Improving the singles rate method for modeling accidental coincidences in high-resolution PET

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Physics

Background:

• Random coincidences are a major source of image degradation in Positron Emission Tomography (PET).
• Accurate correction of random events is essential for high-resolution PET scanners and quantitative analysis.
• The widely used singles rate (SR) method may systematically overestimate randoms, particularly at low energy thresholds.

Purpose of the Study:

• To investigate the performance of the singles rate (SR) method under various conditions (activity, source geometry, energy window).
• To evaluate an alternative 'singles trues' (ST) method for improved randoms estimation.
• To propose and assess an iterative method (STi) using only measurable quantities (prompts and singles) for randoms correction.

Main Methods:

• Investigated the performance of the SR method across different activity levels, source geometries, and energy acceptance windows.
• Developed and tested the 'singles trues' (ST) method, which models true coincidences to refine randoms estimation.
• Proposed and evaluated an iterative version, STi, requiring only prompt and singles count data.

Main Results:

• SR method showed significant overestimation (86-300%) of randoms depending on source geometry.
• ST method reduced deviations to 4-60%, with deviations of 1% or less for conventional energy windows.
• STi deviated slightly from ST at higher activities but reproduced ST results for conventional energy windows, effectively correcting SR overestimations.

Conclusions:

• The ST method significantly improves randoms estimation accuracy compared to SR, especially for conventional energy windows.
• The STi method provides a practical approach for accurate randoms correction using readily available data (prompts and singles).
• Accurate randoms correction using ST or STi is vital for enhancing image quality and quantitative accuracy in high-resolution PET imaging.