To establish a contrast-enhanced ultrasound (CEUS) diagnostic schedule by CEUS analysis of thyroid nodules of C-TIRADS 4. To establish a CEUS-TIRADS diagnostic model to differentiate thyroid nodules (C-TIRADS 4) by combining CEUS with Chinese thyroid imaging reporting and data system (C-TIRADS).
A total of 228 thyroid nodules (C-TIRADS 4) were estimated by CEUS. The arrival time, enhancement degree, enhancement homogeneity, enhancement pattern, enhancement ring, and wash-out time were analyzed in CEUS for all of the nodules. Multivariate factors logistic analysis was performed and a CEUS diagnostic schedule was established. If the nodule had a regular hyper-enhancement ring or got a score of less than 2 in CEUS analysis, CEUS-TIRADS subtracted 1 category. If the nodule got a score of 2 in the CEUS schedule, the CEUS-TIRADS category remained the same as before. If the nodule got a score of more than 2 in the CEUS schedule, CEUS-TIRADS added 1 category. When it reflected an absent enhancement in CEUS, the nodule was judged as CEUS-TIRADS 3. All of the C-TIRADS 4 nodules were re-graded by CEUS-TIRADS. We then compare the diagnosis performance of C-TIRADS, CEUS, and CEUS-TIRADS by sensitivity, specificity, and accuracy.
Among the 228 C-TIRADS 4 nodules, 69 were determined as C-TIRADS 4a, 114 were C-TIRADS 4b, and 45 were C-TIRADS 4c. The sensitivity, specificity, and accuracy of C-TIRADS were 93.1%, 55.3%, and 74.6% respectively. The area under the curve was 0.753. Later arrival time, hypo-enhancement, heterogeneous enhancement, centripetal enhancement, and rapid washout were risk factors of malignancy in multivariate analysis. The sensitivity, specificity, and accuracy of CEUS were 78.7%, 87.5%, and 83.3% respectively. The area under the curve was 0.803. By CEUS-TIRADS diagnostic model combining CEUS with C-TIRADS, a total of 127 cases were determined as malignancy (111 were malignant and 16 were benign) and 101 were diagnosed as benign ones (5 were malignant and 96 were benign). The sensitivity, specificity, and accuracy of CEUS-TIRADS were 95.7%, 85.7%, and 92.1% respectively. The area under the curve was 0.916. The diagnostic performance of CEUS-TIRADS was significantly better than CEUS and C-TIRADS. The difference was statistically significant (P<0.05).
The diagnostic schedule of CEUS could get better diagnostic performance than US in the differentiation of thyroid nodules. The CEUS-TIRADS combining CEUS analysis with C-TIRADS could make up for the deficient sensibility of C-TIRADS, showing a better diagnostic performance than US and CEUS.