Image quality optimization and evaluation of linearly mixed images in dual-source, dual-energy CT

Abstract

In dual-source dual-energy CT, the images reconstructed from the low- and high-energy scans (typically at 80 and 140 kV, respectively) can be mixed together to provide a single set of nonmaterial-specific images for the purpose of routine diagnostic interpretation. Different from the material-specific information that may be obtained from the dual-energy scan data, the mixed images are created with the purpose of providing the interpreting physician a single set of images that have an appearance similar to that in single-energy images acquired at the same total radiation dose. In this work, the authors used a phantom study to evaluate the image quality of linearly mixed images in comparison to single-energy CT images, assuming the same total radiation dose and taking into account the effect of patient size and the dose partitioning between the low-and high-energy scans. The authors first developed a method to optimize the quality of the linearly mixed images such that the single-energy image quality was compared to the best-case image quality of the dual-energy mixed images. Compared to 80 kV single-energy images for the same radiation dose, the iodine CNR in dual-energy mixed images was worse for smaller phantom sizes. However, similar noise and similar or improved iodine CNR relative to 120 kV images could be achieved for dual-energy mixed images using the same total radiation dose over a wide range of patient sizes (up to 45 cm lateral thorax dimension). Thus, for adult CT practices, which primarily use 120 kV scanning, the use of dual-energy CT for the purpose of material-specific imaging can also produce a set of non-material-specific images for routine diagnostic interpretation that are of similar or improved quality relative to single-energy 120 kV scans.

Figure 1

A semianthropomorphic thoracic phantom and three additional attenuation layers were used to represent small, medium, large, and extra-large adults. Two syringes, one with an iodine concentration of 3.5 mg I∕cm³ and the other with 7.0 mg I∕cm³, were placed in the water-filled cardiac regions of the phantoms. The dotted circle on each phantom represents the ROI where the background noise level was measured.

Figure 2

Maximum CNR in dual-energy mixed images as a function of the percent of the total dose used for the 80 kV scans at each of four phantom sizes. The maximum CNR was obtained for each phantom size using the optimal weighting factor from Eq. 4 to form the dual-energy mixed images.

Figure 3

CNR in the dual-energy mixed images (black symbols) and the single-energy images (gray symbols) for the same total radiation dose. The CNR in single-energy images is shown as a function of kV for each phantom size. The CNR in dual-energy mixed images is shown when the 80 kV dose fraction varied from 30% to 70% for each phantom size (top horizontal axis).

Figure 4

Minimum noise in dual-energy mixed images as a function of the fraction of the total dose used for the 80 kV scan at each of four phantom sizes. The minimum noise was obtained for each phantom size by using the optimal weighting factor from Eq. 2 to form the dual-energy mixed images.

Figure 5

Noise in the dual-energy mixed images (black symbols) and the single-energy images (gray symbols) for the same total radiation dose. Noise in single-energy images is shown as a function of kV for each phantom size. Noise in dual-energy mixed images is shown when the 80 kV dose fraction varied from 30% to 70% for each phantom size (top horizontal axis).