The Detection of Infarcted Stroke Tissue via Localised Sodium Concentration Measurements: A Non-Invasive Approach

Abstract

Quantitative 23Na Magnetic Resonance Imaging (qNa-MRI) is a non-invasive technique which has considerable potential for measuring Tissue Sodium Concentration (TSC) changes in pathological brain tissue states such as stroke or tumour. However, the quantification of 23Na with MRI has been hindered by limitations in Signal-to-Noise Ratio (SNR) within the required spatial and temporal resolution constraints. The aim of the work presented here was to develop a 23Na MRI resonator system with an optimal compromise between SNR and B1-field homogeneity to allow for accurate qNa-MRI, to adapt an MRI sequence for short Time to Echo (TE) imaging, to apply the technique to an existing in vivo model of stroke in the rat brain, and to process the data to gain novel insights into the spatio-temporal TSC evolution after ischaemic stroke. The design, development, and characterisation of a transceiver (TXRX) and a transmit-only receive-only (TORO) coil system are described. The developed coils were compared to a commercial double-tuned 23Na/1H TXRX surface coil. As a result, the developed double-tuned TXRX surface coil achieved two-fold SNR improvement in the 23Na channel, while maintaining the high sensitivity in the 1H channel. Furthermore, the developed double-tuned 1H/ 23Na volume resonator achieved a B1-homogeneity better than 5 % across the sample volume. In conjunction with the developed 23Na receive-only surface coil, up to three-fold better SNR was achieved in sample depths of 12 mm, the depth of interest for rat brain imaging. In a second development phase, a 2D-radial sequence was optimised to shorten the TE after the 23Na excitation pulse to below 1 ms. A TSC quantification method was developed for the dual resonator system and tested on phantoms containing known 23Na concentrations. This technique was then applied to the measurement of TSC maps in vivo in a rodent model of cerebral ischaemia with a spatial resolution of 1.2 ?l and a 10 min image acquisition time. Such high spatio-temporal resolution allowed, for the first time, for the study of regionally dependent TSC change in vivo with a quantification accuracy of approximately ? 10 mM for up to eight hours after stroke induction. In contrast to previous studies in this area, it was found that TSC increased immediately in core tissue, but was delayed by up to 4 h in tissue considered to be still-viable but at risk of infarction. For the first time, the bio-energetic cell failure in infarcted stroke tissue was measured and spatial-temporally resolved. An increase in TSC was hypothesised to indicate infarction, and it is therefore felt that elevated TSC could serve as an irreversible and non-invasive marker for tissue damage after the onset of ischaemic stroke.