Optimization of Reconstruction Parameters for Discovery 710 Positron Emission Tomography/Computed Tomography

Area of Science:

• Nuclear Medicine
• Medical Imaging Physics

Background:

• Quantitative accuracy in (18F) fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) imaging is crucial for reliable diagnosis and treatment monitoring.
• Standardized Uptake Value (SUV) quantification is sensitive to image reconstruction parameters, necessitating optimization for consistent results.
• The NEMA image quality phantom is a standard tool for evaluating PET/CT system performance and reconstruction algorithms.

Purpose of the Study:

• To optimize quantitative (18F) FDG PET/CT imaging by assessing the impact of reconstruction parameters on recovery coefficients (RCs).
• To evaluate how matrix size, iterations, subsets, and Gaussian postfilters influence SUV quantification accuracy.
• To provide guidance on selecting optimal reconstruction parameters for improved PET/CT quantitative analysis.

Main Methods:

• Utilized the 'Vue Point FX + Sharp IR' algorithm with 3D-ordered subset expectation maximization (3D-OSEM), time-of-flight, and point spread function technologies.
• Investigated various reconstruction parameters, including matrix sizes (e.g., 256x256), iterations, subsets, and Gaussian postfilters.
• Employed the NEMA image quality phantom to measure recovery coefficients and assess SUV quantification accuracy across different sphere sizes.

Main Results:

• For spheres ≥17 mm, a 256x256 matrix size and mean SUV provided accurate quantification.
• For spheres ≤13 mm, maximum SUV was more suitable, with postfiltering significantly impacting accuracy, especially for small spheres.
• Larger matrix sizes partially mitigated Gibbs artifacts and improved SUV quantification for various sphere sizes.

Conclusions:

• Optimizing PET reconstruction parameters is critical for enhancing SUV quantification accuracy in (18F) FDG PET/CT imaging, particularly for small lesions.
• Adherence to European Association of Nuclear Medicine/EARL guidelines for reconstruction parameter selection is essential for reliable quantitative measurements.
• Careful consideration of matrix size and postfiltering is necessary to ensure precise and reproducible quantitative PET/CT data.