Can computed tomography(CT) texture analysis help to differentiate benign and malignant nodules of thyroid
GUO Wei
LUO Dehong
ZHAO Yanfeng
LI Lin
LIN Meng
HU Lei
ZHAO Xinming
ZHOU Chunwu
Abstract：Objective This study was aimed to determine the ability of texture analysis of contrast-enhanced computed tomography (CT) in differentiating benign and malignant thyroid nodules. Methods The clinical data of 35 patients with 42 thyroid nodules were retrospectively analyzed. Patients were classified as malignant nodules(n=26) or benign nodules (n=16) group based on their histological examination. All nodules were examined by contrast-enhance neck CT before surgery. The DICOM format enhancement CT images (thickness/gap=5 mm) were imported into CT Kinetics software and the parameters of CT texture were calculated automatically with texture and histogram algorithm to generate unfilter fine texture images. The parameters of CT texture, including entropy, skewness, kurtosis, mean pixels and standard deviation (SD) were derived from contrast-enhanced CT images using unfilter fine texture. The diagnostic value of CT texture parameters were examined with Student’s t-test or Mann-Whitney U test, and receiver operating characteristic (ROC) curve analysis was used to determine optimal CT texture parameters in differentiating malignant and benign lesions. Results For malignant thyroid nodules, the values of entropy, skewness, kurtosis, mean pixels and SD were 6.65±0.92, 0.63±1.37, 0.69±1.23, 84.08±23.36 and 18.14±3.31, respectively. For benign thyroid nodules, the values of entropy, skewness, kurtosis, mean pixels and SD were 5.96 ±0.54, 0.59 ±1.42, 0.51 ±1.17, 72.00 ±24.52 and 20.05 ±6.10, respectively. The entropy values of malignant thyroid nodules were significantly higher than that of the benign thyroid lesions (P<0.05);While the values of skewness, kurtosis, mean pixels and SD did not differ significantly (all P>0.05) between the malignant and benign nodules of thyroid. When entropy value of 6.09 was used as a cut-off for differentiating benign from malignant thyroid lesions, the AUC, sensitivity, and specificity in differentiating benign from malignant thyroid lesions were 0.733, 71.3%, and 70.0%, respectively. Conclusions CT texture analysis is helpful in differentiating the benign from malignant thyroid nodules.
Keywords：ThyroidnodulesCT texture analysisDiagnosisdifferentiation
Publication Date：2017-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：4( 3-5,36 )