Technical Implementation of Dual Energy CT (T. Johnson)
Tube potentials and dose
For Dual Energy scanning, the tube potentials are mostly set to 140 and 80 kVp to obtain the largest spectral difference. With the additional filter on the 140 kVp spectrum, 100 kVp can be used as a lower-energy spectrum for some applications. This is especially relevant in the trunk of the body where 80 kVp tube potential can result in insufficient transmission. The protocols can be used to replace routine protocols for the respective application. Regarding radiation exposure, the tube currents are adapted so that the dose is largely equivalent to that of the routine protocol. This implies that the individual 80 and 140 kVp images are noisier than a normal CT image obtained at 120 kVp with the same dose. However, weighted average images are calculated by the image reconstruction system to resemble normal 120 kVp CT images in spectral properties and image noise. These images are immediately available for reading during the clinical routine.
Manufacturer-independent dose measurements with thermoluminiscent detectors in Alderson phantoms have also confirmed that the dose efficiency is very similar, with a marginal increase in noise but also an improved contrast-to-noise ratio at an equivalent dose.
Data handling
The foci of both x-ray tubes of the DSCT scanner are located in the same plane of the z-axis. Projection data are acquired in an angular offset of 90 degrees, and there are no equivalent projections at both potentials. Thus, immediate postprocessing of the projection data is not possible or would be very cumbersome. Postprocessing is therefore performed on the reconstructed images of both acquisitions. Specific Dual Energy CT convolution kernels handle the data in both the 80 and 140 kVp dataset equivalently in order to avoid density changes at object edges that are not caused by differences in spectral behavior.
Image analysis and iodine mapping
The images are further analyzed with a ‘three material decomposition’ (1) implemented in the postprocessing software. The CT numbers are analyzed to assign three constituents of an underlying tissue, e.g. in liver parenchyma, these would be fat, soft tissue, and iodine. The mean CT number at both potentials reflects the relative amount of soft tissue and fat in that voxel. If the CT number at 80 kVp is higher than at 140 kVp, this can be attributed to a certain concentration of iodine. The stronger the enhancement at 80kVp in relation to that at 140 kVp, the higher the concentration of iodine. Thus, the iodine content can be mapped in the image. Also, the iodine map can be subtracted from the normal CT image to obtain a ‘virtual non-contrast’ image.
Related articles: Dual Energy CT – an Introduction, Physical Background, Clinical Applications
References:
1. Johnson TR, Krauss B, Sedlmair M, et al. Material differentiation by dual energy CT: initial experience. Eur Radiol 2007; 17:1510-1517.
