Phantom Design
(Fig.
1 and Table 1)
The liver phantoms consisted of pig liver homogenized with five different fat contents and with three different iron contents.
For each,
saccharated iron oxide (SIO,
20mgFe/ml) was diluted with distilled water to obtain concentrations of 0,
40 and 80 mg of iron per 100 mg of solution.
Fat volume fraction of each was calculated from result of the Soxhlet extraction (g/100g) and specific gravity (g/ml) of fat each phantom.Iron concentration in the liver phantoms were determined using the inductively coupled plasma optical emission spectroscopy (ICP-OEP).
Cylindrical Phantom
– Diameter: 20,
30cm φ
– Length: 8cm
Data Acquired
♦CT
Scanner: 64-MDCTscanner equipped with fast switching dual kV switching
Fast kV switching helical scan mode
80/140kVp,
0.6 sec/rot.
640mA,
Helical Pitch 1.375:1,
5mmTH
40mm beam width (0.625 x 64),
CTDIvol: 15.64 mGy
Conventional helical scan mode for CT value measurement
120kVp,
0.5 sec/rot.
500mA,
Helical Pitch 1.375:1,
5mmTH
40mm beam width (0.625 x 64),
CTDIvol: 13.82 mGy
♦MRI
Scanner:1.5T clinical scanner (GE Healthcare,
Signa HDxt,
WI,
USA)
T2* map Pulse sequence: multiecho fast SPGR
TR 50.8 (ms),
TE (8echo) 3.5 9.5 15.5 21.5 27.5 33.5 39.5 45.5 (ms),
FA 10 (deg),
BW 31.3 (kHz)
FOV 18x18 (cm^2)
Slice Thickness 4.0(mm),
14 slices Gap 0.4 (mm),
Matrix 256×256
NEX 2 time 6:10
MRS Pulse sequence: single shot PRESS
TR 1694 (ms),
TE 28 (ms),
Voxel size 15×15×15 (mm)
NEX 1
Data Analysis
♦Overview of new algorithm for fat quantification for fast kVp-switching DECT
MMD-based liver fat quantification can follow two paths. First,
if contrast agent is present in the DECT image,
a virtual un-enhancement (VUE) (6) algorithm is used to digitally remove the enhancing effect of contrast agent. This is important for CT imaging of the liver since most clinical protocols call for multi-phase (contrast-free and contrast-enhanced) scanning (7,
8). The material triplet used for this VUE step is fat,
blood,
and contrast agent. For contrast-free DECT images,
or for contrast-enhanced DECT images after the application of VUE,
fat quantification is performed through dual material decomposition using fat and healthy liver tissue as the material pair.
♦ROI method
(Fig.
2 and Fig.
3)
Each of fat volume fraction (FVF) obtained from DECT (FVFDECT,
%) and CT attenuation from CT images (HUSECT,
HU) was measured by using the average measurement from three regions-of-interest inside the syringes.
For MRS,
FVF from MRS (FVFMRS,
%) was measured for ROIs inside the syringes and was derived from the areas under the water and fat peaks.
Statistical Analysis
Simple linear regressionanalyses were used to examine the associations between actual measured FVF (FVF true) and three imaging results (HUSECT,
FVFDECT and FVFMRS).
Differences in slopes were tested by using analysis of covariance (ANCOVA).
All analyses were conducted at a nodule-level.
R software (Version 2.12.0: R Foundation for Statistical Computing,
Vienna,
Austria) was used for statistical analysis(9).