Improved Method for Accurate and Efficient Quantification of MRS Data with Use of Prior Knowledge

Abstract

We introduce AMARES (advancedmethod foraccurate,robust, andefficientspectral fitting), an improved method for accurately and efficiently estimating the parameters of noisy magnetic resonance spectroscopy (MRS) signals in the time domain. As a reference time domain method we take VARPRO. VARPRO uses a simple Levenberg–Marquardt algorithm to minimize the variable projection functional. This variable projection functional is derived from a general functional, which minimizes the sum of squared differences between the data and the model function. AMARES minimizes the general functional which improves the robustness of MRS data quantification. The newly developed method uses a version of NL2SOL, a sophisticated nonlinear least-squares algorithm, to minimize the general functional. In addition, AMARES uses a singlet approach for imposition of prior knowledge instead of the multiplet approach of VARPRO because this greatly extends the possibilities of the kind of prior knowledge that can be invoked. Other new features of AMARES are the possibility of fitting echo signals, choosing a Lorentzian as well as a Gaussian lineshape for each peak, and imposing lower and upper bounds on the parameters. Simulations, as well asin vivoexperiments, confirm the better performance of AMARES compared to VARPRO in terms of accuracy, robustness, and flexibility.