A number of acceleration strategies have been proposed for MRI, especially for dynamic imaging, where there is a trade-off between temporal and spatial resolution. To acquire multiple contrasts within one dataset for simultaneous T1 and T2* measurements^1,2,3 additional acceleration is required. Simultaneous multi-slice imaging (SMS) is a complimentary technique^4,5 which can be combined with standard k-space sampling strategies, such as parallel imaging and keyhole methods. CAIPIRINHA-like⁶ signal shifting for more efficient use of coil-sensitivity profiles improves SMS reconstruction⁷. A multi-shot segmented EPI yields sufficient signal also for longer echo-times and allows a flexible acquisition scheme with respect to the desired echo times. Additionally it enables integrated controlled slice-aliasing to reduce g-factor penalty⁵. To satisfy these aspects we introduce a novel incorporation of SMS to a multi-contrast, segmented EPI sequence⁸ (Fig.1) which allows extended spatial coverage in contrast-enhanced dynamic imaging to fully cover larger parts of the body (e.g. leg’s muscles) without exceeding temporal constrains.

One sedated pig was studied with an 1.5T MR-scanner (Aera, Siemens Healthcare/Germany). After bolus-injection of a contrast-agent (CA) (dose of 0.1 mmol/kg at 5ml/s injection-rate of Dotarem, Guebert/France) dynamic images of the hip and right leg were acquired for 11 minutes according to 500 measurements at a temporal resolution of 1.3 s. For a FOV=400x200 mm² at a spatial resolution of 3.1 x 3.1mm² with 24 slices of 5 mm thickness (20% gap, interleaved acquisition scheme to reduce inter-slice crosstalk) could be achieved due to the 4 times (R=MB) accelerated imaging speed. Despite the sampling rate and a reasonable FOV three contrasts (TE1=9 ms, TE2=1.5 ms, TE3=34 ms) with a TR=261 ms could be acquired without additional in-plane acceleration. The 4 simultaneously excited slices (MB=4) were shifted by ¼ FOV along phase-encoding direction (PE) with blipped CAIPI where gradient blips were integrated into multi-shot scheme^8,9. Sagittal slice orientation was chosen such that the aorta and femoral artery were covered for detection of the arterial input function (AIF). For offline SMS reconstruction a 5x5 MB-kernel was trained with MB and Single-Band (SB) data which was acquired before CA administration and with identical imaging parameters (TR_ACS=1500 ms and TE_ACS=TE_SMS). Slice-GRAPPA reconstruction of the SMS data potentially regularizes the SNR of the SMS data according to the SNR of the auto-calibration-signal (ACS)^5,10 which reduces apparent T1 weighting in the SMS images. For comparison two SMS-reconstruction approaches were tested: SB & MB (both TR=1500 ms) and SB & synthetic MB from SB (TR=1500 ms), in both cases 2 prescans were run for steady-state conditions. Time penalty for ACS data acquisition was 20s prior to the dynamic scan. Collecting multiple contrasts in one acquisition allows signal separation into a S₀ (TE=0 ms) and T2* signal by fitting the data to the exponential T2* decay after SMS reconstruction.The application of SMS yields sufficient temporal resolution for AIF detection in a large vessel while anatomical coverage allows perfusion analysis of manually segmented muscles. As addressed by the FLEET-ACS approach¹⁰ echo-spacing is crucial for correct contrast recovery during slice-GRAPPA reconstruction, in addition to echo-spacing also the TR has to be kept constant for MB-kernel training data. The quality of the reconstructed images meets the requirements for post-processing (Fig.2) such that signal separation by fitting the different acquired contrasts (Fig.3) can be successfully performed. By correcting the signal for the T2* contribution a correction of the susceptibility related changes can be achieved which becomes especially important for the distribution phase with high CA concentrations¹. Furthermore T1 effects of CA leakage can be quantified¹¹. Figure 4 shows the S₀(t) signal inside the femoral artery for the first 50 measurements, enabling AIF extraction. In contrast to regions with fast CA-dynamics the S₀(t) time-curve in muscle-tissue represents the extravasation process into the tissue and the wash-out after a maximum CA-concentration has be reached (Fig.5).

By using a segmented EPI acquisition scheme echo-times can be selected flexibly and reasonable signal in late contrasts can be achieved. TR and TE issues in SMS reconstruction have been addressed and the resulting SMS perfusion images show expected signal dynamics. Contrast-enhanced perfusion measurements were performed on a pig’s hip/leg resulting in mainly muscle-perfusion images with separated S₀(t) and T2* maps which allow a more complex pharmacokinetic analysis^1,2. For further acceleration and/or shorter minimal TEs SMS segmented EPI can be combined with (in-plane) undersampling strategies⁵.

It has be demonstrated that the incorporation of SMS into multi-contrast, segmented EPI can overcome existing limitations for dynamic imaging while retaining desired acquisition parameters such as temporal and spatial resolution even for imaging of relatively large FOVs.