Motion effect reduction using high performance combined registration in PET/CT image fusion

2728 Objectives PET/CT images fusion accuracy may suffer from respiratory motion of patient body. The motion effect is even crucial if two sets of PET/CT images need to be fused together (one base line image and one delayed image). To get better clinical images, it is necessary to correct PET data for motion artifacts. In this paper, a combined registration method is introduced to reduce the motion effect in PET/CT images fusion. Methods Generally, motion effect would reduce the fusion quality of PET/CT images of the same patient at different time. If we fuse two sets of PET/CT images to observe the lesion without any motion correction, the position of the lesion is deviated from its true location. In this work, a novel high performance 3D combined registration method is applied to two delayed CT images to obtain the transformation field between two images. The transformation field can be used to correct the deformation between two PET/CT images. The combined registration method contains four main steps: pre-location of organ (e.g. liver), pre-rigid registration of whole image, accurate rigid registration of organ (e.g. liver) and accurate non-rigid registration of whole image. All registration algorithms are based on mutual information. After applying the motion correction method, the motion effect to the PET/CT images could be reduced to proper level. Results The chess fusion results between two delayed PET/CT images show smooth boundary of liver, lung, kidney, spleen and even small vessel. The combined registration method could correct the respiratory motion artifacts of body in a details level. The motion effect is significantly reduced after applying this registration method. Color fusion images of two PET data also show good alignment of organ. Conclusions In this study, the proposed methods reduce mismatch between two sets of PET/CT fusion images due to respiratory motion. The method could be used to register PET and CT images and to reduce the respiratory artifact in PET/CT scan.