AJR: Predictive value of elastography in breast masses may be solidified

An ultrasound image of a suspicious lesion in the left breast in 76-year-old study participant (Zhu et al).

Ultrasound elastography delivered relatively reliable predictions for breast mass malignancies compared with MR diffusion-weighted imaging (DWI), according to a study published in the January issue of the American Journal of Roentgenology.

Elastography and MR DWI are emerging as potential options to enhance the diagnosis of breast tumors, offered lead author Hiroko Satake, MD, of the department of radiology at University School of Medicine in Aichi, Japan, and colleagues. However, neither technique has been widely adopted because of occasional confusion stemming from additional findings.

Satake and colleagues designed the current study to demonstrate the value of combining elastography and MR DWI with existing breast imaging modalities to improve diagnosis as well as to address diagnostic confusion associated with elastography and MR DWI.

The retrospective study focused on 115 BI-RADS category 4 and 5 breast masses evaluated from September 2007 to May 2009. A pair of radiologists reviewed ultrasound elastography and MR DWI images, scoring elastography images according to a five-point scale with scores of one to three classified as benign and four or five as malignant.

For MR studies, “[t]he optimal cutoff level of the continuous ADC (apparent diffusion coefficient) values for differentiating benign versus malignant tumors was determined by estimating the receiver operating characteristic (ROC) curve and identifying the point where the sensitivity and specificity were equal on the curve.”

The researchers reported that 88 of the masses were malignant and 27 benign, with the mean elasticity score for malignant masses (4.1 +/- 0.8) significantly higher than that for benign masses (2.7 +/- 1.1) and the mean ADC for malignant masses significantly lower than that for benign masses.

When the authors undertook univariate analysis, they found that BI-RADS category 5 classification, elasticity scores of four or five and ADC values less than 0.91 x 10^-3mm²/s were statistically significant predictors for malignancy. Multivariate analysis, however, demonstrated that only the BI-RADS category and elasticity score remained statistically significant predictors for malignancy.

For BI-RADS category 4 masses, elasticity provided a statistically significant predictor of malignancy, whereas ADC value did not in the univariate analysis. Furthermore, the multivariate analysis also demonstrated the statistical significance of the elasticity score for BI-RADS category 4 masses, according to Satake and colleagues.

For BI-RADS category 5 masses, neither elasticity nor the ADC value provided a statistically significant predictor of malignancy, reported the researchers, who concluded that: “Our study has shown that BI-RADS category assessment and elasticity score are independently associated with differentiation between benign and malignant lesions of suspicious masses.”

Satake and colleagues suggested that elastography could help differentiate benign from malignant masses for BI-RADS category 4 with the “possibility that ultrasound elastography could be used to prevent unnecessary biopsies.”

The researchers pointed to the need for further study of the diagnostic performance of ADC with separation between BI-RADS 4 and 5 lesions and shared that they are expanding their study to increase the number of cases and include BI-RADS category 3 masses.

Finally, Satake and colleagues stressed the complexity of ultrasound elastography. That is, image acquisition and data quality are highly operator dependent and scoring of elasticity images is not objectively based on quantitative parameters.