Volume 16 Supplement 1

Abstracts of the 17th Annual SCMR Scientific Sessions

Open Access

Super-resolution reconstruction of late gadolinium enhanced MRI from multiple views

  • Qian Tao1,
  • Oleh Dzyubachyk1,
  • Dirk Poot3,
  • Hildo J Lamb1,
  • Katja Zeppenfeld2,
  • Boudewijn PF Lelieveldt1, 4 and
  • Rob J van der Geest1
Journal of Cardiovascular Magnetic Resonance201416(Suppl 1):O40


Published: 16 January 2014


Image resolution is a crucial factor for myocardial scar quantification from late gadolinium enhanced (LGE) MR. The conventional usage of anisotropic short-axis LGE volumes may result in overestimation of the scar size, in particular at the gray zone, due to partial volume effect (PVE). This may impact the diagnostic accuracy of LGE in clinical applications.


A group of 37 post-infarction patients with ventricular tachycardia (VT), who underwent both MRI and a VT catheter ablation procedure with electro-anatomical voltage mapping (EAVM), were included. LGE imaging was performed using a 3D breath-hold inversion recovery turbo-field echo sequence with full ventricular coverage in three views: short-axis (SA), two-chamber, and four-chamber. The three LGE sequences were combined as follows: 1) The inter-scan heart motion was compensated by a joint localized gradient-correlation-based volume registration scheme; 2) The registered volumes were combined into an isotropic volume by the super-resolution reconstruction (SRR) technique. The SRR reconstructed volumes were compared to conventional SA volume using the following measures: the full-width-half-maxima (FWHM)-identified myocardial scar size in terms of core zone and gray zone; the agreement between the scar zones and the gold-standard EAVM data. The agreement was evaluated by the voltage distribution of the mapping points which were back projected onto the normal myocardium, scar core zone and gray zone.


The SRR volume achieved increased resolution in the through-plane direction (Figure 1). Compared to the SA volume, the SRR volume resulted in significantly reduced scar size estimation for both scar core zone (9.6 ± 7.8 ml vs. 8.7 ± 8.3 ml, p < 0.05) and gray zone (11.9 ± 8.2 ml vs. 7.9 ± 6.0 ml, p < 0.05). Average EAVM bipolar voltage for SRR and original SA volume were in normal myocardium 3.7 ± 3.3 mV versus 3.8 ± 3.3 mV (p = NS); in scar core 0.8 ± 0.6 mV versus 1.0 ± 1.0 mV (p = 0.04) and scar gray zone 1.2 ± 0.8 mV versus 1.6 ± 1.7 mV (p = 0.005). SRR volumes showed better agreement with EAVM (Figure 2).
Figure 1

a. The 3D reconstruction and orthogonal planes from the SRR volume. b. The 3D reconstruction and orthogonal planes from the SA volume. Higher quality in SRR is achieved at z-x and z-y planes.

Figure 2

The 3D reconstruction of the myocardial scar, including the core (orange) and gray (yellow) zones, superimposed by the EAVM data. The EAVM colors indicate: healthy myocardium (pink), core scar (red), and gray zone (intermediate colors). a is from the SRR volume, and b is from the SA volume.


We proposed a method to reconstruct a high-resolution isotropic LGE MR volume from three routinely acquired anisotropic views. Improving the through-plane resolution of LGE leads to more accurate myocardial scar assessment with reduced PVE, especially at the clinically significant gray zone.


The work was financially supported by the Dutch Technology Foundation (STW) and EU ITEA2-09039.

Authors’ Affiliations

Department of Radiology, Leiden University Medical Center
Department of Cardiology, Leiden University Medical Center
Departments of Radiology and Medical Informatics, Erasmus Medical Center
Intelligent Systems Department, Delft University of Technology


© Tao et al.; licensee BioMed Central Ltd. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.