Impact of metallic dental implants on CT-based attenuation correction in a combined PET/CT scanner

Our objective was to study the effect of metal-induced artifacts on the accuracy of the CT-based anatomic map as a prerequisite for attenuation correction of the positron emission tomography (PET) emission data. Twenty-seven oncology patients with dental metalwork were enrolled in the present study. Data acquisition was performed on a PET/CT in-line system (Discovery LS, GE Medical Systems, Milwaukee, Wis.). Attenuation correction of emission data was done twice, using an 80-mA CT scan (PET(CT80)) and a (68)Ge transmission scan (PET(68Ge)). Average count in kBq/cc was measured in regions with and without artifacts and compared for PET(CT80) and PET(68Ge). Data analysis of region of interests (ROIs) revealed that the ratio (ROIs PET(CT80)/ROIs PET(68Ge)) and the difference (ROIs PET(CT80) minus ROIs PET(68Ge)) had a higher mean of values in regions with artifacts than in regions without artifacts (1.2+/-0.17 vs 1.06+/-0.06 and 0.68+/-0.67 vs 0.15+/-0.17 kBq/cc, respectively). For most of the studied artifactual ROIs, the PET(CT80) values were higher than those of the PET(68Ge). Attenuation correction of PET emission data using an artifactual CT map yields false values in regions nearby artifacts caused by dental metalwork. This may falsely estimate PET quantitative studies and may disturb the visual interpretation of PET scan.