Short-breath hold cine DENSE
Andrew David Scott1,2, Upasana Tayal1,2, Sonia Nielles-Vallespin1,3, Pedro Ferreira1,2, Xiaodong Zhong4, Frederick Epstein5, Sanjay Prasad1,2, and David Firmin1,2

1NIHR funded Cardiovascular Biomedical Research Unit, Royal Brompton Hospital, London, United Kingdom, 2National Heart and Lung Institute, Imperial College London, London, United Kingdom, 3National Heart Lung and Blood Institute, National Institutes of Health, Bethesda, MD, United States, 4MR R&D Collaborations, Siemens Healthcare, Atlanta, GA, United States, 5University of Virginia, Department of Biomedical Engineering, Charlottesville, VA, United States


Displacement encoding with stimulated echoes (DENSE) can provide valuable strain information, but acquisitions are typically too long for patient cohorts who have difficulty breath holding. In this work we accelerate 2D cine spiral DENSE acquisitions by selectively exciting a small field of view around the heart. We compare strain data derived from DENSE acquired with unaccelerated and up to 2.5x acceleration in a cohort of healthy subjects and show minimal differences when the acquisition is accelerated. We also show an example from a patient with a myocardial infarction where the accelerated DENSE data shows abnormal strain in the infarcted regions.


Displacement encoding with stimulated echoes (DENSE) is an accurate and reproducible1,2 technique for measuring myocardial strain throughout the cardiac cycle. The strain maps provided could be valuable in dilated cardiomyopathy (DCM) patients where strain has a prognostic value3. However, DENSE acquisitions typically require long breath holds which are difficult for DCM patients or navigator gating which is time consuming and prone to failure. In this work we accelerate DENSE acquisitions by selectively exciting a volume of tissue around the heart. This allows fewer spiral interleaves to be acquired without aliasing and, therefore a shorter breath hold. We compare accelerated and unaccelerated DENSE in healthy volunteers and demonstrate the utility of accelerated DENSE in an example patient.


A cine spiral DENSE sequence4 was modified to selectively excite a reduced field of view. A slice selective gradient was added to the first and second RF pulses on the read and phase axes respectively (figure 1). An improved excitation profile without increasing TE was achieved by making the 1st pulse asymmetric (peak at 81% of duration) and the 2nd pulse a time-reversed copy of the first.

In-vivo 2D cine DENSE was performed in 8 normal subjects (Siemens Skyra 3T) in a mid short-axis slice. Images were acquired with variable flip angle (20o max), TE=1ms, fat suppression, 3.5x3.5mm2 spatial resolution, 1282 reconstruction matrix, 8mm slice thickness, TR=15ms, 30ms temporal resolution, 2 spirals/frame, 6ms spiral duration, 2 direction simple encoding (+reference) at 0.06cycles/mm, CSPAMM and through-plane dephasing artifact suppression5 (0.08cycles/mm). Long, medium and short acquisitions were performed with square fields of view/breath hold durations of 360mm2/20RR-intervals (RR), 224mm2/14RR and 120mm2/8RR, respectively, including 2RR-intervals for a B0 field map in each case. The long acquisition used a similar field of view to previous work performed without the zonal excitation and was used as a reference4. Images were processed using the DENSE analysis tool from the University of Virginia6. Peak and time-to-peak radial and circumferential strains measured globally over the left ventricle were compared between protocols using a Wilcoxon signed rank test (p<0.05 threshold) and by calculating the root mean square error between the strain-time curves.

Mid-ventricular short axis spiral cine DENSE using the medium breath hold protocol was also performed in one patient diagnosed with dilated cardiomyopathy.


Figure 2 shows example magnitude images in one example volunteer acquired using long, medium and short acquisitions and the corresponding strain curves are shown in figure 3. There is good agreement between the strain curves from all three acquisitions. Global peak strain, time to peak strain and the differences between acquisitions (long acquisition as reference, expressed as bias and root mean square error) are summarized in the table for radial and circumferential strain for all 8 volunteers. While there is a significant under-estimation (p<0.01) of peak radial strain using the medium acquisition, the magnitude of the difference is small (0.01). There were no other significant differences and the medium and short breath hold acquisitions appear to be equally accurate.

