Reduced field-of-view 3D stack-of-spirals perfusion imaging with high spatiotemporal resolution
Yang Yang1, Li Zhao2, Xiao Chen3, Kelvin Chow4, Peter W. Shaw4, Jorge A. Gonzalez4, Frederick H. Epstein1,5, Craig H. Meyer1,5, Christopher M. Kramer4,5, and Michael Salerno1,4,5

1Biomedical Engineering, University of Virginia, Charlottesville, VA, United States, 2Radiology, Beth Israel Deaconess Medical Center & Harvard Medical, Boston, MA, United States, 3Medical Imaging Technologies, Siemens Healthcare, Princeton, NJ, United States, 4Medicine, University of Virginia, Charlottesville, VA, United States, 5Radiology, University of Virginia, Charlottesville, VA, United States


3D CMR perfusion imaging enables whole ventricular coverage at the same cardiac cycle permitting quantification of ischemic burden of patients being evaluated for coronary artery disease. Current 3D techniques have limited spatial-temporal resolution. We developed an efficient outer-volume suppressed 3D Stack-of-Spiral perfusion sequence with motion-guided compressed sensing reconstruction which can acquire 20 partitions with 2 mm in-plane and 4 mm through-plane resolution with a temporal foot print of 180 ms. A pilot study of 10 subjects using this technique demonstrates clinically acceptable image quality.


First-pass contrast-enhanced myocardial perfusion CMR has proven to be a powerful noninvasive technique for evaluating coronary artery disease1. Current techniques image a limited number of 2D slices, and small, but clinically relevant perfusion defects may be missed. 3D perfusion imaging is potentially advantageous for quantifying ischemic burden by covering the whole ventricle at the same cardiac phase, but current techniques have limited spatiotemporal resolution. Considering the heart only occupies a small portion of the chest, reduced field-of-view (rFOV) perfusion imaging with 2D outer volume suppression2 (OVS) significantly improves sampling efficiency and has enabled single-shot 2D spiral perfusion imaging3. We hypothesized that the application of OVS to 3D stack-of-spiral (SoS) perfusion imaging could significantly reduce the temporal footprint while increasing the in-plane and through-plane spatial resolution as compared to previous 3D approaches.


2D OVS preparation was incorporated into a 3D centrically ordered SoS perfusion sequence (Fig 1a). The OVS consists of a non-selective adiabatic BIR-4 tip-down pulse, a 2D spiral spatially selective tip-back pulse and a spoiler to suppress signal outside the heart (Fig 1b). The 3D sequence parameters included: FOV 170×170 mm2, TE 1.0 ms, TR 9.0 ms, saturation recovery time 150 ms, flip angle 35o, 2×2 mm in-plane resolution, 20 partitions with 4 mm thickness, 180ms temporal footprint. For each partition, single 8 ms spiral with a dual density design with a broad Fermi shape transition4 was implemented and the spiral trajectory was rotated by golden angle through time to generate an incoherent k-t sampling pattern. The proposed sequence was performed in 10 healthy subjects with a 0.075 mmol/kg Gd-DTPA bolus on a 1.5T Avanto Siemens scanner. The data was reconstructed with two approaches: 1) (thin-slice) 2×2×4 mm with 180 ms temporal footprint and 2) (high-temporal resolution) 2×2×8 mm with 90 ms temporal footprint using only the central 10 partitions. Block LOw-rank Sparsity with Motion guidance (BLOSM)5 combined with SENSE was used for image reconstruction. Images were graded on a 5 point scale (5 excellent, 1 poor) by a single cardiologist. 2×2×4 mm data-sets were also interpolated to an isotropic 2x2x2 mm resolution for display in arbitrary image orientations.


