Iterative image reconstruction for positron emission tomography based on a detector response function estimated from...

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Reconstruction

Background:

• Accurate system models are crucial for high-quality positron emission tomography (PET) image reconstruction.
• Current methods often rely on Monte Carlo simulations for the sinogram blurring matrix, which can be computationally intensive and may not fully capture physical effects.
• Direct measurement of blurring is challenging due to the need for collimated sources.

Purpose of the Study:

• To develop and validate a novel method for estimating the 2D blurring kernels and sinogram blurring matrix in PET imaging using uncollimated point source measurements.
• To improve the accuracy of the PET system model by incorporating physically measured blurring effects.
• To enhance the quality of reconstructed PET images in terms of resolution and contrast.

Main Methods:

• Acquired point source measurements with high count statistics on a microPET II scanner using a high-precision motion stage.
• Developed a monotonically convergent iterative algorithm to estimate the detector blurring matrix, leveraging rotational symmetry and detector block structure.
• Integrated the measured sinogram blurring matrix into a maximum a posteriori (MAP) iterative image reconstruction algorithm.

Main Results:

• The proposed method demonstrated improved resolution and contrast ratio in reconstructed images compared to reconstructions using Monte Carlo-based blurring matrices or no detector response model.
• The method accurately accounts for physical effects in photon detection that are difficult to model with simulations.
• The computational reconstruction time was not significantly affected, as the blurring component represents a small fraction of the total reconstruction time.

Conclusions:

• The proposed method provides a more accurate system model for PET image reconstruction by using measured blurring kernels.
• This technique offers advantages over Monte Carlo simulations, including easier application to transformed data and better modeling of physical processes.
• The method is applicable to various PET scanners for both human and animal imaging, promising enhanced diagnostic capabilities.