Manjiri K. Dighe MD, Presenter: Research Grant, General Electric Company

Jeff Thiel, Abstract Co-Author: Nothing to Disclose

Theodore J. Dubinsky MD, Abstract Co-Author: Nothing to Disclose

To assess intra-reader agreement in evaluation of the color elastograms in thyroid elastography using Shear-Wave elastography

After IRB approval, elastograms were obtained by Shear-wave elastography from 77 individuals, with a total of 96 thyroid nodules. Elastography data was acquired without any external compression using the Supersonics Aixplorer machine. Each nodule had multiple color elastograms saved per nodule. 1 reader blinded to the final cytopathology results was asked to score the elastogram images based on a standard 5-point scale. The reader was then asked to rescore the elastograms after a period of 15 days and was also blinded to the prior results. Results were also compared to the cytopathology diagnosis based on Bethesda classification. In addition, since each nodule had multiple images, variability of scoring within a single nodule was also evaluated.

There was overall good intra-reader agreement in scoring the elastograms with a concordance correlation coefficient of 0.83, a weighted kappa of 0.71 with a 95% confidence interval of 0.66 to 0.76 and a mean Intraclass correlation coefficient (ICC) of 0.83 and 0.90. The Area under the curve for accurate diagnosis of the nodule as being benign or malignant was 0.82. The variability in the scoring between the benign nodules was less than that in the malignant nodules with a variance of 0.64 and 1.38 respectively.

Our study indicates that there was good overall intra-reader agreement for qualitative scoring of the elastograms. The overall variance in the scoring of benign nodules was less than that in the malignant nodules. There could be bias in this since the number of malignant nodules in our study was smaller than benign nodules. Since Shearwave elastography provides quantitative values for the stiffness in the nodule, there would be less of a discordance and less variance compared to scoring color elastograms for thyroid nodule elastography.

Decreasing the intra-reader variability in thyroid elastography is important. We also evaluated the variance in scoring each individual nodule since in every exam multiple images are acquired per nodule and it is important to know which particular score should be assigned to a nodule to be able to provide accurate diagnosis.

Dighe, M, Thiel, J, Dubinsky, T, Intra-reader Agreement for Color Based Elastograms in Thyroid Elastography.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14016776.html