Improving diffusion MRI using simultaneous multi-slice echo planar imaging

K Setsompop, J Cohen-Adad, B A Gagoski, T Raij, A Yendiki, B Keil, V J Wedeen, L L Wald, K Setsompop, J Cohen-Adad, B A Gagoski, T Raij, A Yendiki, B Keil, V J Wedeen, L L Wald

Abstract

In diffusion MRI, simultaneous multi-slice single-shot EPI acquisitions have the potential to increase the number of diffusion directions obtained per unit time, allowing more diffusion encoding in high angular resolution diffusion imaging (HARDI) acquisitions. Nonetheless, unaliasing simultaneously acquired, closely spaced slices with parallel imaging methods can be difficult, leading to high g-factor penalties (i.e., lower SNR). The CAIPIRINHA technique was developed to reduce the g-factor in simultaneous multi-slice acquisitions by introducing inter-slice image shifts and thus increase the distance between aliased voxels. Because the CAIPIRINHA technique achieved this by controlling the phase of the RF excitations for each line of k-space, it is not directly applicable to single-shot EPI employed in conventional diffusion imaging. We adopt a recent gradient encoding method, which we termed "blipped-CAIPI", to create the image shifts needed to apply CAIPIRINHA to EPI. Here, we use pseudo-multiple replica SNR and bootstrapping metrics to assess the performance of the blipped-CAIPI method in 3× simultaneous multi-slice diffusion studies. Further, we introduce a novel image reconstruction method to reduce detrimental ghosting artifacts in these acquisitions. We show that data acquisition times for Q-ball and diffusion spectrum imaging (DSI) can be reduced 3-fold with a minor loss in SNR and with similar diffusion results compared to conventional acquisitions.

Copyright © 2012 Elsevier Inc. All rights reserved.

