It has been widely accepted that increasing the optical density of clinical mammograms would possibly achieve higher contrast. Using 5 X-ray mammography units, series of images with optical densities ranging from 0.7 to 2.1 were obtained to investigate the dependence of radiographic contrast on image optical density. The technical factors investigated were an RMI-156 phantom, a moving grid, Mo/Mo, 0.3 mm focus, 28 kV, MinR 2000/MinR 2000, processing time of 90 s and processing temperature of 32-34.2°C. Pixel values measured on digitized images were converted to optical density using a calibration curve, and radiographic contrast was evaluated for a disk, calcification specks and masses.
The results of 5 series of images showed a consistent relationship between radiographic contrast and optical density, indicating that the radiographic contrast would be effectively improved for test objects with a high object contrast, such as disks or specks, by increasing the optical density of the images. On the other hand, the gain in radiographic contrast was small for objects with a low object contrast. These experimental results were supported by theoretical estimation based on the film characteristic curve. The correlation coefficient between measured and estimated contrasts was 0. 987 as an average of 5 series of images.
It is concluded that increasing the image optical density has only limited efficacy for detecting low-contrast breast masses embedded in glandular tissue.
View full abstract