Optimisation of T₂*-weighted MRI for the detection of small veins in multiple sclerosis at 3 T and 7 T.

T₂*-weighted magnetic resonance imaging at 7 T has recently been shown to allow differentiation between white-matter multiple sclerosis lesions and asymptomatic white-matter lesions, by the presence or absence of a detectable central blood vessel. The aim of the present work is to improve the technique by increasing the sensitivity to veins at both 3 T and 7 T, and to assess the benefit of ultra-high-field imaging. Signal-to-noise ratio (SNR) measurements and simulations are used to compare the sensitivity of magnitude T₂*-weighted and susceptibility-weighted images for the detection of small veins (<1 pixel in diameter), both with and without the use of gadolinium. The simulations are used to predict the optimal scanning parameters in order to increase the sensitivity to these veins at both field strengths, and to reduce the inherent dependence on vessel orientation. The sensitivities of the sequences at both field strengths are compared, theoretically and experimentally, in order to quantify the benefit of imaging at ultra-high-field. Subjects with multiple sclerosis (MS) are scanned at both field strengths, using the optimised sequence parameters, as well as those used in previously published work, and the optimisation is shown to improve the detection of veins within lesions.