Matching breast masses depicted on different views a comparison of three methods

Abstract

Rationale and objectives: Computerized determination of optimal search areas on mammograms for matching breast mass regions depicted on two ipsilateral views remains a challenge for developing multiview-based computer-aided detection (CAD) schemes. The purpose of this study was to compare three methods aimed at matching CAD-cued mass regions depicted on two views and the associated impact on CAD performance.

Materials and methods: The three search methods used (1) an annular (fan-shaped) band, (2) a straight strip perpendicular to the estimated centerline, and (3) a mixed search area bound on the chest wall side by a straight line and an annular arc on the nipple side, respectively. An image database of 200 examinations with positive results depicting the masses on two views and 200 examinations with negative results was used for testing. Two performance assessment experiments were conducted. The first investigated the maximum matching sensitivity as a function of the search area size, and the second assessed the change in CAD performance using these three search methods.

Results: To include all 200 paired mass regions within the search areas, maximum widths were 28 and 68 mm for the use of the straight strip and the annular band search methods, respectively. When applying a single-image-based CAD scheme to this image database, 172 masses (86% sensitivity) and 523 false-positive (FP) regions (0.33 per image) were detected and cued. Among the positive findings, 92 were cued by the CAD system on both views, and 80 were cued on only one view. In an attempt to match as many of the 172 CAD-cued masses (true-positive [TP] regions) on two views by incrementally reducing the CAD threshold inside the different search areas, the CAD scheme generated 158 TP-TP paired matches with 14 TP-FP paired matches, 142 TP-TP paired matches with 30 TP-FP paired matches, and 146 TP-TP paired matches with 26 TP-FP paired matches, using the methods involving the straight strip, the annular band, and the mixed search areas, respectively. Using the straight strip search method, the CAD also eliminated 25% of FP regions initially cued by the single-image-based CAD scheme and generated the lowest case-based FP detection rate, namely, 15% less than that generated by the annular band method.

Conclusions: This study showed that among these three search methods, the straight strip method required a smaller search area and achieved the highest level of CAD performance.

Figure 1

Distribution of mass region size (mm ²) as computed by the CAD scheme on the CC and corresponding MLO views for the entire dataset (a) and subset of masses with smaller sizes (b).

Figure 2

Distributions of mass region contrast values (a) and shape factor ratio (b) as computed by the CAD scheme for all regions depicted on the CC and corresponding MLO views.

Figure 3

Demonstration of the three search methods for matching a mass depicted on two views. A mass region with two associated distances (Euclidian distance $\left[d_1^E\right]$ and projected distance along the centerline that is perpendicular to the chest wall) between the nipple and the center of the mass as depicted on a CC view are shown in (a). The three types of search areas are defined on the MLO view are represented as an annular band (b), a straight line based strip (c), and a mixed area with a staight line based boundary in the posterior and an arc in the front (d).

Figure 4

A Comparison of two methods designed to modify the estimated centerline on CC views. Shown are the initial (horizontal) centerline (a), the adjusted centerline connecting the nipple and the center pixel along the edge of imaged breast area (b), the adjusted centerline generated by the iterative rotation method (c), and the two resulting (adjusted) centerlines generated by the two methods (d).

Figure 5

Comparison of two search straight strips including that one is defined based on a horizontal centerline assuming that the chest wall is parallel to the image edge on CC view (a) and one is defined based on a modified centerline symmetrically dividing the breast image on the CC view (b).

Figure 6

Histograms of the number of incrementally matched masses as a function of the search area width (mm) for straight line based and the annular band based search methods.

Figure 7

Two performance curves for the sensitivity levels as a function of search area width (mm) using the straight line based strip and the annular band search methods.

Figure 8

The case and region based FROC performance curves as well as the operating threshold line of original single-image based CAD scheme when applied to the test image dataset of 200 positive and 200 negative examinations with 1600 images.