A Conway-Maxwell-Poisson (CMP) model to address data dispersion on positron emission tomography

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Statistical Modeling

Background:

• Positron emission tomography (PET) relies on radioactive decay, typically modeled by Poisson statistics.
• PET data can deviate from Poisson distributions due to physical factors and data corrections.
• Accurate statistical modeling is crucial for effective PET image reconstruction and noise reduction.

Purpose of the Study:

• To introduce and validate the Conway-Maxwell-Poisson (CMP) distribution for modeling PET data.
• To assess a novel method for estimating CMP parameters (λ and ν).
• To demonstrate the utility of CMP in analyzing raw and corrected PET data.

Main Methods:

• Utilized the Conway-Maxwell-Poisson (CMP) distribution with parameters λ and ν.
• Developed and assessed a simple, efficient parameter estimation method.
• Validated the method using Monte Carlo simulations and experimental PET phantom data.

Main Results:

• The CMP distribution effectively models deviations from Poisson statistics in PET data.
• The proposed estimation method is accurate across various activity levels.
• CMP parameters successfully detected statistical deviations in both raw and corrected PET data.

Conclusions:

• The CMP distribution offers a robust statistical framework for PET data analysis.
• The developed method accurately quantifies deviations from Poisson statistics.
• CMP implementation can enhance quantitative accuracy in PET imaging, particularly for low-count dynamic studies.