List-mode-based reconstruction for respiratory motion correction in PET using non-rigid body transformations
F Lamare1, M J Ledesma Carbayo, T Cresson
1INSERM, U650, Laboratoire du Traitement de l'Information Médicale, Brest, F-29200 France.
Physics in Medicine and Biology
|September 1, 2007
Summary
This study introduces an elastic transformation method for correcting respiratory motion in positron emission tomography (PET) imaging. The new technique improves image quality by utilizing all available data, offering more uniform correction across the lungs.
Area of Science:
- Medical Imaging
- Physics
- Computer Science
Background:
- Respiratory motion degrades image quality in emission tomography.
- Current methods like gated acquisitions yield low signal-to-noise ratio frames.
- A need exists for improved respiratory motion correction techniques.
Purpose of the Study:
- To implement and evaluate an elastic transformation method for respiratory motion correction in list-mode PET reconstruction.
- To compare the elastic transformation method with affine transformation and uncorrected images.
Main Methods:
- Developed an elastic transformation algorithm using B-spline basis functions for respiratory motion.
- Integrated elastic transformations into a one-pass list mode expectation-maximization (OPL-EM) algorithm.
- Evaluated the method using NCAT phantom data and Monte Carlo simulations.
Main Results:
- Both elastic and affine transformations improved lung field motion artifacts.
- Elastic transformation provided more uniform correction across different lesion sizes and locations.
- The elastic transformation method leverages all available data for reconstruction.
Conclusions:
- Elastic transformation in list-mode PET reconstruction is effective for correcting respiratory motion artifacts.
- This method offers superior and more uniform correction compared to affine methods.
- The technique enhances image quality in emission tomography by utilizing comprehensive respiratory data.
