Related Experiment Video
Updated: Mar 28, 2026
15:18
Near Infrared Optical Projection Tomography for Assessments of β-cell Mass Distribution in Diabetes Research
Published on: January 12, 2013
17.0K
Performance of 3DOSEM and MAP algorithms for reconstructing low count SPECT acquisitions.
Willem Grootjans1, Antoi P W Meeuwis2, Cornelis H Slump3
1Department of Radiology and Nuclear Medicine, Radboud university medical center, Nijmegen, The Netherlands; Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands.
Zeitschrift Fur Medizinische Physik
|January 4, 2016
Summary
Modern 3D ordered subset expectation maximization (3DOSEM) algorithms improve low count SPECT imaging but introduce noise and distortions. Regularization techniques can mitigate these artifacts, enhancing image quality for clinical applications.
Area of Science:
- Medical Imaging
- Nuclear Medicine
- Image Reconstruction
Background:
- Low count Single Photon Emission Computed Tomography (SPECT) is crucial for whole-body imaging and dose reduction.
- Evaluating advanced reconstruction algorithms is vital for optimizing low count SPECT performance.
Purpose of the Study:
- To assess the performance of 3D ordered subset expectation maximization (3DOSEM) and maximum a posteriori (MAP) algorithms for low count SPECT image reconstruction.
- To compare these algorithms against traditional methods using phantom studies.
Main Methods:
- Utilized NEMA NU2 IQ phantom with varying radiotracer concentrations.
- Reconstructed images using Siemens Flash3D (3DOSEM) and Hermes Hybrid Recon (3DOSEM, MAP).
- Quantified image quality using contrast recovery coefficient (CRC), background variability (N%), contrast-to-noise ratio (CNR), and aspect ratio (AR).
Main Results:
- 3DOSEM algorithms showed improved contrast and resolution over FBP, OSEM, and 2DOSEM.
- Correlated noise patterns and image distortions degraded 3DOSEM image quality, with increased AR and decreased CNR.
- MAP algorithms with anatomical priors yielded sharp images with low distortion and noise, though some artifacts persisted.
Conclusions:
- 3DOSEM algorithms for low count SPECT are susceptible to correlated noise and distortions, impacting image quality.
- Artifacts observed in phantom studies are relevant to clinical SPECT imaging.
- MAP algorithms with appropriate priors offer a promising approach for high-quality low count SPECT reconstruction.
