Foto 7

Santarelli MF, Della Latta D., Scipioni M., Positano V., Landini L., "A Conway–Maxwell–Poisson (CMP) model to address data dispersion on positron emission tomography.", Computers in Biology and Medicine. 77(1): 90-101

Written by

Abstract: Positron emission tomography (PET) in medicine exploits the properties of positron-emitting unstable nuclei. The pairs of γ-rays emitted after annihilation are revealed by coincidence detectors and stored as projections in a sinogram. It is well known that radioactive decay follows a Poisson distribution; however, deviation from Poisson statistics occurs on PET projection data prior to reconstruction due to physical effects, measurement errors, correction of deadtime, scatter, and random coincidences. A model that describes the statistical behavior of measured and corrected PET data can aid in understanding the statistical nature of the data: it is a prerequisite to develop efficient reconstruction and processing methods and to reduce noise.

The deviation from Poisson statistics in PET data could be described by the Conway-Maxwell-Poisson (CMP) distribution model, which is characterized by the centring parameter λ and the dispersion parameter ν, the latter quantifying the deviation from a Poisson distribution model. In particular, the parameter ν allows quantifying over-dispersion (ν<1) or under-dispersion (ν>1) of data. A simple and efficient method for λ and ν parameters estimation is introduced and assessed using Monte Carlo simulation for a wide range of activity values.

The application of the method to simulated and experimental PET phantom data demonstrated that the CMP distribution parameters could detect deviation from the Poisson distribution both in raw and corrected PET data. It may be usefully implemented in image reconstruction algorithms and quantitative PET data analysis, especially in low counting emission data, as in dynamic PET data, where the method demonstrated the best accuracy.

 

Keywords:{Positron emission tomography (PET), Conway-Maxwell-Poisson (CMP) distribution, Maximum likelihood (ML) estimation, Sinograms, Poisson statistic deviation}

Filehttp://dx.doi.org/10.1016/j.compbiomed.2016.08.006