Local Motion Correction and Influence of Attenuation Correction in Combined Positron Emission and X-Ray Computed Tomography

Abstract

Movement in the area of the lung and the upper abdomen leads to motion blurring in PET images and to misalignment between PET data and CT data used for attenuation correction. This leads to errors in the quantification of the PET data, especially for lesion size and activity uptake. To minimize motion artifacts in the PET images, a local motion correction based on list-mode data was performed. The remaining problem of possible misalignment between PET and the CT based attenuation map, was studied using Monte Carlo simulations. An emission data driven correction of the attenuation map was used to reconstruct the PET data. This was performed for nine patients with solitary lung lesions. Using only the local motion correction a change of measured lesion volume of up to 27 % (mean 14.5 %) and of measured tracer uptake of up to 13 % (mean 4.7 %) was found. If additionally the data driven corrected attenuation map was applied, the changes were up to 54 % (mean 19.7 %) for the lesion volume and up to 36 % (mean 16.7 %) for the tracer uptake. This shows that local motion correction in combination with the data driven attenuation map improves the PET/CT data quantification of lesions in areas with movement.