Identification of Contrast Media in PET/CT Using Dual Energy CT

Abstract

Iodine-containing contrast media facilitate interpretation of CT images. In PET/CT such CT images are often converted to 511 keV attenuation maps to correct the PET images for tissue attenuation. Current conversion methods based on single-energy CT do not account for contrast media, and may therefore generate artifacts in the PET images. A method is presented in which fractional differences of CT images at two energies (DECT) are interpreted, using well-known properties of dual-energy X-ray imaging, to identify regions of contrast media. The identified regions can then be assigned soft-tissue attenuation values to minimize effects of contrast media on the PET attenuation-corrected images. Two cylinder phantoms, including a bone insert and inserts containing iodine-based contrast, were imaged in a PET/CT scanner at 140, 120, 100 and 80 kVp. DECT fractional difference images were created for 120, 100 and 80 kVp relative to 140 kVp. Fractional differences ((HU <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">80</sub> -HU <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">140</sub> )/HU <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">140</sub> ) were close to 0.9 for contrast, regardless of concentration, and 0.47 for bone. This is due to steeper decrease in CT attenuation with increasing energy for iodine, compared to bone, and potentially allows automatic detection of contrast-containing tissue. Results were corroborated with theoretical calculations. DECT permits reliable separation of contrast from soft tissue and bone on CT. This identification of contrast could be used to avoid contrast errors in PET/CT attenuation correction by preventing erroneous automatic conversion of contrast containing regions to inaccurate 511 keV attenuation coefficients.