System resolution versus image uncertainty for positron emission tomography scanners

Area of Science:

• Medical Imaging Physics
• Nuclear Medicine Technology
• Quantitative Imaging Analysis

Background:

• Quantitative measures in Positron Emission Tomography (PET) imaging are susceptible to statistical uncertainty.
• Scanner system parameters, particularly spatial resolution, critically influence this uncertainty.
• Accurate prediction of uncertainty is essential for reliable quantitative PET analysis.

Purpose of the Study:

• To compare predicted statistical uncertainty with measured values in PET imaging.
• To experimentally evaluate the impact of varying spatial resolution on quantitative measures.
• To validate theoretical predictions of uncertainty in PET reconstructions.

Main Methods:

• Utilized a dual-ring PET prototype with adjustable detector configurations (1-5.5 mm FWHM resolution).
• Imaged a Micro Jaszczak hot-spot phantom with varying rod diameters (1.2-4.8 mm).
• Reconstructed images using different detector combinations and evaluated statistical properties, including signal-to-noise ratio (SNR).

Main Results:

• Measured SNRs for high, medium, and low resolution configurations generally aligned with predictions.
• A 1 million event dataset was used for analysis across different resolutions.
• The SNR advantage of high-resolution systems was experimentally confirmed.

Conclusions:

• The unique PET tomograph configuration enabled direct experimental comparison of spatial resolution effects.
• Higher spatial resolution in PET systems provides a significant SNR advantage.
• This advantage holds true even for object features larger than the system's detector resolution.