Scanning and Image Reconstruction
A 320-multi detector CT (Aquilion One Vision Edition,
Toshiba Medical Systems) was used as the CT device and an Aquarius iNtuition Server (TeraRecon,
Inc.) was used as the workstation.
The imaging conditions that were applied comprised capturing images by adjusting the dose measurement in the tube current product time (mAs) values through three steps (20,
35,
and 60 mAs).
We utilized a helical scanning system with the following attributes: collimation: 80 × 0.5 mm; pitch: 0.813; slice thickness: 5.0 mm; number of scans: 5; reconstruction kernel: FC13; matrix size: 512 × 512; reconstruction field of view (field of view; FOV): 20.0 cm.
With respect to image reconstruction,
we evaluated four types of images,
consisting of FBP images and FBP images processed through the TV method (three stages of intensity; level 1,
3,
and 5).
Phantom
We used a Catphan phantom (The Phantom Laboratory,
Salem,
NY),
with a diameter of 15 mm and a depth of 110 mm.
Modulation transfer function
The MTF that was employed consisted of evaluating the high contrast section of the phantom (CTP404).
Measurements were taken after establishing the 2.5 mm diameter Teflon section at the rectangular region of interest (Width512,
Height40).
Figure 1 displays the range of values obtained from the plot profile sought by the MTF after Fourier transformation.
Noise power spectrum
We evaluated NPS at the uniform portions of the phantom (CTP486).
We obtained the NPS by conducting five slit races using a 1 × 30 pixel virtual slit with the phantom center in a 256 × 256 range (Fig.
2).
Contrast-to-Noise Ratio
We evaluated the CNR at the low-contrast portions of the phantom (CTP515).
We set the region of interest at 14 mmφ and the signal region at 15 mmφ at the 1.0% contrast module portion.
We set the background region as the region separated from the signal by 18 mm (Fig.
3).
The CT values obtained were calculated from the CNR derived from standard deviation (SD).
Observer study
In order to evaluate low-contrast detectability,
we performed a visual assessment using a medical image interpretation monitor (Fig.
3).
We utilized a phantom contrast differential of 1.0% in the visual assessment.
We presented five images taken for each imaging condition set that was applied,
determined the detectable sizes for FBP images and images processed using the TV method,
and calculated the average limit on detection size for a group of five observers (15 mm; 1 pt,
9 mm; 2 pt,
8 mm; 3 pt,
7 mm; 4 pt,
6 mm; 5 pt,
5 mm; 6 pt,
4 mm; 7 pt,
3 mm; 8 pt,
2 mm; 9 pt).