Anatomy-guided brain PET imaging incorporating a joint prior model

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Reconstruction

Background:

• Positron Emission Tomography (PET) imaging relies on accurate image reconstruction.
• Integrating anatomical information can improve PET image quality and diagnostic accuracy.
• Current reconstruction methods may have limitations in handling complex anatomical-functional relationships.

Purpose of the Study:

• To develop and evaluate a novel Maximum a Posteriori (MAP) framework for PET image reconstruction.
• To incorporate co-registered anatomical images using a new anato-functional joint prior.
• To assess the performance of the proposed method against conventional reconstruction algorithms.

Main Methods:

• Proposed a MAP framework utilizing a novel anato-functional joint prior.
• Employed a hyperbolic potential function based on anatomical image intensity differences.
• Penalization based on intensity differences in reconstructed PET images.
• Validated using simulated (18)FDG PET data and a clinical Florbetapir PET study.

Main Results:

• The joint prior MAP (JP-MAP) reconstruction algorithm quantitatively enhanced reconstructed PET images in simulations.
• JP-MAP outperformed conventional 3D MLEM and 3D MAP methods in visual and quantitative accuracy.
• Demonstrated improved noise versus activity performance and robustness to MR-PET mis-registrations.
• Showed definitive improvements in computational performance compared to other MR-guided MAP methods.

Conclusions:

• The proposed JP-MAP method effectively integrates anatomical information for superior PET image reconstruction.
• This approach offers enhanced quantitative and visual accuracy, particularly beneficial for FDG and Florbetapir PET imaging.
• JP-MAP presents a computationally efficient and robust alternative to existing PET reconstruction techniques.