Related Experiment Video
Updated: May 22, 2026
08:30
X-ray Dose Reduction through Adaptive Exposure in Fluoroscopic Imaging
Published on: September 11, 2011
Acquisition time/dose reduction in pediatric PET imaging using patch-based deep learning
Chenyang Han1, Andrew T Trout2,3,4, Andi Li1
1Department of Biomedical Engineering, University of Cincinnati, 2901 Woodside Drive, Cincinnati, OH, 45219, USA.
EJNMMI Physics
|May 21, 2026
Summary
Patch-based deep learning (PDL) significantly reduces noise in pediatric PET imaging, enabling shorter scan times or lower radiation doses. This method maintains image quality and quantitative accuracy, minimizing training data requirements.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Radiology
Background:
- Deep learning (DL) shows potential for dose reduction in pediatric PET.
- Conventional DL requires extensive datasets for generalizability.
- Patch-based DL (PDL) learns from single high-quality scans to reduce training burden.
Purpose of the Study:
- To develop and evaluate a PDL approach for pediatric whole-body PET.
- To improve image quality and quantitative accuracy with reduced radiation dose or scan time.
- To minimize the need for large, diverse training datasets.
Main Methods:
- A PDL model was trained on a single high-quality pediatric PET/CT dataset.
- Reduced-count data (20s/bed) were simulated by truncating standard 90s/bed acquisitions.
- The PDL model was tested on a NEMA phantom and 88 pediatric patient datasets.
Main Results:
- PDL preserved spatial resolution and contrast while reducing noise in phantom studies.
- Observer studies found PDL-enhanced 20s/bed images comparable or superior to standard 90s/bed images in 85% of cases.
- Quantitative analysis showed minimal bias in SUV measurements, outperforming regularized reconstruction.
Conclusions:
- PDL enables significant reduction in pediatric PET acquisition time (to 20s/bed) or administered activity.
- The method maintains spatial resolution, contrast, quantitative accuracy, and diagnostic quality.
- PDL offers a promising solution for low-dose, time-efficient pediatric PET imaging.
