Image acquisition:
All scans were performed with a Gammex 464 phantom (Gammex,
Madison WI) with a plastic 50-ml tube in its air cavity,
filled with a sample simulating haemorrhage (blood),
diluted iodine and a haemorrhage-iodine (blood-iodine) mixture consecutively.
Each sample was scanned in the phantom with the following scan settings on the clinical IQon SDCT scanner: 200 mAs,
120 kVp,
pitch 1,
gantry rotation time 330 ms and detector collimation 64 x 0.625 mm.
Images were reconstructed using filtered back projection (FBP) with spectral and iterative reconstruction setting (iDose4,
Philips Healthcare,
Best,
Netherlands) level 0.
Reconstructed slice thickness was 3 mm using the standard kernel settings.
All scans were performed three times to account for inter-scan variability.
Sample preparation:
Previous studies show that the density of blood is primarily determined by the haemoglobin concentration (14).
In turn,
changes in hematocrit are reflected in the density of a haemorrhage seen on conventional CT (15).
We collected discarded packed red blood cells (pRBC) and prepared samples to a range of 0,
5,
15,
25,
35,
45,
55,
65,
75 and 85 % hematocrit (HCT).
The pRBC were prepared using phosphate buffered saline (PBS) to prevent cytolysis.
Each sample was scanned with our scan previously described scan protocol.
Diluted iodine (Optiray 350,
Ioversol,
Covidien,
Mansfield,
Massachusetts) in PBS,
with concentrations ranging 0 – 14 mg/ml (0.00,
0.70,
1.40,
2.10,
2.80,
2.50,
4.20,
4.90,
5.60,
7.00,
14.00 mg/ml),
were scanned using our previous described protocol.
We matched the iodine attenuation with the densities found for 85,
75,
65,
55,
45 and 35% HTC to create iodine dilutions with matching densities at those HTC levels.
Blood-iodine mixtures with matching densities were scanned with following proportions: 2/3 blood – 1/3 iodine,
½ blood – ½ iodine,
1/3 blood – 2/3 iodine.
Image analysis:
All image analysis was performed using the Spectral Diagnostic Suite (SDS) (Philips Healthcare,
Best,
Netherlands).
Each diluted blood,
diluted iodine and blood-iodine mixture sample was analysed by placing a circular 2 cm2 Region of Interest (ROI) centrally in the inserts.
Mean attenuation in Hounsfield Units (HU) in the conventional CT and VNC image using the “VNC map” (Fig.
1).
Statistical analysis:
Statistical analysis was performed using SPSS 21.0 (IBM,
Armonk,
NY,
USA).
Mean attenuation in the conventional and VNC image were compared for correlation by using Pearson’s correlation coefficient.
Attenuation measurements are reported as relative retention of attenuation on VNC compared to conventional images (%),
by mean +/- 95% confidence interval (CI).
Categories (3/3 blood,
2/3 blood – 1/3 iodine,
½ blood – ½ iodine,
1/3 blood – 2/3 iodine,
3/3 iodine) are compares by Mann-Whitney U test.
VNC and conventional attenuation values were analysed for accuracy in detecting haemorrhage,
and estimating haemorrhagic component size.
A p-value of <0.05 was considered statistically significant.