Time-domain semi-parametric estimation based on a metabolite basis set

Abstract

A novel and fast time-domain quantitation algorithm--quantitation based on semi-parametric quantum estimation (QUEST)--invoking optimal prior knowledge is proposed and tested. This nonlinear least-squares algorithm fits a time-domain model function, made up from a basis set of quantum-mechanically simulated whole-metabolite signals, to low-SNR in vivo data. A basis set of in vitro measured signals can be used too. The simulated basis set was created with the software package NMR-SCOPE which can invoke various experimental protocols. Quantitation of 1H short echo-time signals is often hampered by a background signal originating mainly from macromolecules and lipids. Here, we propose and compare three novel semi-parametric approaches to handle such signals in terms of bias-variance trade-off. The performances of our methods are evaluated through extensive Monte-Carlo studies. Uncertainty caused by the background is accounted for in the Cramér-Rao lower bounds calculation. Valuable insight about quantitation precision is obtained from the correlation matrices. Quantitation with QUEST of 1H in vitro data, 1H in vivo short echo-time and 31P human brain signals at 1.5 T, as well as 1H spectroscopic imaging data of human brain at 1.5 T, is demonstrated.

On the use of Cramér–Rao minimum variance bounds for the design of magnetic resonance spectroscopy experiments

“…Given that metabolite signals in a proton MR spectrum usually have considerable overlap that makes the quantification difficult, generally, one of three different approaches is taken: 1) use of a single non-specific one-dimensional spectrum (e.g. a localized short echo time (TE) spectrum) followed by linear combination model fitting based on prior knowledge about the constituent metabolites and spectral parameters (Provencher, 1993;Ratiney et al, 2005;Slotboom et al, 1998;Wilson et al, 2011), or 2) use of a dedicated (socalled editing) one-dimensional experiment optimized for exclusive or selective sensitivity for a single metabolite of interest, usually followed by simple model peak fitting or signal integration (Allen et al, 1997), or 3) use of a standard localized two-dimensional MR spectrum followed by peak integration (Thomas et al, 1996;Thomas et al, 2001) or prior knowledge fitting Gonenc et al, 2010;Kiefer et al, 1998;Kreis et al, 2005;; Thomas et al, 2008;van Ormondt et al, 1990;Vanhamme et al, 1999). In cases 1 and 3, the choice of experimental parameters like TE and repetition time (TR) is most often based on general considerations about maximum signal for given relaxation times, insensitivity to changes in relaxation times or arguments about minimization of macromolecular baseline contributions, while in case 2 the signal yield of wanted and unwanted metabolites and their relative overlap is modeled based on quantum mechanical simulations or solution measurements.…”

Quantification Improvements of 1H MRS Signals

“…On the other hand, adjusting the first order phase corresponds to the modification of the begin time of the MR signal (Chen et al, 2002). Although some quantification (Poullet et al, 2007;Provencher, 2001;Ratiney et al, 2005) are able to take into consideration phase distortions and provide accurate metabolite estimates even if the spectra are not zero-phased, other quantification methods, such as peak integration, require zero-phased spectra in order to obtain reliable metabolite estimates.…”

Temperature dependence of 1H NMR chemical shifts and its influence on estimated metabolite concentrations

“…However, since the analysis of short TE 1 H MR spectra of the brain is often hampered by severe signal overlap, the use of prior knowledge on the chemical shifts and the J-coupling constants for all relevant metabolites is of central importance. Thus, well established quantification programs such as LCModel [8], QUEST [9,10], or AQSES [11] use a model function for each metabolite to minimise the number of variables during the fitting procedure. These model functions are either measured on phantom solutions or simulated using published values of chemical shifts and J-coupling constants as prior knowledge [12,13].…”