Optimization of image quality and radiation dose in CR ammography

Objective: To investigate the difference of image quality between automatic exposure control (AEC) and manual exposure in CR digital mammography, and to observe the effect of image post-processing on improvement of image quality and dose reduction. Methods: CDMAM 3.4 phantom and TORMAM phantom were selected and exposed in a CR system in 5 different phantom thickness. Under each thickness there were 4 dose levels (once automatic and 3 times manual exposures, which corresponded to 80%, 60% and 40% dose). Three observers interpreted all the images. Each image was scored twice with unprocessed images and post-processed image. The differences of image quality were compared between 3 lower dose images and 100% dose image. Results: The results of image quality figure (IQF) of CDMAM phantom showed the image quality of post-processed images were higher than unprocessed images. Part of the lower dose images exceeded AEC dose image in image quality through post-processing. The dose decreased from 20% to 60% through image post-processing in different equivalent breast thickness. From the results of TORMAM phantom, for the detection of filament details and circular details, the scores increased after post-processing (P<0.001). However for the partical details, the effect of image post-processing on image quality was not obvious (P=0.74). Conclusion: In CR digital mammography, the image quality can be improved through image post-processing. Radiation dosage can decrease through the image post-processing without reduce of image quality.