Upon review of the patient data, a wall motion abnormality and late gadolinium enhancement (LGE) indicated that the diagnosis was incorrect and the patient had an extensive infarct. Figure 4 shows the strain curves derived from DENSE with the corresponding LGE images. Radial and circumferential strain is reduced with abnormal strain patterns in the infarcted regions of myocardium, while the remote regions show relatively normal strains.

Discussion and conclusion

Spiral cine DENSE imaging can be accelerated by up to a factor of 2.5 by selectively exciting and imaging a small field of view around the heart. The associated loss of signal to noise ratio is partly compensated for by imaging at 3T. This improvement allows 2D acquisitions to be performed in a medium (~14s) or even short breath hold (~8s) which most patients could sustain. The technique was demonstrated in a patient with myocardial infarction. Segments of the heart with reduced strain were found to correspond to regions of the heart with late gadolinium enhancement. In future the same approach may be used to accelerate 2D long axis and 3D spiral cine DENSE acquisitions.


This work was performed at the National Institute for Health Research funded Cardiovascular Biomedical Research Unit at the Royal Brompton Hospital and Imperial College London.


1. Kim D, Gilson WD, Kramer CM, Epstein FH. Myocardial Tissue Tracking with Two-dimensional Cine Displacement-encoded MR Imaging: Development and Initial Evaluation. Radiology 2004;230(3)862-871.
2. Haggerty CM, Kramer SP, Binkley CM et al. Reproducibility of cine displacement encoding with stimulated echoes (DENSE) cardiovascular magnetic resonance for measuring left ventricular strains, torsion, and synchrony in mice. J Cardiovasc. Magn. Res. 2013;15(71).
3. Buss SJ, Breuninger K, Lehrke S et al. Assessment of myocardial deformation with cardiac magnetic resonance strain imaging improves risk stratification in patients with dilated cardiomyopathy. Eur Heart J Cardiovasc Imaging. 2015;16(3):307-315.
4. Zhong X, Spottiswoode BS, Meyer CH et al. Imaging three-dimensional myocardial mechanics using navigator-gated volumetric spiral cine DENSE MRI. Magn Reson Med. 2010;64(4):1089-1097.
5. Zhong X, Spottiswoode BS, Cowart EA et al. Selective suppression of artifact-generating echoes in cine DENSE using through-plane dephasing. Magn Reson Med 2006;56:1126-1131.
6. Spottiswoode BS, Zhong X, Hess AT et al. Tracking myocardial motion from cine DENSE images using spatiotemporal phase unwrapping and temporal fitting. IEEE Trans Med Imaging. 2007;26(1):15-30.


Figure 1: Sequence schematic for the cine spiral DENSE sequence with zonal excitation.

Figure 2: Example magnitude images acquired in one example healthy subject using each of the 3 DENSE acquisitions. The line styles reference the plots shown in figure 3.

Figure 3: Strain curves corresponding to the images shown in figure 2. Data from the long, medium and short acquisitions are shown as solid, dashed and dotted lines respectively. The coloured lines show the mean strain values from 6 equal angle segments and the black lines show the global short axis mean values. There is a good agreement between all three acquisitions.

Figure 4: Spiral cine DENSE in a patient with a myocardial infarction acquired using the medium breath hold protocol. Radial and circumferential strain (A and B) is reduced in segments where there is late gadolinium enhancement (LGE) (D). The location of the short axis LGE image and the DENSE images is shown by the green lines on the long axis LGE images. The corresponding myocardial segments for the plots shown in A and B are coloured accordingly on the magnitude DENSE image shown in C.

Table: Global strain results.
Bias: medium or short – long.
RMSE: Root mean square error between the global strain curves.

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