Event-by-event respiratory motion correction for PET with 3-Dimensional internal-external motion correlation

Respiratory motion during PET/CT imaging can cause substantial image blurring and underestimation of tracer concentration for both static and dynamic studies. In this study, we developed an event-by-event respiratory motion correction method that used 3-Dimensional internal-external motion correlation in listmode reconstruction. In the proposed method, the positional translation of the internal organ during respiration was first determined from the reconstructions of 8 phase-gated images. A level set (active contour) segmentation method was used to segment the targeted internal organ, and the centroids of the segmented organ were determined. The mean displacements of the external respiratory signal acquired by the Anzai system that corresponded to each phase frame of the respiratory cycle were determined. Three linear correlation functions between the centroids of the internal organ (x,y,z) and the Anzai mean displacements were established. An internal organ motion file was then generated by applying the correlation functions to the entire Anzai trace onto each dimension to guide event-by-event motion correction in listmode reconstruction. The proposed method was evaluated with a NEMA phantom driven by a QUASAR respiratory motion platform and three human pancreas studies with the tracer [¹⁸F]FP(+)DTBZ used to study pancreatic beta cell for diabetes acquired on a Siemens mCT PET/CT scanner. Our results show that the proposed method improved the average contrast ratio recovery for the hot spheres in the NEMA phantom by 7% and 33% compared to the reconstructions with only 1D motion correction and no motion correction, respectively, and yielded comparable contrast ratio recovery to the static scan data without increasing image noise. In the human studies, the proposed method improved the image quality by reducing the blurring of pancreas and kidney cortex caused by respiratory motion, and also demonstrated the effectiveness on compensating for intra-gate motion. The measurements of the time-activity-curves showed that the proposed method yielded 18% and 11% increases of the tracer concentration in the pancreas and kidney cortex, respectively. The initial results suggest that the proposed method effectively compensated for respiratory motion while preserving all the counts, and is applicable to dynamic studies.