Quantitative myocardial perfusion PET parametric imaging at the voxel-level

Area of Science:

• Nuclear Medicine
• Cardiovascular Imaging
• Medical Physics

Background:

• Quantitative myocardial perfusion (MP) PET offers potential for early atherosclerosis detection and coronary artery disease (CAD) assessment.
• Accurate quantification of myocardial blood flow (MBF) and flow reserve (MFR) is crucial but challenging with noisy dynamic PET data, especially using short-lived tracers like rubidium-82 (82Rb).
• Existing methods often require heavy post-smoothing, which degrades important functional information.

Purpose of the Study:

• To develop a robust voxel-level quantitative MP-PET framework for improved visualization and quantification of MBF and MFR.
• To address the challenge of high noise levels in 82Rb dynamic PET images.
• To reduce noise without significant loss of spatial resolution for more accurate physiological and pathological assessments.

Main Methods:

• A novel methodology termed 'physiological clustering' was developed.
• This approach leverages functional similarity between voxels to penalize kinetic deviations from physiological partners.
• The method was validated through extensive simulations, including various perfusion defect types, and clinical studies.

Main Results:

• Physiological clustering significantly reduced noise while preserving spatial resolution compared to traditional post-smoothing.
• The method demonstrated superior noise-versus-bias performance and enhanced recovery of perfusion defects (quantified by CNR).
• Parametric images generated using physiological clustering proved robust even with high noise levels in voxel time-activity curves.

Conclusions:

• Physiological clustering is a feasible and robust approach for voxel-level quantitative MP-PET.
• It overcomes the limitations of high noise in 82Rb PET, enabling more accurate MBF and MFR quantification.
• This methodology enhances the diagnostic potential of MP-PET for evaluating coronary artery disease and microvascular function.