Image quality of digital direct flat-panel mammography versus an indirect small-field CCD technique using a high-contrast phantom

Abstract

Objective: To compare the detection of microcalcifications on mammograms of an anthropomorphic breast phantom acquired by a direct digital flat-panel detector mammography system (FPM) versus a stereotactic breast biopsy system utilizing CCD (charge-coupled device) technology with either a 1024 or 512 acquisition matrix (1024 CCD and 512 CCD).

Materials and methods: Randomly distributed silica beads (diameter 100-1400 μm) and anthropomorphic scatter bodies were applied to 48 transparent films. The test specimens were radiographed on a direct digital FPM and by the indirect 1024 CCD and 512 CCD techniques. Four radiologists rated the monitor-displayed images independently of each other in random order.

Results: The rate of correct positive readings for the "number of detectable microcalcifications" for silica beads of 100-199 μm in diameter was 54.2%, 50.0% and 45.8% by FPM, 1024 CCD and 512 CCD, respectively. The inter-rater variability was most pronounced for silica beads of 100-199 μm in diameter. The greatest agreement with the gold standard was observed for beads >400 μm in diameter across all methods.

Conclusion: Stereotactic spot images taken by 1024 matrix CCD technique are diagnostically equivalent to direct digital flat-panel mammograms for visualizing simulated microcalcifications >400 μm in diameter.

Figure 1

Radiological image of a laser-printer-film covered with different silica beads containing a polymethylmethacrylate (PMMA) sheet of 1.5-cm thickness (Plexiglas, Degussa) and a 1.5-com thick layer of ground meat as scattering bodies. For the raters' orientation, a metal wire was used to divide the phantom into 4 quadrants. Direct flat-panel detector mammography scanned the whole phantom (a). The indirect CCD technique only produced spot images of the phantom's 4 quadrants (c and d). Quadrant IV contains 49 lobular microcalcifications of 300–599 μm in diameter.

Figure 2

Mean absolute interrater variability of the 4 radiologists' ratings from an experimentally preset reference values for each quadrant of the 48 universal laser printer films are presented here in box plots for the variables “number” and “size” separately for each rater. Statistical outliers are labeled with a dot (•). In the global comparison of all simulated microcalcifications, the diagnostic accuracy of direct digital flat-panel detector mammography system versus the small-field CCD stereotactic breast biopsy system utilizing was comparable, and independently of number, size, or shape of the silica beads.

Figure 3

(a) Box plot presenting the interrater variability of the 4 radiologists' from the experimentally preset gold standard as a function of size and number of simulated microcalcifications scattered on the 48 films. Statistical outliers are labeled with an asterisk (∗) or a dot (•). Criterion: “Number" of detectable microcalcifications as a function of the experimentally preset size of silica beads. (b) Box plot presenting the interrater variability of the 4 radiologists' from an experimentally preset gold standard as a function of size and number of simulated microcalcifications scattered on the 48 films. Statistical outliers are labeled with an asterisk (∗) or dots (•). Criterion: “Size” of detectable microcalcifications as a function of the experimentally preset size of the silica beads.