Skip to main content

Advertisement

  • Poster presentation
  • Open Access

Principle component analysis of myocardial strain to quantify left ventricular dyssynchrony

  • 1,
  • 1,
  • 1,
  • 2 and
  • 1
Journal of Cardiovascular Magnetic Resonance201315 (Suppl 1) :P74

https://doi.org/10.1186/1532-429X-15-S1-P74

  • Published:

Keywords

  • Cardiac Resynchronization Therapy
  • Mechanical Dyssynchrony
  • Myocardial Strain
  • Left Ventricular Dyssynchrony
  • Left Ventricular Segment

Background

Cardiac resynchronization therapy (CRT) is effective for selected heart failure (HF) patients, but is associated with a 30-40% nonresponse rate. Identification of CRT responders may be improved with myocardial strain imaging. The circumferential uniformity ratio estimate (CURE )1 measures mechanical dyssynchrony by Fourier series fitting of myocardial strains over space, but requires user interaction to define a range of cardiac phases over which CURE is calculated (time dependence). We hypothesize that principal component analysis (PCA) can quantify dyssynchrony in myocardial strain in a data-driven, time-independent manner that does not require any subjective user assessments of strain data.

Methods

Dyssynchronous HF was induced in canines (N=5) with tachycardia pacing and left bundle branch ablation (LBBB-HF), while synchronous HF with narrow QRS (NQRS-HF) was induced in canines with tachycardia pacing only (N=5). Four normal canines were also studied.

Spiral cine DENSE MRI was performed on a 1.5 T scanner (Avanto, Siemens) in all 14 canines. 2D myocardial motion was tracked in a mid-ventricular short-axis plane from DENSE images acquired using the following parameters2: interleaves=6, TR=17 ms, TE=1.9 ms, slice thickness=8 mm, excitation flip angle=15°, in-plane resolution=2.8 x2.8 mm and displacement-encoding frequency= 0.1 cycles/mm.

DENSE images were analyzed3 to calculate left ventricular (LV) circumferential strain (Ecc), and PCA was applied to the Ecc-time curves for a 24-segment LV model. Specifically, the LV Ecc curves were decomposed spatially into principal component basis vectors. The PCA-based metric for measuring LV dyssynchrony, termed First Order Regional Conformity Estimate (FORCE), was calculated as |sum(PCL1)|/sum(|PCL1|) where PCL1 represents the loadings of the first principal component basis vector. Both FORCE and CURE range from 0 (dyssynchrony) to 1 (perfect synchrony). The Kruskal-Wallis one-way ANOVA test was used to compare FORCE and CURE among the three groups in pairwise fashion.

Results

Figure 1 shows the spatial distribution of PCL1 for the Ecc of example LBBB-HF and NQRS-HF canines. The PCL1 of LV Ecc in LBBB-HF canines varied widely over LV segments, whereas PCL1 showed little variation over LV segments in NQRS-HF and normal canines. As shown in Figure 2, FORCE and CURE were both markedly different in LBBB-HF versus NQRS-HF (p<0.05) and LBBB-HF versus normal (p<0.05). Also, FORCE was significantly greater than CURE for NQRS-HF (p<0.05), indicating even better identification of synchrony than CURE.
Figure 1
Figure 1

Bulls-eye plots of PCL1 as a function of LV segments for I) a LBBB-HF and II) a NQRS-HF canine; A-Anterior, AL-Anterolateral, IL-Inferolateral, I-Inferior, IS-Inferoseptal, AS-Anteroseptal

Figure 2
Figure 2

Boxplot comparisons of I) CURE and II) FORCE for LBBB-HF,NQRS-HF and Normal dogs

Conclusions

PCA, using the promising new FORCE parameter, effectively and automatically identifies mechanical dyssynchrony in HF in a data-driven and completely time-independent fashion. Further clinical evaluation of FORCE for prediction of CRT response is warranted.

Funding

This work was funded by the American Heart Association 12GRNT12050301 and NIH Heart, Lung and Blood Institute (NHLBI) T32HL007284.

Authors’ Affiliations

(1)
Biomedical Engineering, University of Virginia, Charlottesville, VA, USA
(2)
Department of Medicine, University of Virginia, Charlottesville, VA, USA

References

  1. Bilchick, et al: . JACC Imaging. 2008Google Scholar
  2. Budge, et al: . JACC Imaging. 2012Google Scholar
  3. Spottiswoode, et al: . IEEE Trans Med Imaging. 2007Google Scholar

Copyright

© Ramachandran et al; licensee BioMed Central Ltd. 2013

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.

Advertisement