Skip to main content

Late Gadolinium Enhancement imaging using stack of stars and compressed sensing


To develop and test a 3D hybrid radial(stack of stars) acquisition scheme with compressed sensing(CS) reconstruction for high resolution late gadolinium enhancement(LGE) imaging of the left ventricle(LV) and the left atrium(LA).


LGE imaging is the gold standard for identifying viable tissue and fibrosis in the LV [1]. It has also been recently applied to imaging of the LA to identify regions of RF ablation [2] and fibrosis [3]. While a Cartesian acquisition scheme is commonly used, it can take a long time [1] and can be sensitive to motion leading to inconsistent image quality [3, 4]. A radial acquisition scheme can be advantageous due to its robustness to motion and undersampling. Here a hybrid radial acquisition is combined with CS to improve LGE imaging efficiency.


A kz-first segmented acquisition was implemented on a Siemens 3T scanner. This scheme results in having consistent in-plane data with inconsistencies along kz leading to improved image quality as shown with simulation studies in [5]. As well, each plane is rotated by an angle based on golden ratio - Figure 1. This rotation helps exploit sparsity also along the z-dimension - a CS reconstruction with a 3D total variation constraint was used. Acquired data was first interpolated onto a 3D Cartesian grid and minimization of the cost function for CS was done with a gradient descent scheme.

Figure 1
figure 1

Schematic of kz-first hybrid radial lines after the inversion pulse are first acquired along kz before acquiring kx-ky. The kx-ky planes are color coded to illustrate consistency in the data.

The methods were tested for imaging the LV of three rabbits and for imaging the LA of three patients at least three months post-ablation. The scan parameters for the rabbit studies were TR=2.4 ms,TE=2.7ms,voxel size=0.9x0.9x2mm3,TI=300-350ms. For the patient studies the scan parameters were TR=3.4ms,TE=1.83ms,voxel size=1.25x1.25x2.5mm3,TI=300-400ms;36 slices with 144 rays in each plane were acquired in 1.25-2min.


Figure 2 shows the results from three rabbit studies. Enhancement in the lateral free wall corresponds to tissue necrosis from a ligation of a coronary artery. Axial slices of LA from three patient studies are shown in Figure 3. Enhancement corresponds to ablated regions in the atrium.

Figure 2
figure 2

Results lf LGE imaging of the LV with proposed scheme. One representative slice from each of the three studies is shown. Arrow in the image points to enhancement.

Figure 3
figure 3

Results of LGE imaging of the LA. One representative slice from each of the three studies is shown. Arrows in the images point to enhancement.


The 3D stack of stars acquisition in conjunction with CS offers a promising alternative for rapid high resolution LGE imaging. More patient studies are required to validate the proposed framework.


  1. Simonetti OP, et al: Radiol. 2001, 218: 215-223.

    Article  CAS  Google Scholar 

  2. Peters DC, et al: Radiol. 2007, 243: 690-695. 10.1148/radiol.2433060417.

    Article  Google Scholar 

  3. Oakes RS, et al: Circulation. 2009, 119: 1758-1767. 10.1161/CIRCULATIONAHA.108.811877.

    Article  PubMed Central  PubMed  Google Scholar 

  4. Peters DC, et al: JACC Cardiovasc Imaging. 2009, 2: 308-316. 10.1016/j.jcmg.2008.10.016.

    Article  PubMed Central  PubMed  Google Scholar 

  5. Adluru G, et al: ISMRM. 2010, 1285-

    Google Scholar 

Download references

Author information

Authors and Affiliations


Rights and permissions

Open Access This article is published under license to BioMed Central Ltd. This is an Open Access article is distributed under the terms of the Creative Commons Attribution License ( ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Cite this article

Adluru, G., Chen, L., Kim, SE. et al. Late Gadolinium Enhancement imaging using stack of stars and compressed sensing. J Cardiovasc Magn Reson 13 (Suppl 1), P162 (2011).

Download citation

  • Published:

  • DOI: