Signal-to-noise per unit time optimization for in vivo single-voxel proton magnetic resonance spectroscopy of the brain: Theoretical formulation and experimental verification at two field strengths

Abstract

Signal-to-noise ratio optimization, regarding repetition time selection, was explored mathematically and experimentally for single-voxel proton magnetic resonance spectroscopy. Theoretical findings were benchmarked against phantom measurements at 1.5 Tesla and localized in vivo proton brain spectra acquired at both 1.5 Tesla/3.0 Tesla. A detailed mathematical description of signal-to-noise ratio per unit time was derived, yielding an optimal repetition time of 1.256 times the metabolite longitudinal relaxation time. While long-repetition-time acquisitions minimize longitudinal relaxation time contributions, a repetition time of ~1.5s results in maximum signal-to-noise ratio per unit time, which can in turn be invested into smaller voxel sizes. The latter is of utmost importance in brain oncology, allowing accurate spectroscopic characterization of small lesions.