Three-dimensional mapping of mechanical activation patterns, contractile dyssynchrony and dyscoordination by two-dimensional strain echocardiography: rationale and design of a novel software toolbox

Abstract

Background: Dyssynchrony of myocardial deformation is usually described in terms of variability only (e.g. standard deviations SD's). A description in terms of the spatio-temporal distribution pattern (vector-analysis) of dyssynchrony or by indices estimating its impact by expressing dyscoordination of shortening in relation to the global ventricular shortening may be preferential. Strain echocardiography by speckle tracking is a new non-invasive, albeit 2-D imaging modality to study myocardial deformation.

Methods: A post-processing toolbox was designed to incorporate local, speckle tracking-derived deformation data into a 36 segment 3-D model of the left ventricle. Global left ventricular shortening, standard deviations and vectors of timing of shortening were calculated. The impact of dyssynchrony was estimated by comparing the end-systolic values with either early peak values only (early shortening reserve ESR) or with all peak values (virtual shortening reserve VSR), and by the internal strain fraction (ISF) expressing dyscoordination as the fraction of deformation lost internally due to simultaneous shortening and stretching. These dyssynchrony parameters were compared in 8 volunteers (NL), 8 patients with Wolff-Parkinson-White syndrome (WPW), and 7 patients before (LBBB) and after cardiac resynchronization therapy (CRT).

Results: Dyssynchrony indices merely based on variability failed to detect differences between WPW and NL and failed to demonstrate the effect of CRT. Only the 3-D vector of onset of shortening could distinguish WPW from NL, while at peak shortening and by VSR, ESR and ISF no differences were found. All tested dyssynchrony parameters yielded higher values in LBBB compared to both NL and WPW. CRT reduced the spatial divergence of shortening (both vector magnitude and direction), and improved global ventricular shortening along with reductions in ESR and dyscoordination of shortening expressed by ISF.

Conclusion: Incorporation of local 2-D echocardiographic deformation data into a 3-D model by dedicated software allows a comprehensive analysis of spatio-temporal distribution patterns of myocardial dyssynchrony, of the global left ventricular deformation and of newer indices that may better reflect myocardial dyscoordination and/or impaired ventricular contractile efficiency. The potential value of such an analysis is highlighted in two dyssynchronous pathologies that impose particular challenges to deformation imaging.

Figure 1

Shortening and stretching patterns in a patient with LBBB (left) and a normal individual (right) represented by a series of colour-coded bulls-eyes representing deformation-rate at 25 time points throughout the entire cardiac cycle for each of the 36 segments. Yellow = shortening, blue = stretching, green = no deformation/diastasis. AS = anteroseptum, ANT = anterior, LAT = lateral, POST = posterior, INF = inferior, SEPT = septum. In the normal individual (right), throughout most of the cardiac cycle all segments deform in phase (indicated by the same colour in all segments per frame). In LBBB (left) at many time points simultaneously occurring back-and-forth shortening (yellow) and stretching (blue) between septal regions and ventricular free wall can be appreciated.

Figure 2

Shortening and stretching patterns in a patient with LBBB (left) and a normal individual (right) represented by a two-dimensional M-mode map representation of the same data as in figure 1: the temporal information can now be continuously plotted over time (left to right within each plot). Vertical lines represent event timing markers. Spatial representation is less optimal: each of the 6 walls is plotted separately (from top to bottom) with each level plotted from base (top) to apex (bottom) within the separate plots. As in figure 1, shortening and stretching are markedly inhomogeneous in LBBB (left) compared to NL (right).

Figure 3

Global strain plot in a patient with LBBB (left) and in a normal individual. All 36 separate tracings are displayed as well as the average shortening curve representing global ventricular longitudinal shortening (thick yellow curve). Timing event markers are indicated in vertical dashed lines with MVC: mitral valve closure, AVO: aortic valve opening, AVC: aortic valve closure, MVO: mitral valve opening, and AWO: onset of mitral flow A-wave. Note the highly variable timing, shape and amplitude of the segmental deformation curves in this patient with LBBB (left) compared to the normal pattern (right).

Figure 4

ISF-plot with timing markers in a patient before (A) and after (B) CRT. To calculate ISF (in this article between AVO and AVC), all shortening (in blue) and all lengthening strain-rates (in yellow) are summed and integrated over the desired time period. Overlapping areas (light green) indicate simultaneous shortening and stretching between different segments (dyssynergy). A/Note large areas of overlap between AVO and AVC in this patient with LBBB, resulting in a high ISF. B/Immediately after CRT, ISF has decreased (less overlap) because little lengthening occurs during ejection and the amount of shortening (blue area) has increased. This suggests conversion of internal strain into external (global) strain.

Figure 5

Rationale for the use of ESR. Representative deformation traces of the septal (green) and lateral wall (red) and global ventricular deformation (light grey) obtained by MR-tagging before (A) and 8 weeks after the induction of left bundle branch block (B) in a dog with on Y-axis shortening in % (data from reference 14). AVC = time of aortic valve closure; # and dashed line from AVC to Y-axis denote the end-systolic value, * denotes the peak value of deformation. From A to B: LBBB induces a marked reduction in septal strain peak amplitude (green *) and in particular in the end-systolic value (green #). The peak deformation amplitude of the lateral wall (red *) occurs after AVC and has increased (-11.0% to -13%) but the end-systolic value has changed less. This means that a hypothetically perfect resynchronization (backwards from B to A) would consist of an relative increase in septal deformation with little change in the lateral contribution in this period. Hence, to estimate how much function can improve by resynchronization, ESR only takes differences between peak and end-systolic values of early shortening into account (i.c.green* and green #).

Figure 6

Colour-coded bulls-eye projections of mechanical activation times (top row; A to D) and of timing of peak shortening (bottom row; A₁-D₁), with the vector magnitude and direction plotted in the centre (small blue arrows). The thick black line on the left of each bulls-eye defines the attachment of the inferior right ventricular wall between the inferior and the septal wall (wall segmentation: see figure 1). For clarity, all plots have an equal scale: 0 ms to 180 ms for mechanical activation times (top row) and 100 to 550 ms for peak times (bottom row). Example A and A₁: normal volunteer. Examples B, B₁ and C, C₁: patient with an inferoseptal and with an anterolateral bypass, respectively (* = invasively determined bypass localization). D: patient with LBBB. Note the large onset delay vector (D) and even larger peak shortening vector (D₁) pointing from the septum to the lateral wall.

Figure 7

Comparison of the ISF-plot obtained in a normal volunteer (A) and in a patient with WPW (B). Note the increased area of overlap (light green areas) around the isovolumic periods (between MVC-AVO and between AVC-MVO) and in particular the onset of vigourous shortening (onset of high blue spike, red arrow) starting during the end of atrial filling, long before MVC (red dashed line).