Purpose: Primary and secondary hypervascularized liver tumors may be treated with transarterial chemoembolization (TACE). The purpose of this study was to experimentally quantify the effect of different chemoembolization materials on the PET activity concentration in PET/CT.
Materials and methods: Different concentrations of lipiodol, tungsten, tantalum, and a different number of platinum coils embedded in a carrier substance were placed in a liver phantom. An insert filled with only the carrier substance served as the negative control. The liver phantom was placed in a body phantom. The liver phantom was filled with 63.3 KBq [18-F]-Fluor-2-deoxy-D-glucose (FDG)/ml water, the body phantom was filled with 19.7 KBq FDG/ml water. PET/CT was performed and PET attenuation correction was performed based on the CT data. We defined: Activity concentration over embolization material (kBq/ml) approximately measured activity concentration; activity concentration over negative control (kBq/ml) approximately real activity concentration. An overestimation of the activity concentration was quantified by the following ratio: Activity concentration overestimation = activity concentration over embolization material (kBq/ml)/activity concentration over negative control (kBq/ml).
Results: All chemoembolization materials led to an overestimation of the PET activity concentration when using CT information for PET attenuation correction. The extent of overestimation is dependent on the concentration and the density of the chemoembolizing agent. PET activity overestimation was 11-151% with lipiodol, 34-1827% with tungsten, 16-1205% with tantalum, and 4-29% with platinum coils.
Conclusion: Conventional chemoembolization materials cause an overestimation of the PET activity concentration in CT-based attenuation-corrected PET/CT images. This is of importance for the clinical routine since activity concentration quantification may not be used in the presence of chemoembolizing agents for imaging follow-up. If an increased FDG uptake is detected after transarterial chemoembolization, non-attenuation-corrected PET images must be assessed in addition to the attenuation-corrected images in order to differentiate artificially increased tracer uptake from a true increase in activity concentration of the tracer. The use of non-attenuating chemoembolizing materials (e.g. drug-eluting beads) for TACE may serve as an alternative to avoid embolization-associated PET artifacts.