Yajing Zhang1, Jiazheng Wang2, and Yishi Wang2
1Philips Healthcare, Suzhou, China, 2Philips Healthcare Greater China, Beijing, China
Synopsis
We demonstrated the feasibility of 5.3 times
acceleration for 3D T2-weighted breast imaging with a 1 mm isotropic resolution
at 1.5 T, using Compressed SENSE. This technique showed either better image
quality or enhanced imaging speed when compared to traditional SENSE with
different imaging schemes.
Introduction
Compressed
sensing is one of the state-of-the-art acceleration technologies for MRI
acquisition and reconstruction [1]. Compressed SENSE technology is built on the
combination of compressed sensing and SENSE [2] theories by leveraging a balanced
under-sampling acquisition scheme with variable density and incoherent sampling
pattern [3]. Iterative reconstruction is used to solve an inverse problem with
the priori information of coil sensitivities and a sparsity constraint. This
study aimed to demonstrate the feasibility of compressed SENSE for 3D
T2-weighted breast imaging.Method
The whole-breast imaging was performed on an Ingenia Prodiva 1.5 T MR scanner (Philips
Healthcare, Suzhou, China) on healthy volunteers with informed consent, study
approved by a local IRB. A 7-channel phase-arrayed breast coil was used for a
3D turbo spin echo (TSE) acquisition with TSE factor 66. The imaging FOV =
360x360 mm2, voxel size = 1mm isotropic, number of slices = 200, and
TE/TR = 170/1300 ms. Compressed SENSE was applied with 5.3 times acceleration
to enhance the imaging efficiency (hereafter CS-SENSE). To examine the image
quality with comparable acceleration, scans were also performed with 1) 5 times
SENSE acceleration in phase encoding direction (hereafter SENSE-1), 2) 2.6 and
2.0 times SENSE acceleration in both phase encoding and slice encoding
directions (hence a total acceleration factor of 5.2, hereafter SENSE-2), and
3) 2.6 and 2.0 times SENSE acceleration in both phase encoding and slice
encoding directions with oversampling in the slice encoding direction (hereafter
SENSE-3). Other imaging parameters were the same as abovementioned. Results
Imaging time for
CS-SENSE, SENSE-1, SENSE-2, and SENSE-3 were 2m39s, 2m40s, 2m32s, and 5m17s,
respectively. Figure 1 illustrates a comparison of CS-SENSE and SENSE-1 using
the same net acceleration factor of about 5. Severe aliasing artifact was
observed in the SENSE-1 image while not significant in the image from CS-SENSE.
In-plane artifacts were reduced in SENSE-2 when compared to that in SENSE-1, while
the through-plane artifact started to present (Figure 2). Discussion
Our results
showed that, under the same acceleration factor of about 5, 3D T2-weighted
images can be obtained using compressed SENSE without observable imaging
artifacts, at a relatively low field (1.5 T) with a moderate quality coil (7
channels). At the same time, it posed a challenge for traditional SENSE to
achieve such a high acceleration factor, where the acceleration factor is heavily
limited by the number and alignment of coil elements. When the SENSE
acceleration was distributed into two directions, folding-artifacts started to
appear in the slice direction, which could be removed by oversampling in the
same direction at the cost of doubled imaging time. A further study should be
performed to verify the clinical significance of compressed SENSE for
T2-weighted breast imaging.Conclusion
For 3D T2-weigted
breast imaging, image acquisition can be highly accelerated by compressed SENSE
at 1.5 T.Acknowledgements
No acknowledgement found.References
[1] Lustig M,
Donoho D, Pauly JM. Sparse MRI: the application of compressed sensing for rapid
MR imaging. Magn Reson Med. 2007; 58: 1182-1195
[2] Pruessmann
KP, Weiger M, Scheidegger MB, Boesiger P. SENSE: sensitivity encoding for fast
MRI. Magn Reson Med. 1999; 42: 952–962
[3] Bratke G, Rau R, Weiss K, et al. Accelerated MRI
of the lumbar spine using compressed sensing: quality and efficiency. J Magn
Reson Imaging 2018; DOI: 10.1002/jmri.26526