The Value Of Conventional Ultrasonography And Ultrasound Elastography In The Differential Diagnosis Of Benign And Malignant Thyroid Nodules

Objective:Ultrasound elastography(UE) is a novel and promising imaging tool, which has been applied to breast, thyroid, and prostate tissues. The aim of this study was to compare the value of UE including elasticity score(ES) and strain ratio(SR) on differential diagnosis of benign and malignant thyroid nodules with that of conventional ultrasonography(CU).Methods: This was a retrospective study conducted in a single center. Between June 1, 2013 and November 3, 2014, 416 cases of thyroid nodule operated and histopathologically confirmed in Cangzhou central hospital. The CU and UE data were analyzed, and compared with the histopathological results. Finally, 108 nodules from 98 patients consistent with the inclusion criteria were included in the current study.The 108 nodules were subdivided into 3 groups according to the size,with the diameter of less than 1 centimeter as group 1, 1 to 2 centimeters as group 2, and over 2 centimeters as group 3 respectively. According to Park’s thyroid imaging reports and data system, CU features of grade 1~3 were classified as benign, and those of grade 4~5 as malignant. ES uses a five score scale, with 0~2 score as benign, and 3~4 score as malignant. The receiver operating characteristic curve of the efficacy for diagnosing malignant thyroid nodules by SR, ES, and CU was constructed respectively. The threshold of SR for diagnosis of malignant thyroid nodule was determined according to the highest point of Youden index(sensitivity + specificity-1), SR ≥ the best threshold diagnosis of malignant, SR < the best threshold diagnosis of benign.Combined application of CU, ES, and SR was performed, and with those nodules classified as malignant if they were diagnosed as malignant by any of the above three methods, and otherwise classified as benign. SPSS17.0 software was used for statistical analysis. The chi square test was used for comparison of the difference of CU features and ES between benign and malignant thyroid nodules, and U-test was used for comparison of the difference of SR between benign and malignant thyroid nodules. The sensitivity, specificity and accuracy for the differential diagnosis between benign and malignant thyroid nodules of different size group and overall sample using CU, ES, SR and combined application of the above three methods were compared respectively according to the gold standard of histopathological diagnosis. The chi square test of row times column was used for comparison of the diagnostic efficacy for malignant thyroid nodules among the four methods, and Bonferroni method was used for pairwise rates comparison of multiple samples. The area under curve(AUC) was calculated. Z test was used for the comparison of the AUC between SR and CU, between SR and ES, and between ES and CU respectively. P<0.05 was defined as statistical significant.Results: Among the 108 thyroid nodules, 50 were histopathologically proved as malignant and 58 as benign. 53 nodules were diagnosed as malignant and 55 as benign according to the CU diagnosis. The difference of the CU features between benign and malignant nodules was statistically significant(x2=31.171, P=0.000). 52 nodules were diagnosed as malignant and 56 as benign according to the ES diagnosis. The difference of the ES between benign and malignant nodules was statistically significant(x2=54.271, P=0.000). The SR values of malignant nodules were significantly higher than those of benign nodules(3.47±0.989 vs 1.79±0.875), and the difference was statistically significant(U=5.5,P<.001). The best threshold of SR for the diagnosis of malignant nodule was 2.15. The ROC AUC of SR(0.911) was greater than that of ES(0.888) and CU(0.815) for diagnosis of malignant thyroid nodule. The difference of the ROC AUC was statistically significant(Z =2.4, P<0.05) between SR and CU, and were not statistically significant between SR and ES(Z=0.532, P>0.532) and between ES and CU(Z =1.403, P>1.403). When used alone, the sensitivity(96.0%) and specificity(82.7%)in diagnosing malignant thyroid nodules were higher for SR than those for ES and CU. Combined application of the three methods achieved higher sensitivity(98.0%), specificity(89.6%) and accuracy(93.5%) in diagnosing malignant thyroid nodules than those of the separate application with any of the three methods. For the overall sample, the difference of diagnostic efficacy for differentiate benign and malignant thyroid nodules was statistically significant among the three methods and their combined application( x2=14.132, P<0.05). The difference of diagnostic efficacy for differentiate benign and malignant thyroid nodules was statistically significant between CU and SR,and between CU and combined application of the three methods, and were not statistically significant between CU and ES, between ES and SR, between ES and combined application of the three methods, and between SR and combined application of the three methods. In comparison of the different size groups, the diagnostic efficacy for differentiate benign and malignant thyroid nodules was higher for the combined application of the three methods than the SR, and then ES, with CU the lowest. For Group 1, the difference of diagnostic efficacy for differentiate benign and malignant thyroid nodules was statistically significant among the three methods and their combined application,and the difference of diagnostic efficacy for differentiate benign and malignant thyroid nodules was statistically significant between CU and SR,and between CU and combined application of the three methods, and were not statistically significant between CU and ES, between ES and SR, between ES and combined application of the three methods, and between SR and combined application of the three methods. For Group 2 and 3, the difference of diagnostic efficacy for differentiate benign and malignant thyroid nodules were not statistically significant among the three methods and their combined application.Conclusions:1 All of the CU, ES, and SR had a high accuracy for the differential diagnosis of benign and malignant thyroid nodules, with SR had the highest diagnosis efficiency, ES the next, and CU the lowest.2 This study showed that the best threshold of SR in diagnosing malignant thyroid nodule is 2.15, with a very high diagnostic efficacy, and the AUC of the ROC was 0.911.3 Although the diagnostic efficiency of ultrasound elastography(ES, SR) is better than that of CU for the differential diagnosis of benign and malignant thyroid nodules, it will still be affected by nodule size, depth of the position, the subjective judgment of operator and the factors of instrument, etc.4 Combined application of the three methods achieved higher diagnostic efficiency in diagnosing malignant thyroid nodules than those of the separate application with any of the three methods of CU, ES, and SR, especially for small nodules. So it will be helpful for both CU and UE to consult with each other in differential diagnosing of benign and malignant thyroid nodules.