Related Experiment Video
Updated: May 22, 2026
Hybrid PET/MRI Imaging of Alzheimer's Disease Based on 18F-AV-1451
Published on: April 18, 2025
Realistic PET image synthesis from MRI for automated inference of brain atrophy and Alzheimer's
Brandon Theodorou1,2, Anant Dadu2,3, Brian Avants4
1Department of Computer Science, University of Illinois at Urbana-Champaign, Urbana, IL, USA.
Abstract:
Positron emission tomography (PET) is a crucial tool in medical imaging diagnostics but remains costly and less accessible than alternatives like X-ray and MRI. To address this, we propose MRI2PET, a 3D diffusion-based model that generates AV45-PET scans from T1-weighted MRI images. MRI2PET incorporates style-transferred pre-training and a Laplacian pyramid loss to leverage unpaired MRI data and structural correspondences between modalities while simultaneously emphasizing crucial details. Using the ADNI dataset, we demonstrate MRI2PET produces realistic PET images and improves downstream clinical classification. Notably, augmenting PET-only training data with MRI2PET-synthesized scans increases AUROC from 0.688 ± 0.014 to 0.780 ± 0.005 when classifying into one of cognitively normal, mild cognitive impairment, and Alzheimer's disease groups. These results highlight MRI2PET's ability to generate high quality, clinically informative PET scans from widely available MRI, offering an accessible, cost-effective approach to enhance machine learning performance and expand diagnostic imaging workflows.
Insights
This study introduces MRI2PET, a novel AI model generating PET scans from MRI data. This innovation enhances diagnostic accuracy for conditions like Alzheimer's disease, making advanced imaging more accessible.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Neuroscience
Background:
- Positron emission tomography (PET) is vital for diagnostics but faces accessibility and cost challenges compared to MRI.
- Developing cost-effective and accessible imaging solutions is crucial for widespread clinical application.
Purpose of the Study:
- To develop MRI2PET, a 3D diffusion model for generating AV45-PET scans from T1-weighted MRI images.
- To assess the clinical utility of MRI2PET-generated PET scans for downstream diagnostic tasks.
Main Methods:
- Utilized a 3D diffusion-based generative model (MRI2PET) trained on T1-weighted MRI data.
- Incorporated style-transferred pre-training and Laplacian pyramid loss for enhanced image generation.
- Evaluated performance using the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset.
Main Results:
- MRI2PET successfully generated realistic and clinically informative PET images from MRI data.
- Augmenting training data with synthesized PET scans improved classification accuracy for cognitive impairment and Alzheimer's disease (AUROC increased from 0.688 to 0.780).
- The model demonstrated improved downstream clinical classification performance.
Conclusions:
- MRI2PET offers a cost-effective and accessible method for generating high-quality PET scans from readily available MRI data.
- This approach can significantly enhance machine learning model performance in diagnostic imaging.
- The findings support the integration of MRI-based PET synthesis into clinical workflows for broader diagnostic capabilities.
