Transforming Static CT in Gated PET/CT Studies to Multiple Respiratory Phases

Abstract

Discrepancy between CT and PET data due to motion in PET/CT studies is a significant problem. PET is acquired over a period of time whereas CT is static. Thus attenuation correction of PET data with CT leads to wrong quantification and can result in wrong diagnosis and treatment. A solution based on respiratory gating and optical flow algorithms is presented. A combined local-global optical flow algorithm is used for motion estimation on PET data, the static CT is then transformed to different phases of the respiratory cycle. Corresponding PET and CT data is thus obtained for better attenuation correction. The results are verified on both software phantom and real patient data