Construction and analysis of digital mammography database: Analysis of CBT among multiple facilities

We analyzed the compression pressures in 2772 mammography images of 807 patients acquired by digital mammography equipment at four facilities. The analysis included the average compression pressure at all facilities, difference in compression pressure at each facility, differences between the pressures used by radiological technologists in the same facility, and difference attributed to the breast structure. We also analyzed the effects of the compression pressure on the breast thickness and mean glandular dose (MGD) at each facility. The median values of the compression pressure and breast thickness for the 2772 images at all facilities were 86.5 N and 43 mm, respectively. The compression pressures differed among the institutions. The maximum difference in the median pressures among the four facilities was 38.6 N, while the difference in the breast thickness was 6 mm. The radiological technologists working at the same facility used almost the same compression pressure. However, differences between the compression pressures used by different radiological technologists were observed. The compression pressure in a dense breast was smaller than that in a non-dense breast. The difference in the compression pressure affected the breast thickness and MGD. The results of this analysis could be utilized for an optimal imaging in future digital mammography.

[Construction and Analysis of a Digital Mammography Database: Analysis of Breast Compression Pressures at Four Facilities].

This study aims to evaluate the indices of glandular dose heterogeneity in full-field digital mammography. The distributions of granular dose (GD) in a breast phantom with a skin layer of 4 mm were determined using the Monte Carlo method with simulated X-ray fluence spectra. First, the GD to air kerma (GD/Kair) volume histogram was obtained from the GD distributions, which were indicated by the glandular volume (%) as a function of GD/Kair. The GD indices, namely, the maximum glandular dose (GD2%) and glandular volume percentage (%) receiving at least the mean glandular dose (MGD) (VMGD) were calculated from the GD/Kair volume histogram. Next, the scatter plots of GD2%/MGD and VMGD were drawn as functions of the normalized mean glandular dose (DgN). Finally, (GD2%)i and (VMGD)i were obtained from the relationship between the GD indices and DgN for 596 clinical irradiation cases based on individual irradiation conditions. The values of GD2%/MGD were more affected by breast thickness than glandularity and tube voltage, and they decreased according to the power law of DgN for all the target/filter combinations. The values of VMGD were proportional to DgN and decreased with increase in the compressed breast thickness (CBT). The values of (MGD)i and (GD2%)i for 596 clinical irradiation cases were estimated to range from 0.6 to 3.0 mGy and 1.1 to 7.0 mGy, respectively, and (VMGD)i was in the range 32%-48%. (GD2%)i and (VMGD)i are mainly affected by breast thickness. These indices are useful for the evaluation of glandular dose heterogeneity in mammography.

Yoshiharu Higashida

Papers

Construction and analysis of digital mammography database: Analysis of CBT among multiple facilities

[Construction and Analysis of a Digital Mammography Database: Analysis of Breast Compression Pressures at Four Facilities].

Indices for the evaluation of glandular dose heterogeneity in full-field digital mammography.