Related Experiment Video
Updated: Jul 16, 2026
10:02
Whole-body PET/MRI of Pediatric Patients: The Details That Matter
Published on: December 19, 2017
Image optimization for low-dose 18F-FDG breast PET/MRI using deep learning: a pilot study
Quinton J Keigley1, Leah C Henze Bancroft1, Kelley Salem1
1Department of Radiology, University of Wisconsin School of Medicine and Public Health, 600 Highland Avenue, Madison, WI, 53792, USA.
EJNMMI Physics
|July 15, 2026
Summary
Low-dose 18F-FDG breast PET/MRI with denoising achieves acceptable image quality and lesion conspicuity. This approach may facilitate wider clinical use of this advanced breast cancer imaging technique.
Area of Science:
- Medical Imaging
- Nuclear Medicine
- Oncology
Background:
- Positron emission tomography with magnetic resonance imaging (PET/MRI) offers noninvasive molecular characterization for breast cancer.
- Improving diagnostic accuracy, staging, and treatment assessment are key goals for personalized breast cancer care.
- Reducing radiation dose in 18F-FDG PET/MRI is crucial for broader clinical adoption.
Purpose of the Study:
- To evaluate the diagnostic image quality and lesion conspicuity of low-dose 18F-FDG breast PET/MRI.
- To assess the impact of a deep learning-based denoising algorithm on image quality and radiotracer uptake.
- To compare simulated low-dose (SLD) and denoised (DN) PET/MRI datasets with full-dose images.
Main Methods:
- A pilot study analyzed 23 women with primary invasive breast cancer undergoing 18F-FDG breast PET/MRI.
- Simulated low-dose (SLD) images were created by random undersampling, approximating a 90% reduction in injected activity.
- A deep learning-based denoising (DN) algorithm was applied to SLD data; image quality, lesion conspicuity, and SUV values were assessed by three readers.
Main Results:
- Denoised (DN) datasets showed significantly improved image quality, sharpness, and signal-to-noise ratio compared to simulated low-dose (SLD) images (p < 0.001).
- DN images were rated as equivalent or slightly inferior to full-dose images for diagnostic quality in 89.9% of reads.
- Readers preferred DN images over SLD images, and quantitative image noise was significantly reduced in DN datasets (p < 0.0001).
Conclusions:
- Acceptable diagnostic image quality and lesion conspicuity can be achieved with low-dose 18F-FDG breast PET/MRI using denoising.
- This denoising technique shows promise for reducing radiation exposure while maintaining diagnostic performance.
- Further prospective validation is warranted to confirm the clinical utility of low-dose 18F-FDG breast PET/MRI with denoising.
![Visualization and Quantification of Brown and Beige Adipose Tissues in Mice using [18F]FDG Micro-PET/MR Imaging](/_next/image?url=https%3A%2F%2Fcloudfront.jove.com%2FCDNSource%2Fteasers%2F62460.jpg&w=3840&q=50)