Tripled Readout Slices in Multi Time-Point pCASL Using Multiband Look-Locker EPI

Abstract

Multi time-point pseudo-continuous arterial spin labelling (pCASL) with a Look-Locker EPI readout can sample the signal curve of blood kinetics at multiple time points after the labelling pulse. However, due to signal relaxation of labelled blood, the number of readout slices is limited. The aim of this study is to employ a multiband excitation technique to triple the number of readout slices in multi time-point pCASL. The multiband technique, along with 2-fold in-plane parallel imaging, was incorporated into the Look-Locker EPI for the multi time-point sampling of blood kinetic behaviour following the pCASL labelling scheme. The performance evaluation of the multiband and the single-band techniques were performed on four healthy subjects using a 32-channel head RF coil at 3T. Quantitative perfusion maps were analysed using a combination of labelling with and without flow suppression gradients. The perfusion maps provided by the multiband accelerated multi time-point pCASL were in good agreement with the conventional single-band technique. Multiband acceleration caused SNR loss but offered quantitative perfusion maps in 6.23 min with 18 slices compared with 6 slices within the same time period for the single-band method. As conclusion, the multiband technique can successfully triple the number of readout slices while achieving comparable perfusion data in the same measurement time as the conventional single-band readout.

Fig 1. Schematic representation of the slice groups selected using MB RF pulses.

The slice groups were recorded in the ascending direction. Three slices, shown in the same line style, are selected simultaneously by one MB RF pulse.

Fig 2. (a) EPI-based images acquired with single and multiband excitations. All the images were reconstructed after the unfolding procedure for the 2-fold in-plane parallel imaging. With the use of the MB technique, the number of slices was increased to 18 (cf. 6 in the SB data). Using MB excitation with an MB factor of 3 and blipped-CAIPI, 3 simultaneously excited slices were folded together with a phase shift. After reconstruction with the 'Slice-GRAPPA' algorithm, these 3 folded slices were clearly separated. (b) Corresponding g-factor maps from the MB technique.

Fig 3. (a) Multi time-point perfusion-weighted images without crusher gradients (top row) and with crusher gradients (bottom row) with SB and MB, respectively. (b) Plot of the averaged signals from the whole brain over the time TI in ms. (c) Signal fitting of the arterial and tissue signals to the kinetic models.

Fig 4. (a) Quantitative CBF ranging from 0 to 140 mL/100g/min when using the SB and MB techniques. Increased brain coverage in sagittal orientation (arrows) was observed with the MB method. (b) Corresponding histograms of the CBF from SB (top) and MB (bottom).

Fig 5. CBF maps acquired employing SB and MB methods on four subjects.

Fig 6. Averaged CBFs of grey matter and white matter with the SB and MB methods with respect to the subject number.

The error bar indicates the standard deviation.