MRI images converted to PET scans to improve dementia diagnosis accuracy
Bridging Modalities on the Cortex: Surface-based MRI to PET Translation with a Diffusion Bridge
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Dementia diagnosis can benefit from PET scans that show brain activity, but PET scans are expensive and use radiation. The researchers created a computer method to predict PET scans just from MRI brain scans by focusing on the brain’s surface shape. Their new approach makes synthetic PET images that help doctors detect dementia more accurately than MRIs alone, and nearly as well as real PET scans. This works even on different patient groups without retraining the model.
What this means in practice
- •For hospital imaging teams: Create synthetic PET scans from MRI data to improve dementia diagnosis accuracy without additional radiation or cost.
- •For medical ai developers: Incorporate surface-based diffusion models using spherical vision transformers to enhance multi-modal brain image synthesis.
Authors
Yitong Li, Alexandra Samoylova, Fabian Bongratz, Timo Grimmer, Dennis M. Hedderich, Igor Yakushev, Christian Wachinger
Abstract
Cortical hypometabolism measured by Fluorodeoxyglucose Positron Emission Tomography (FDG-PET) is a highly sensitive biomarker for dementia diagnosis. However, high costs, radiation exposure, and limited accessibility constrain its clinical utility. While cross-modal synthesis from Magnetic Resonance Imaging (MRI) offers a promising alternative, existing volumetric generation methods do not explicitly account for the highly folded cortical geometry, where disease-related patterns predominantly reside. To address this, we introduce a novel surface-based diffusion bridge framework DB-SUiT for MRI-to-PET translation that operates natively on the cortical manifold. A conditional Spherical U-shaped vision Transformer (SUiT) is specifically designed to model the intricate cross-modal relationships while preserving surface topology. It combines spherical convolutional encoders for multi-scale surface feature extraction with bottleneck Transformers to capture long-range spatial dependencies, while incorporating demographic and subcortical conditions to refine the synthesis. Evaluated on two datasets, including subjects with different dementia types, DB-SUiT demonstrates high-fidelity synthesis that substantially outperforms other baselines. In automated dementia classification, synthesized PET surfaces improve performance over MRI by 14.2% and PET volumes by 11.3%, approaching the performance of real PET surfaces. In a blinded reader study, synthetic PET achieved 85.5% diagnostic accuracy, compared with 75.8% for MRI and 95.2% for real PET. This further demonstrates cross-cohort and cross-pathology generalization, as the model was evaluated without retraining on an external cohort that included a dementia subtype not represented during training. Our code is available at https://github.com/ai-med/DB-SUiT.