrlm_Run Length Non Uniformity Normalized）joint diagnosis,the AUC could be raised to 0.863,the sensitivity and specificity could be increased to 90.20% and 78.80%,respectively.Conclusion: Spectral CT quantitative parameters and texture feature analysis based on the venous phase have certain value in the differential diagnosis of benign and malignant thyroid nodules,and their combination can improve the diagnostic efficiency of benign and malignant thyroid nodules.Value of Multi-parameter Energy Spectrum CT Combined with Tumor Morphological Features in Predicting Cervical Lymph Node Metastasis of Thyroid Papillary Carcinoma Objective:To investigate the value of energy spectrum CT quantitative parameters and morphological characteristics in preoperative prediction of cervical lymph node metastasis of thyroid papillary carcinoma（PTC）Methods: Retrospective analysis was performed on the preoperative case data and energy spectrum CT images of 88 patients with PTC confirmed by surgical pathology.All patients underwent lymph node dissection in the central region of the neck.Based on surgical pathology,they were classified as lymph node metastasis group（49 cases）and non-lymph node metastasis group（39 cases）.The morphological features of primary lesions included the number,capsule invasion,nodular length,nodular location,and calcification were assessed.CT quantitative parameters of energy spectrum include primary venous iodine concentration（IC）,standardized iodine concentration（NIC）and energy spectrum curve slope（λHU）.The （?）~2 test or t test were used to compare the difference of qualitative characteristics and quantitative parameters of PTC primary lesions between the two groups.The binary logistic regression analysis was used to build models based on morphological features,quantitative CT parameters and their combination.The receiver operating characteristic curve（ROC）was described,and then the area under ROC curve（AUC）,sensitivity and specificity were measured to evaluate the diagnostic efficacy.Results: There were significant differences in the number of lesions,capsule infiltration,nodular length and age between metastatic group and non-metastatic group（P < 0.05）.The IC,NIC and λHU of the primary lesion in the group with lymph node metastasis were than larger in the group without lymph node metastasis,there was a statistical difference（P < 0.05）.The AUC value of morphological features for predicting cervical lymph node metastasis of PTC was 0.836,and the sensitivity and specificity were 81.83%and 79.49%,respectively.The AUC value of multi-parameter energy spectrum CT in the predictions of PTC lymph node metastasis was 0.718,the sensitivity was 63.27%,and the specificity was 76.92%.The AUC value of PTC combined with morphological characteristics and quantitative parameters of primary spectral CT for predicting cervical lymph node metastasis was 0.884,the sensitivity was 91.84%,and the specificity was 76.92%.The diagnostic efficiency of the combination of them was superior to the morphological features and quantitative parameters of energy spectrum CT.Conclusion: The quantitative parameters of energy spectrum CT and morphological characteristics of thyroid papillary carcinoma have certain clinical value in predicting cervical lymph node metastasis,and the combined diagnostic efficacy of the two methods is further improved.It is beneficial for clinicians to design a more reasonable and personalized lymph node treatment plan before surgery and avoid unnecessary surgical trauma.