Search

High resolution 3D PET scanners with high count rate performance, such as the quad-HIDAC, place new demands on image reconstruction algorithms due to the large quantities of high precision list-mode data which are produced. Therefore a reconstruction algorithm is required which can, in a practical time frame, reconstruct into very large image arrays (submillimetre voxels, which range over a large field of view) whilst preferably retaining the precision of the data. This work presents an algorithm which meets these demands: Regularized One-Pass List-mode EM (ROPLE). The algorithm operates directly on list-mode data, passes through the data once only, accounts for finite resolution effects in the system model and also includes regularization. The algorithm performs multiple image updates during its single pass through the list-mode data, corresponding to the number of subsets that the data have been split into. The algorithm has been assessed using list-mode data from a quad-HIDAC, and is compared to the analytic reconstruction method 3D RP.

Dynamic (4-D) PET imaging is a routinely used technique applied in clinical studies, and is now also finding application for high resolution animal PET systems. The usual method of implementing the image reconstruction involves separating the acquired data out into lower count discrete time frames, resulting in increased image noise. This paper presents two further developments for one-pass 4-D list-mode expectation-maximization (EM) reconstruction, which has already demonstrated enhanced reconstruction quality for image sequences compared to conventional strategies, as well as significantly reducing reconstruction time. The technique, which is an extension of one-pass 3-D list-mode EM, filters a previously reconstructed frame, which is then used as a starting image for the following frame, thus including prior object information. A new approach to post-reconstruction smoothing is also proposed, which in some cases is able to reduce noise without loss of resolution. Using measured quad-HIDAC data from a moving point source and a specially designed liquid flow experiment, and a rat scan, the two approaches are compared to the uniform image initialization method in terms of spatial and temporal resolution, noise, and accuracy of time-activity curves. Preliminary results, using resolution recovery, indicate the potential of achieving ultra-high spatial resolution with good temporal resolution, arising from improving the use of otherwise wasted prior information.

It is widely recognized that scattered radiation is a significant issue in 3-D positron emission tomography (PET). This work proposes a method to address the modeling and correction of scatter in PET studies when using the 3-D list-mode expectation-maximization (EM) algorithm. The rationale of this technique is to directly incorporate an estimated scatter function into the system model of the 3-D list-mode EM algorithm. In this initial work, the scatter model assumes that the scatter distribution can be described by, a convolution of the true activity estimate with a scatter kernel. For small-volume studies acquired with the quad-HIDAC, it is expected that convolution with a shift-invariant scatter kernel will be adequate to describe the broad scatter distributions encountered. It is anticipated that the method proposed here will offer an improvement upon techniques that use subtraction of the scatter background estimate.

High-resolution three-dimensional (3-D) positron emission tomography (PET) scanners with high count rate performance, such as the quad-high density avalanche chamber (HIDAC), place new demands on image reconstruction algorithms due to the large quantities of high-precision list-mode data which are produced. Therefore, a reconstruction algorithm is required which can, in a practical time frame, reconstruct into very large image arrays (submillimeter voxels, which range over a large field of view) whilst preferably retaining the precision of the data. This work presents an algorithm which meets these demands: one-pass list-mode expectation maximization (OPL-EM) algorithm. The algorithm operates directly on list-mode data, passes through the data once only, accounts for finite resolution effects in the system model, and can also include regularization. The algorithm performs multiple image updates during its single pass through the list-mode data, corresponding to the number of subsets that the data have been split into. The algorithm has been assessed using list-mode data from a quad-HIDAC and is compared to the analytic reconstruction method 3-D reprojection (RP) with 3-D filtered backprojection.

The quad-HIDAC PET scanner offers unparalleled high resolution PET imaging capability (