Multi-resolution Imaging For Total-body PET/CT

The application of PET has been limited by the trade-offs between SNR(Signal-to-Noise Ratio),scanning time and radiation dose for a long time.Starting from 2016,the two-meter-long total-body PET scanner has been in developing.The two-meter-long total-body PET scanner,which is an innovation to explore the maximum potential of PET,will possess a range of capabilities currently unavailable with the multi-bed multi-pass scanning system.Based on the two-meter-long total-body PET scanner,we expect to reconstruct multiple ROI images with different parameters(voxel size,iteration number and post-filter strength,etc.)simultaneously,and the multi-resolution reconstruction algorithm has been proposed.(1)We proposed the multi-resolution reconstruction algorithm to reconstruct multiple ROI images with different parameters in one process.The images could be displayed separately or fusioned.(2)We proposed the multi-resolution reconstruction algorithm with the classical Siddon algorithm,and the algorithm has been validated with the 2D digital data.(3)To improve the reconstructing speed,we proposed the Parallel ray-tracing algorithm for GPU implementation.Based on this algorithm,the multi-resolution reconstruction algorithm has been also validated by the data from multi-bed PET/CT scanner and total-body PET/CT scanner.(4)NEMA IQ test has been shown to illustrate the consequence between image quality and reconstructing parameters,we expect to choose the suitable reconstructing parameters for each ROI.