Scatter modeling for 3-D PET list-mode EM reconstruction

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 he 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.