Purpose: Spectral Detector CT (SDCT) generates virtual mono-energetic (ME) images at different photon energies (keV). The purpose of this study is to evaluate accurate detection of very small blood clots by spectral analysis of mean Hounsfield Unit (HU) as a function of the energy of the ME images.
Material and Methods: Phantoms were constructed using clotted swine blood, placed in 8mm tubes (n=6). Heparinized blood containing 0.5, 1.0 and 2.0 mg/ml iodine (Iomeron 350 mg/ml) was added to the tubes. Control tubes (n=6) were filled with blood with the same iodine concentrations, without clot. The tubes were located in a water phantom, which was scanned with a 64-slice Dual-layer detector CT, at 120 kV, 250 mAs. In addition, an anthropomorphic phantom containing Iodine solutions of various known concentrations (2-20 mg/ml) was scanned. Mono-energetic images at 40, 50, 65, 80 and 200keV were generated for both phantoms. Software was developed to calculate the relationship between Hounsfield Units (HU) of pixels containing iodine solutions in the ME images at 65 keV and 200 keV. This relationship created a spectral map that uniquely characterized the material in the pixel, independent of its concentration. For any given image of the clots, the software generated an iodine image by searching and identifying pixels, which fit into the spectral map equation of iodine and displayed them in different color.
Results: Clot detection was significantly better in the iodine images, generated by the spectral map of the ME images, when compared to the conventional images. The non-clot containing tubes demonstrated a homogeneous map on the iodine images. The number of clots and their size could be easily determined in the iodine images. In the images with the lowest iodine concentration, clot visualization was slightly reduced but was much better than the conventional images.
Conclusion: Iodine images based on spectral maps generated by ME images improved clot visualization in a phantom model, compared to conventional CT images. The results imply that spectral detector CT with the aid of iodine images permit contrast dose reduction while increasing reader confidence of clot detection.
