Computerized characterization of breast masses using three-dimensional ultrasound images

Breast ultrasound can potentially increase the accuracy of computerized discrimination of malignant and benign masses. Newly developed three-dimensional (3-D) ultrasound techniques provide statistically richer information than conventional two-dimensional (2-D) ultrasound, and may therefore be better-suited for computerized statistical classification techniques. In this study, we investigated the feasibility of classifying solid breast masses using features extracted from 3-D ultrasound images. Our data set consisted of seventeen biopsy-proven masses. Eight of the masses were malignant and nine were benign. The masses were identified by an experienced breast radiologist in the 3-D volume, and a 3-D ellipsoid containing the mass was defined. Spatial gray level dependence features were extracted from 2-D slices in three regions, which were (i) the interior of the ellipse; (ii) a disk-shaped region at the upper periphery of the ellipse; and (iii) a disk-shaped region at the lower periphery of the ellipse. 2-D analysis was performed by evaluating the classification accuracy of the features extracted from each slice. 3-D analysis was performed by first averaging feature values from different slices into a single 3-D feature, and then evaluating the classification accuracy. The best texture feature in this study achieved a classification accuracy of A(Z)=0.97 for both 3-D and 2-D analysis. Our results indicate that the performance of 3-D analysis is comparable to that of 2-D analysis using the best available slice. Since the best 2-D slice for texture analysis may not be known a-priori, this preliminary study suggests that 3-D ultrasound may be beneficial for computerized breast mass characterization.