Joint reconstruction of activity and attenuation map using LM SPECT emission data

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Computational Imaging

Background:

• Single photon emission computed tomography (SPECT) imaging correction relies on computed tomography (CT) scans.
• CT scans increase patient radiation dose, cost, and hardware complexity.
• Attenuation in SPECT is linked to Compton scattering, suggesting scattered photons may provide attenuation data.

Purpose of the Study:

• To investigate joint reconstruction of activity and attenuation maps using SPECT list-mode (LM) data, including scattered photons.
• To assess the feasibility of eliminating CT scans for attenuation correction in SPECT.
• To explore the fundamental limits of information retrieval from LM SPECT data.

Main Methods:

• Developed a path-based formalism for processing scattered photon data.
• Derived analytic expressions for the Cramér-Rao bound (CRB) of activity and attenuation map estimates.
• Proposed a maximum-likelihood (ML) scheme using LM SPECT data for joint reconstruction.
• Developed an expectation-maximization (EM) algorithm to compute the ML solution.

Main Results:

• Demonstrated the potential for joint reconstruction of activity and attenuation maps from LM SPECT data.
• Provided a framework to analyze the fundamental limits of this joint reconstruction.
• Introduced an ML-based EM algorithm for practical implementation.

Conclusions:

• Joint reconstruction using scattered photon data in SPECT is a promising approach.
• This method could potentially replace CT scans, reducing radiation exposure and costs.
• Further research can optimize the reconstruction algorithms and validate clinical applicability.