Related Experiment Video
Updated: May 5, 2026
A Custom Multiphoton Microscopy Platform for Live Imaging of Mouse Cornea and Conjunctiva
Published on: May 17, 2020
The performance of MLEM for dynamic imaging from simulated few-view, multi-pinhole SPECT
Dan Ma1, Paul Wolf, Anne V Clough
1Department of Biomedical Engineering, Case Western University, Cleveland, OH 44105 and formerly with the Department of Biomedical Engineering, Marquette University, Milwaukee, WI, 53201, ( dxm302@case.edu ).
Initializing Maximum Likelihood Expectation Maximization (MLEM) reconstructions with a prior image improves spatial accuracy in dynamic SPECT imaging but can decrease temporal accuracy. This method enhances image quality for stationary small-animal SPECT systems using limited angular views.
Area of Science:
- Nuclear Medicine
- Medical Imaging
- Image Reconstruction
Background:
- Stationary small-animal SPECT systems are advancing for rapid dynamic imaging using limited angular views.
- Reconstructing time-activity curves (TACs) from limited data presents challenges in accuracy and spatial depiction.
Purpose of the Study:
- To quantify the performance of Maximum Likelihood Expectation Maximization (MLEM) for TAC reconstruction from limited angular views in stationary SPECT.
- To evaluate the benefits of initializing MLEM with a prior image (MLEM_IG) derived from late-study data.
Main Methods:
- Simulations were used to assess MLEM and MLEM_IG performance for dynamic TAC reconstruction.
- Studies involved static and dynamic phantoms, including modeling 99mTc-HMPAO uptake in rat lung.
- Quantification focused on spatial depiction accuracy, contrast recovery, and estimation of washin/washout parameters.
Main Results:
- MLEM_IG improved spatial accuracy and contrast depiction in both static and dynamic phantom studies.
- MLEM reconstructions accurately reproduced TAC temporal dynamics, while MLEM_IG showed some deviations in washin parameters.
- The benefit of MLEM_IG decreased with an increasing number of views, and it introduced an additive offset in dynamic images.
Conclusions:
- MLEM reconstruction from limited views with a uniform initial guess accurately captures TAC temporal dynamics but compromises spatial accuracy.
- Initializing MLEM with a prior image enhances spatial accuracy but can reduce temporal accuracy, highlighting a trade-off.
- The findings provide insights into optimizing reconstruction strategies for dynamic SPECT imaging with limited data acquisition.
