Eung Tae Kim MD, Presenter: Nothing to Disclose

Jeong Seon Park MD, Abstract Co-Author: Nothing to Disclose

Kwang Gi Kim PhD, Abstract Co-Author: Nothing to Disclose

Soo-Yeon Kim MD, Abstract Co-Author: Nothing to Disclose

To evaluate the computer-aided diagnosis (CAD) of US elastography for classification of benign and malignant thyroid nodules

Between October 2008 and February 2010, real-time thyroid US elastography was performed in 488 patients who had scheduled for US-guided fine needle aspiration. We included consecutive patients who had thyroid nodules with surgically proven histopathology (n=77), or had at least twice results of benign thyroid cytology (n=146). Finally, we included 223 patients (M: F=27:196; mean age, 50.8 years) with 230 thyroid nodules (89 malignant, 141 benign). Two radiologists retrospectively reviewed elastogram in consensus and assigned elasticity score from 1 to 5. For each representative elasticity image, ROI was drawn around the mass margin by another radiologist and a score for each pixel was assigned from 0 for the greatest strain to 255 for no strain. Elasticity features—mean hue histogram value (MHHV), skewness, kurtosis, difference histogram variation, edge density, and run length—were computed to evaluate the findings of benign and malignant nodules. Best classifiers were evaluated by using forward logistic regression analysis. Receiver operating characteristic (ROC) analysis was used to evaluate diagnostic performances and the area under the curve (AUC) was compared between CAD and elasticity scores.

The AUC values of the CAD and elasticity scores were 0.95 and 0.91, respectively (p = 0.10). The sensitivity, specificity, positive and negative predictive values were 93%, 83%, 78%, and 95% for computer-assisted classifiers, and 92%, 81%, 75%, and 94% for elasticity scores at the cut off value between 2 and 3.

CAD of US elasticity images has at least equal diagnostic performance compared to radiologist’s elasticity-scoring in the classification of benign and malignant thyroid nodules.

Computer-aided analysis of US elasticity images has the potential to aid in the classification of benign and malignant thyroid nodules.

Kim, E, Park, J, Kim, K, Kim, S, Computer-aided Diagnosis of Ultrasound Elastography for Classification of Benign and Malignant Thyroid Nodules.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9010649.html