Article Text


005 Cardiac MRI: understanding myocardial motion to predict remodelling pre cardiac resynchronisation therapy
  1. S G Duckett1,
  2. W Shi2,
  3. X Zhuang2,
  4. A Shetty1,
  5. M Ginks3,
  6. C A Rinaldi3,
  7. G Carr-White3,
  8. D Rueckert2,
  9. R S Razavi1
  1. 1Kings College London, London, UK
  2. 2Imperial College London, London, UK
  3. 3Guy's and St Thomas' Hospital, London, UK


Introduction A significant number of patients undergoing Cardiac Resynchronisation Therapy (CRT) do not remodel. Assessing global dyssynchrony has the potential to improve patient selection. We developed a framework for comparing measures of myocardial motion from cardiac magnetic resonance (CMR) imaging and evaluated the potential of these techniques to improve patient selection.

Methods 48 patients recruited, (43 males, 63.8±13.9 years), NYHA class 2.9±0.5, ejection fraction 25±9%. Patients had LBBB (QRS 154±24 ms). Acute haemodynamic response was measured at time of implant with a pressure wire in the LV measuring change in dP/dtmax. A >10% increase in LV-dP/dtmax from baseline was considered an acute response. Decrease in end systolic volume (ESV) ≥15% at 6 months was used to determine remodelling. CMR was performed prior to CRT. A novel framework was developed. Key steps included: (1) detection of heart and myocardium segmentation from anatomical CMR cine images; (2) detection of endo and epi-cardial surfaces for wall thickening computation; (3) extraction of deformation fields within the myocardium for strain computation. A systolic dyssynchrony index (SDI) was produced for all parameters which included volume change, muscle thickening, radial, circumferential, longitudinal strain and combined strain. High SDI denoted dyssynchrony. Results Pre-implant ESV 175±64 ml, post-implant ESV 155±68 ml (p<0.01). 20 (44%) patients remodelled. We found a strong relationship between volume derived SDI and acute response (p=0.008) and remodelling (p<0.001) (Abstract 005 figure 1). We found a weaker relationship with remodelling and muscle thickening SDI (p=0.001) and no relationship with a SDI derived from strain indexes (Abstract 005 figure 2). Volume SDI ≥10% was highly sensitive (0.94) and specific (0.87) for predicting remodelling. Volume SDI was far superior for predicting remodelling than any other method. The intra-observer average difference for volume SDI was 0.04±0.3% and COV was 1.8±1.2% and the inter-observer average difference was 0.55±1.4% and COV was 4.2±4.6%.

Abstract 005 Figure 1

Shows the ANOVA plots for acute response and remodelling for QRS duration, volume and muscle thickening derived SDI.

Abstract 005 Figure 2

Shows the ANOVA plots for acute response (top row) and remodelling (bottom row) for SDI derived from various types of strain.

Conclusion A volume derived SDI from cine CMR strongly predicts remodelling post CRT. It is a highly reproducible measurement that has significant potential clinical implications in the future.

  • Cardiac resynchronisation therapy
  • dyssynchrony
  • remodelling

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.