Fig 1c shows the spatial profile of the 2D OVS with the 100 mm rFOV design. The 1D spatial profile across the center (Fig 1d), illustrates the stopband at approximately ±400 mm, which was large enough to suppress the signals outside heart to prevent spatial aliasing. Fig 2 shows perfusion images at a single time frame during first-pass of contrast from one healthy volunteer using the 2 different strategies: a) 10 slices with 2×2×8 mm and 90 ms temporal footprint presented high SNR and good quality perfusion images; b) 20 slices with 2×2×4 mm and 180 ms temporal footprint showed high through-plane resolution images with similar image quality (3.6±0.6 vs 3.4±0.6 p=NS). By 2x sinc-interpolation of the 2x2x4 mm dataset, 40 slices with isotropic 2 mm resolution can be displayed (Fig 3) enabling isotropic visualization of perfusion in short and long axis image orientations (Fig 4).


OVS enabled high resolution thin-slice (4 mm) 3D perfusion data to be acquired in only 180 ms, while high temporal-resolution 3D data with 8mm slices can be acquired in only 90 ms. 3D centrically ordered allows for a flexible reconstruction strategy to produce either relative higher SNR but lower through-plane resolution (8 mm) perfusion images or lower SNR but higher through-plane (4 mm) resolution. The higher through-plane resolution could reduce partial volume effects and provide better depiction of the apical slices. The isotropic reconstruction provides the ability to reformat the data in arbitrary slice orientations. The short temporal footprint 3D spiral acquisition should have reduced sensitivity to cardiac motion as compared to techniques with a longer temporal footprint resulting in sharper depiction of trabeculae and papillary muscles.


We demonstrated the successful application of rFOV 3D SoS perfusion techniques. The high sampling efficiency using OVS enables 3D imaging with an excellent combination of high in-plane and through-plane spatial resolution and a short temporal footprint of 180 ms. The technique can also achieve 3D perfusion coverage with an 8mm slice thickness in 90 ms, a temporal footprint shorter than most clinically available 2D perfusion pulse sequences. Further validation will be required in patients undergoing adenosine stress CMR.


Grant funding provided by NIH K23 HL112910 and T32 EB003841


1. Schwitter J, Nanz D, Kneifel S, Bertschinger K, Buchi M, Knusel PR, Marincek B, Luscher TF, von Schulthess GK. Assessment of myocardial perfusion in coronary artery disease by magnetic resonance: a comparison with positron emission tomography and coronary angiography. Circulation 2001;103(18):2230-5.

2. Smith TB, Nayak KS. Reduced field of view MRI with rapid, B1-robust outer volume suppression. Magn Reson Med 2012;67(5):1316-23.

3. Yang Y, Zhao L, Chen X, Shaw P, Gonzalez JA, Epstein FH, Meyer CH, Kramer CM, Salerno M. Reduced Field-Of-View Single-Shot Spiral Perfusion Imaging. In Proceedings of the 23rd Annual Meeting of ISMRM, Toronto, Canada, 2015. Abstract 1005.

4. Yang Y, Kramer CM, Shaw PW, Meyer CH, Salerno M. First-pass myocardial perfusion imaging with whole-heart coverage using L1-SPIRiT accelerated variable density spiral trajectories. Magn Reson Med 2015.

5. Chen X, Salerno M, Yang Y, Epstein FH. Motion-compensated compressed sensing for dynamic contrast-enhanced MRI using regional spatiotemporal sparsity and region tracking: block low-rank sparsity with motion-guidance (BLOSM). Magn Reson Med 2014;72(4):1028-38.


Figure 1. a) Schematic of rFOV 3D SoS pulse sequence with centric reordering. b) RF shape and gradient waveforms of the OVS module. c) 2D spatial profile of OVS design with rFOV = 100mm. d) 1D spatial profile across the center of c) to show the stopband is around ±400mm.

Figure 2. Perfusion images from one healthy volunteer. a) 10 slices of 2×2×8 mm resolution with 90 ms temporal footprint. b) 20 slices of 2×2×4 mm resolution with 180 ms temporal footprint.

Figure 3. 2x interpolated 40 slices of 2×2×2mm resolution with 180 ms temporal footprint.

Figure 4. Long axis view perfusion images from isotropic resolution images.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)