MRI tractography-guided PET image reconstruction regularisation using connectome-based nonlocal means filtering.

Area of Science:

• Neuroimaging
• Medical Physics
• Biomedical Engineering

Background:

• Positron Emission Tomography (PET) and Diffusion MRI (dMRI) offer complementary insights into neurodegenerative diseases like Alzheimer's.
• dMRI provides crucial data on brain microstructure and structural connectivity (SC).
• Integrating dMRI information into PET reconstruction could enhance image quality but remains underexplored.

Purpose of the Study:

• To develop and evaluate a novel method (CONN-NLM-OSLMAP) for incorporating dMRI-derived SC information into PET image reconstruction.
• To regularize PET images using complementary connectivity data from dMRI.
• To assess the method's effectiveness in noise reduction and contrast enhancement.

Main Methods:

• Proposed a CONNectome-based non-local means one-step late maximum a posteriori (CONN-NLM-OSLMAP) method.
• Integrated dMRI-derived SC information directly into the PET iterative image reconstruction process.
• Evaluated the method using a realistic tau-PET/MRI simulated phantom.

Main Results:

• The CONN-NLM-OSLMAP method demonstrated superior noise reduction and lesion contrast improvement compared to alternative methods.
• Achieved the lowest overall bias in PET image reconstruction.
• Showed more effective and targeted denoising and regularization by incorporating SC information.

Conclusions:

• The proposed method effectively integrates dMRI structural connectivity into PET image reconstruction.
• This integration significantly enhances PET image quality, offering improved diagnostic potential for neurodegenerative conditions.
• Demonstrated the feasibility and effectiveness of using complementary SC information for PET image regularization.