Meeting Banner
Abstract #3738

Quantifying uncertainty in kinetic modelling parameters of hyperpolarized dynamic nuclear polarization data through the applicaton of Bayesian Inference fitting techniques

Samira Kazan 1 , Steven Reynolds 2 , Gillian Tozer 1 , Martyn Paley 2 , and Michael Chappell 3,4

1 CR-UK/YCR Sheffield Cancer Research Centre, University of Sheffield, Sheffield, South Yorkshire, United Kingdom, 2 Academic Unit of Radiology, University of Sheffield, Sheffield, South Yorkshire, United Kingdom, 3 Oxford Centre for Functional MRI of the Brain, University of Oxford, Oxfordshire, United Kingdom, 4 Department of Engineering Science, University of Oxford, Oxfordshire, United Kingdom

Dynamic nuclear polarization (DNP) is a novel technique for increasing the sensitivity of magnetic resonance spectroscopy and imaging. Intravenous administration of hyperpolarized pyruvate provides a means for quantifying pyruvate-lactate interconversion in living tissues via MRS/MRSIand in oncology, is a potential marker for the efficacy of anti-cancer drugs. The rate constant for pyruvate to lactate conversion, kpl, requires mathematical models to extract kinetic parameters, from the spectroscopy data and the quantification of such parameters depends on the mathematical model and the fitting approach used. In this study we determine whether a Bayesian fitting method (previously adopted in the quantification of perfusion from Arterial Spin Labeling) offers improvements in accuracy and robustness compared to common fitting methods such as Nelder-Mead when applied to the quantification of pyruvate to lactate rate constants. Additionally, we use the estimates of the uncertainty in kpl obtained from the Bayesian method to assess whether the variations kpl are the result of fitting error or inter-group variability.

This abstract and the presentation materials are available to members only; a login is required.

Join Here