Astrocytic tracer dynamics estimated from [1-11c]-acetate pet measurement

Andrea Arnold, Daniela Calvetti, Albert Gjedde, Peter Iversen, Erkki Somersalo

Research output: Contribution to journalArticle

Abstract

We address the problem of estimating the unknown parameters of a model of tracer kinetics from sequences of positron emission tomography (PET) scan data using a statistical sequential algorithm for the inference of magnitudes of dynamic parameters. The method, based on Bayesian statistical inference, is a modification of a recently proposed particle filtering and sequential Monte Carlo algorithm, where instead of preassigning the accuracy in the propagation of each particle, we fix the time step and account for the numerical errors in the innovation term. We apply the algorithm to PET images of [1-11C]-acetate-derived tracer accumulation, estimating the transport rates in a three-compartment model of astrocytic uptake and metabolism of the tracer for a cohort of 18 volunteers from 3 groups, corresponding to healthy control individuals, cirrhotic liver and hepatic encephalopathy patients. The distribution of the parameters for the individuals and for the groups presented within the Bayesian framework support the hypothesis that the parameters for the hepatic encephalopathy group follow a significantly different distribution than the other two groups. The biological implications of the findings are also discussed.

Original languageEnglish (US)
Pages (from-to)367-382
Number of pages16
JournalMathematical Medicine and Biology
Volume32
Issue number4
DOIs
StatePublished - Dec 1 2015
Externally publishedYes

    Fingerprint

Keywords

  • Parameter estimation
  • Particle filters
  • PET imaging
  • Sequential monte carlo (SMC)
  • Tracer kinetics

ASJC Scopus subject areas

  • Applied Mathematics
  • Modeling and Simulation
  • Biochemistry, Genetics and Molecular Biology(all)
  • Environmental Science(all)
  • Immunology and Microbiology(all)
  • Neuroscience(all)
  • Pharmacology

Cite this