Figure 1.
Time flow of CT scans at different dose levels.
First, all native scans were performed. Afterwards, contrast enhanced scans were acquired, in a random order of the different dose levels. There was a gap of 30 minutes between the scans. 100 mAs scan ( = base of the simulations) was achieved twice in order to minimize discrepancies between originals and simulations caused by differences in contrast accumulation.
Table 1.
Synopsis of the images created for individual image assessment (original versus simulated).
Figure 2.
Mean values of image noise (left side) and CT values (right side) of characteristic tissues (bone, fat, fluid, lung, and muscle) for original (black bars) and simulated (grey bars) transverse slice images (3 mm slice thickness) at different dose levels (10, 20, 40, 60, and 80 mAs).
There were no significant differences with mean discrepancies of −1.2% in image noise and −0.2% in CT values between simulated and original images. Error bars indicate standard deviation.
Table 2.
Image noise for different anatomical regions in original and simulated images at different dose levels.
Figure 3.
Differences of the mean and confidence interval for noise (a) and CT values (b) for bone, fat, fluid, lung, and muscle at different dose levels.
Figure 4.
Coronal reformations with lung window settings (left side) showed similar image appearance of the lungs in simulated (right column) and original (left column) images.
Simulated (right column) versus original (left column) images of the abdomen in transverse orientation are shown in the middle part. Image noise and streak artifacts are increased in lower dose images without visually detectable differences between original and simulated images. Also the simulated (right column) and original (left column) reformations of the abdomen including the kidneys in coronal orientation matched closely (right side).
Table 3.
Subjective image evaluation.