Motivation: In diagnosing Parkinson's disease (PD), MRI provides clear structural information and is radiation-free, but not capture functional data like PET. Goal(s): Develop a method to convert MRI images to PET using a conditional diffusion model, aiming to improve PD diagnosis. Approach: We developed a conditional diffusion-based method that uses MRI images and basic patient information as inputs to generate corresponding PET images. Results: Our model demonstrated high accuracy and robustness compared to conventional synthetic image methods. Impact: This development is valuable for the diagnosis of PD in patients, offering a potential tool to enhance diagnostic capabilities.
Liu et al. (Tue,) studied this question.