Transmission-less attenuation and scatter compensation in myocardial perfusion SPECT with list-mode data

Washington University, Saint Louis MO

Investigators

Linked publications & trials

Abstract

Summary Myocardial perfusion single-photon emission computed tomography (SPECT) (MPS) has an important and well- validated role in the diagnosis of coronary artery disease (CAD). Attenuation and scatter compensation (ASC) in MPS has been shown to provide improved diagnostic accuracy for CAD. Conventional methods for ASC require an independent measurement of the attenuation map using a transmission scan such as an x-ray computed tomography (CT) scan. These CT-based attenuation compensation (CTAC) methods suffer from several issues such as misalignment of SPECT and CT scans, higher radiation dose, increased acquisition time, scanning costs, and hardware complexity. Further, a majority of SPECT systems, including many of the emerging solid-state-detector-based MPS systems that have demonstrated the ability to provide MPS images at low dose, do not have CT-imaging capability. Thus, a method to perform transmission-less ASC is poised to have exceptional impact for CAD diagnosis. The existing methods for transmission-less ASC are often too slow and inaccurate. These methods either discard the scattered photons or ignore the energy attribute of these photons. Since the probability of Compton scatter depends on the attenuation coefficient of the tissue and since there is a relationship between the energy of the scattered photon and the scattering angle, scattered photons can provide information about the attenuation distribution. In fact, we have shown that scattered photons acquired in list-mode (LM) format and including the energy information contain information to estimate the attenuation distribution. Based on this premise, our overall goal is use the entire SPECT LM dataset, including the scattered photon data, to develop a transmission-less ASC method. Our first aim is to develop an iterative statistical method based on inverting the models used for photon transport, including scatter, to jointly reconstruct the activity and attenuation distribution using the LM SPECT emission data. The method will be implemented on a high-performance computing system with several algorithmic and computational optimizations. We hypothesize that the developed method will yield performance that is not inferior to that of CTAC methods for defect-detection and quantitation tasks in MPS. To test this hypothesis, we will objectively evaluate the method for defect- detection tasks using simulation and phantom studies, and quantitation tasks with existing patient data. A successful demonstration of the proposal objectives will establish a foundation for clinical applicability of the developed method and stimulate additional studies on the extension of the proposed method for quantitative SPECT imaging applications and for emerging technologies such as solid-state-detector-based MPS imaging systems. By enabling ASC in the absence of a transmission scan for the existing and emerging imaging technologies and applications, the proposed method has the potential to have tremendous clinical impact.