Figures

Fig. 1
Fig. 1
Left: Modulate SLR 90° and 180° RF pulses for simultaneous multi-slice excitation. Right: comparison of the resulting excitation profiles of 1× standard and a 3× simultaneous multi-slice 90°–180°–180° excitation at on resonance and at 50 Hz.
Fig. 2
Fig. 2
Comparison of FOV/2 ghost correction methods for blipped-CAIPI acquisition. A) Collapsed 3 slice-accelerated acquisition with FOV/2 inter-slice image shift, B)–D) image artifact in the separated center slice as a percent signal change, for standard, tailored, and tailored with two GRAPPA kernels reconstruction methods. Significant inter-slice ghost artifact (mainly from the top slice) can be observed with standard ghost correction reconstruction. Minor reduction of this artifact is achieved via the tailored ghost correction method, while a major improvement is provided by the addition of the two GRAPPA kernels method. E) Flow diagram of the tailored ghost+two GRAPPA kernels method. F) Shows the justification for the two GRAPPA kernel method, where (i) shows the application of a GRAPPA kernel to the even (blue) and odd (yellow) lines of the collapsed k-space data in the presence of even/odd phase imperfection, and (ii) illustrates the differences in k-space coverage in the aligned k-space co-ordinate for the even and odd line application of the kernel. With these differences in k-space coverage, a different kernel should be use for the odd and even line. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 3
Fig. 3
Results from 3× slice-accelerated blipped-CAIPI acquisition with FOV/2 inter-slice shift. A) Unfolded images of an aliased slice group and the corresponding Monte-Carlo generated SNR ratio maps, where the SNR retention is close to 100% in all locations B) bootstrap metrics comparison. Left: 95% confidence interval of the primary (CI 1) and the secondary (CI 2) fiber directions and Jenson-Shannon Divergence (JSD) of the ODF of the Q-ball reconstruction for i) 3× slice-accelerated (4 min), ii) non-accelerated (12 min) and iii) two averages of 3× slice-accelerated (6 min) acquisitions. C) The corresponding voxel histograms of the confidence intervals and JSD of the three acquisition schemes and of iv) 1.5 averages of 3× slice-accelerated acquisition (5.5 min) and v) non-accelerated acquisition with reduced FOVz (with 21 slices to achieve a matching TR to the 63 slices 3× slice-accelerated acquisition, 4 min). The angular uncertainties and JSD measures of the 3× slice-accelerated acquisition were very similar to that of the reduced FOVz non-accelerated acquisition. The performances of these acquisitions were marginally worse than that of the standard non-accelerated acquisition which a longer TR and slightly higher signal. The 1.5 and 2 averages of 3× slice-accelerated acquisitions provided progressively better performance; illustrating the gain in SNR per unit time of the blipped-CAIPI acquisition.
Fig. 4
Fig. 4
Results from 3× slice and 2× in-plane accelerated blipped-CAIPI acquisition with FOV/4 inter-slice shift. A) Unfolded images of an aliased slice group and the corresponding Monte-Carlo generated SNR ratio. B) Bootstrap metrics comparison left: 95% confidence interval of the primary (CI 1) and the secondary (CI 2) fiber directions and JSD of the ODF of the Q-ball reconstruction i) 3× slice and 2× in-plane accelerated (3.3 min), ii) 2× in-plane accelerated (10 min) and iii) three averages of 3× slice and 2× in-plane accelerated (10 min) acquisitions. C) The corresponding voxel histograms of the confident intervals and JSD of the three acquisition schemes and of iv) two averages of the 3× slice and 2× in-plane accelerated (6.6 min) acquisition. The uncertainty measures of the 3× slice and 2× in-plane accelerated (3.3 min) acquisition are higher than that of the 2× in-plane accelerated (10 min) acquisition; highlighting an SNR reduction per shot. The 2 and 3 averages of the 3× slice and 2× in-plane accelerated acquisition (6.6 and 10 min) both provide superior performances in comparison to the 2× in-plane accelerated acquisition, with performance clearly improving with more averages. This illustrates the gain in SNR per unit time of the blipped-CAIPI acquisition.
Fig. 5
Fig. 5
Voxel histogram plots of the confident intervals and JSD of three independent bootstrap datasets of the 3× slice and 2× in-plane accelerated acquisition. Good agreement between the uncertainty measures from the three independent datasets can be observed.
Fig. 6
Fig. 6
Comparison of the general fractional isotropy (GFA), b=0 image, b=3000 image, and Q-ball based orientation distribution function of a zoomed in region (orange square) of the 2× in-plane accelerated (10 min), and 3× slice and 2× in-plane accelerated (3.3 min) acquisitions. Similar results are observed.
Fig. 7
Fig. 7
Comparison of the tractography results of 1× (45 min) vs. 3× (15 min) slice accelerated DSI acquisitions (256 directions). The top panel shows whole-brain tractography results while the bottom panel shows the tracts that reside within a 17.5 mm coronal slab. The 1× and 3× tractography results appear to be very similar. These results also highlight the benefit of DSI acquisition in the reconstruction of multiple fiber orientations within voxels.
Fig. 8
Fig. 8
A) Axial view of white-matter pathways labeled from streamline DSI tractography in 1× (45 min) and 3× (15 min) data. Visible in this view are the forceps minor and major of the corpus callosum, the anterior thalamic radiations, the cingulum, the superior longitudinal fasciculus, and the superior endings of the corticospinal tract. B) Average FA (left) and volume in number of voxels (right) for each of the 18 labeled pathways, as obtained from the 1× (green) and 3× (yellow) data sets. Intra-hemispheric pathways are indicated by “L-” (left) or “R-” (right). The pathways are: corpus callosum — forceps major (FMAJ), corpus callosum — forceps minor (FMIN), anterior thalamic radiation (ATR), cingulum-angular (infracallosal) bundle (CAB), cingulum-cingulate gyrus (supracallosal) bundle (CCG), corticospinal tract (CST), inferior longitudinal fasciculus (ILF), superior longitudinal fasciculus-parietal bundle (SLFP), superior longitudinal fasciculus-temporal bundle (SLFT), uncinate fasciculus (UNC). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 9
Fig. 9
Results from the 1418-direction DSI acquisition (Rsl×Rinplane =3×2) acquired using the CONNECTOM gradient and a 64-channel coil array. Left: diffusion weighted images at different b values, illustrating good SNR and contrast in a single shot up to b=10,500 s/mm2right: tractography result of tracts that reside within a 16 mm coronal slab.

Source: PubMed

Nadchodzące badania kliniczne

Ustawienia plików cookie

Używamy plików cookie, aby zapewnić najlepszą jakość korzystania z naszej witryny. Aby uzyskać więcej informacji, przeczytaj uważnie naszą Politykę prywatności i pliki cookie. ...>>>Możesz zmienić swoje preferencje lub wycofać zgodę w dowolnym momencie, usuwając pliki cookie ze swojej strony internetowej lub komputera, zgodnie z opisem w polityce. Zaakceptuj go i wybierz swoje preferencje, zaznaczając odpowiednie pola w sekcji „Zarządzaj ustawieniami” poniżej. Przeczytaj uważnie nasze Warunki korzystania z usług, zanim zaczniesz korzystać z jakiejkolwiek części tej witryny.
«Zarządzaj ustawieniami