Quantitative analysis shows that contrast medium in positron emission tomography/computed tomography may cause significant artefacts

OBJECTIVESAttenuation correction algorithms are required for accurate quantification of PET data and for mapping of radioactive tracers. Modern PET systems incorporate computed tomography (CT) systems to perform attenuation correction. However, high-density media, such as contrast agents, may introduce potentially clinically significant artefacts in PET images when CT-based attenuation correction algorithms are used. Although various groups have investigated this issue, no study has quantitatively assessed the clinical significance of these artefacts by comparing artefact and lesion standardized uptake values (SUVs) in controlled phantom experiments. Furthermore, previous studies have focussed on the effects of increasing the concentration of contrast medium, without investigating the effects of increasing its transaxial area. This study quantifies the clinical significance of increasing the concentration and transaxial area of contrast agents and evaluates a commercially available contrast agent correction algorithm. METHODSImages of a phantom containing background activity, a volume of contrast agent and varying sizes of hot lesions were acquired using clinical acquisition protocols. Quantitative analysis was performed on transaxial image slices of PET data. RESULTSThe densest medium caused a 125% SUVmean increase in the area containing, and immediately adjacent to, contrast medium when compared with a reference water phantom. As the transaxial area of the contrast medium increased, artefacts appeared as a ring of activity around the periphery of the contrast medium. The contrast correction algorithm reduced these artefacts to within ±39% of the reference results. CONCLUSIONOral and IV contrast agents can cause clinically significant artefacts in CT-based attenuation-corrected PET images and should be used with caution.