Related Experiment Video
Updated: Jun 5, 2026
08:40
Positron Emission Tomography Imaging for In Vivo Measuring of Myelin Content in the Lysolecithin Rat Model of Multiple Sclerosis
Published on: February 28, 2021
MR-Guided PET Denoising and Resolution Enhancement Improves Visual Interpretation and Preserves Quantitative Behavior
Caroline Szujewski1, Timothy M Shepherd2, Munir Ghesani3
1Microstructure Imaging, Inc., 370 Jay St FL7, Brooklyn, 11201, NY, USA.
Medrxiv : the Preprint Server for Health Sciences
|June 4, 2026
Summary
A novel MR-guided PET method enhances amyloid PET image quality and reader accuracy, particularly for low amyloid burden cases. This improves diagnostic confidence and reduces false negatives in Alzheimer's disease detection.
Area of Science:
- Neuroimaging
- Radiochemistry
- Biomarker Discovery
Background:
- Amyloid-β PET is crucial for Alzheimer's disease diagnosis and therapy evaluation.
- Image interpretability is limited by low spatial resolution and partial volume effects, especially in low/borderline amyloid burden cases.
- Discrepancies between visual reads and quantitative metrics (SUVr, Centiloid) complicate amyloid PET interpretation.
Purpose of the Study:
- To evaluate a vendor-neutral MR-guided PET (MRG) method for denoising and resolution enhancement.
- To assess if MRG improves image interpretability and reader performance across multiple amyloid tracers.
- To determine if MRG preserves established quantitative biomarkers and increases concordance between visual and quantitative reads.
Main Methods:
- Compared standard (STN) and MRG PET images for four tracers ([18F]florbetapir, [18F]florbetaben, [18F]flutemetamol, [11C]PiB).
- Assessed quantitative equivalence using Standardized Uptake Value ratio (SUVr) and Centiloid scores.
- Conducted a blinded multi-reader study evaluating image quality, artifacts, confidence, and binary amyloid positivity for three tracers.
Main Results:
- MRG preserved quantitative SUVr and Centiloid metrics (R² > 0.90) across all tracers without bias.
- MRG significantly improved concordance between visual reads and quantitative burden measures.
- The multi-reader study showed MRG significantly enhanced image quality, reduced artifacts, and increased reader confidence (p < 0.0001).
- Reader accuracy improved from 0.89 to 0.94, and false-negative rates decreased from 0.08 to 0.04, most notably in low-burden cases.
Conclusions:
- MRG denoising and resolution enhancement improve perceived image quality, reader confidence, and accuracy for amyloid PET.
- The method preserves standard quantitative behavior (SUVr/Centiloid) across tracers.
- MRG enhances cortical definition in challenging low-burden cases, potentially reducing visual-quantitative discordance and supporting confident interpretation near the positivity threshold.