Optimized sampling for high resolution multi-pinhole brain SPECT with stationary detectors

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Neurology

Background:

• Single-photon emission computed tomography (SPECT) is crucial for diagnosing neurological and psychiatric conditions.
• Traditional SPECT has limitations in resolution and sensitivity.
• The G-SPECT-I scanner offers high-resolution, focused brain imaging.

Purpose of the Study:

• To develop and evaluate optimized bed translation strategies for whole brain imaging using the G-SPECT-I scanner.
• To minimize scan time by reducing the number of patient bed translations.
• To assess the impact of reduced sampling on image quality and diagnostic accuracy.

Main Methods:

• Simulations of noise-free and noisy SPECT scans were performed.
• Reduced numbers of bed positions (18 and 12) were compared to a reference scan (128 positions).
• Brain uptake ratios (BURs) and asymmetry indices (AIs) were calculated in 51 regions of interest.

Main Results:

• Decreasing bed positions to 18 resulted in minimal deviations (2.2% for BUR, 1.5% for AI) from the reference.
• Using 12 positions led to slightly larger deviations (2.9% for BUR, 2.7% for AI).
• Sampling-induced deviations were significantly smaller than noise-induced deviations.

Conclusions:

• Optimized protocols with 18 or 12 bed positions enable fast whole brain SPECT imaging with G-SPECT-I.
• Reduced scan times (20-30 seconds overhead) are achievable with negligible impact on diagnostic metrics.
• G-SPECT-I facilitates rapid, motion-corrected SPECT protocols for clinical use.