Optimization of Block Sequential Regularized Expectation Maximization Reconstruction Parameters for Striatal SPECT...

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Radiochemistry

Background:

• Striatal SPECT imaging is crucial for diagnosing neurological disorders.
• The block sequential regularized expectation maximization (BSREM) method offers improved image reconstruction.
• Optimizing BSREM parameters is essential for accurate quantitative analysis in SPECT.

Purpose of the Study:

• To determine the optimal noise-suppression (β) and edge-preservation (γ) parameter values for BSREM reconstruction.
• To evaluate image quality and diagnostic accuracy for ¹²³I-labeled SPECT imaging of the striatum.
• To establish optimal reconstruction settings for cadmium-zinc-telluride camera-based SPECT.

Main Methods:

• A striatum phantom with varying ¹²³I concentrations was used.
• SPECT data were acquired using a cadmium-zinc-telluride camera and reconstructed with BSREM.
• 42 images were generated by systematically varying β and γ values, followed by quantitative and visual assessment.

Main Results:

• Increased β values reduced image noise (percent coefficient of variation).
• Increased γ values enhanced image contrast (percent contrast).
• Optimal parameter pairs (β=0.03, γ=5 and β=0.05, γ=10) provided diagnostically suitable images with accurate striatal tracer depiction.

Conclusions:

• The study identified optimal β and γ values for BSREM reconstruction in striatal SPECT.
• These optimized parameters ensure accurate visualization of tracer uptake in striatal regions.
• The findings support improved diagnostic accuracy for neurological conditions using SPECT/BSREM.