Tissue-Dependent and Spatially-Variant Positron Range Correction in 3D PET.

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Computational Science

Background:

• Positron range (PR) limits PET image resolution, particularly with radionuclides like 82Rb, 124I, and 68Ga.
• Existing PR modeling methods are often approximations and fail to account for material boundaries, leading to artifacts in regions like lungs.

Purpose of the Study:

• To develop and implement an efficient method for accurate, spatially-variant positron range corrections in PET image reconstruction.
• To minimize the impact of PR on PET image quality and reduce artifacts, especially at tissue interfaces.

Main Methods:

• Pre-computing voxel-dependent PR kernels using CT or segmented images and Monte Carlo simulations.
• Convolving images with these kernels during the forward-projection step of iterative reconstruction algorithms.
• Implementing the method on a preclinical Argus PET/CT scanner.

Main Results:

• Achieved artifact-free, positron range-corrected PET images, even with activity at high-density boundaries.
• The method introduces only a modest increase in overall reconstruction time.
• Demonstrated effectiveness in improving image quality for isotopes with large PR.

Conclusions:

• The proposed method accurately incorporates spatially-variant PR corrections, enhancing PET image quality.
• This approach is applicable to various PET scanners and beneficial for clinical studies using isotopes with significant positron ranges.
• The technique effectively mitigates artifacts caused by positron travel through heterogeneous materials.