RSNA 2009 

Abstract Archives of the RSNA, 2009


SSK02-04

Comparison of Two Commercially Available Sonoelastography Systems in Distinguishing Benign from Malignant Breast Masses

Scientific Papers

Presented on December 2, 2009
Presented as part of SSK02: Breast Imaging (Ultrasound)

Participants

Nariya Cho MD, Presenter: Nothing to Disclose
Woo Kyung Moon MD, Abstract Co-Author: Nothing to Disclose

PURPOSE

To retrospectively compare the diagnostic performance of two commercially available sonoelastography systems in distinguishing benign from malignant breast masses.

METHOD AND MATERIALS

Between August 2008 and November 2008, 64 women (mean age 50 years, range 28 – 78 years) with 74 breast masses (mean size 16 mm, range 3-50 mm) (23 malignant, 51 benign) underwent sonoelastographic examinations with both systems (A and B) by one radiologist prior to biopsy. Probability of malignancy based on conventional US findings was recorded prior to sonoelastography. Real time imaging files obtained from the two systems were saved as video clips, then masked, randomized and separately analyzed by two radiologists without information of final histology. For the A images, the lesion width ratio obtained by dividing the strain image measurements by the conventional US image measurements was calculated. For the B images, the elasticity score (1-5) based on the degree and distribution of strain was given in consensus. Diagnostic performance of the two systems in distinguishing benign from malignant masses was compared using receiver operating characteristic (ROC) curve analysis and McNemar’s test.

RESULTS

The area under the ROC curve for the B system (Az=0.896) was higher than that of the A system (Az=0.722) (difference between areas 0.175, 95% CI 0.0589-0.291, P=0.003) and similar to that of conventional US (Az=0.916) (difference between areas 0.0196, 95% CI -0.0747-0.114, P=0.684). The best cut-off values, yielding the maximal sum of sensitivity and specificity, were between width ratios of 0.99 and 1 and elasticity scores of 3 and 4. The sensitivity of the A system was higher than that of the B system [87.0% (20 of 23) vs. 78.3% (18 of 23), P = 0.006] and the specificity of the B system was higher than that of the A system [90.2% (46 of 51) vs. 64.7% (33 of 51), P = 0.004].

CONCLUSION

The B system showed better overall diagnostic performance and specificity, and the A system showed better sensitivity in distinguishing benign from malignant breast masses.

CLINICAL RELEVANCE/APPLICATION

Understanding the characteristics of the two commercially available sonoelastography systems can be helpful in optimizing the diagnostic criteria for each system.

Cite This Abstract

Cho, N, Moon, W, Comparison of Two Commercially Available Sonoelastography Systems in Distinguishing Benign from Malignant Breast Masses.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8008435.html