The influence of the region of interest width on two-dimensional speckle tracking-based measurements of strain and strain rate.

AIMS: There are significant variations in the published normal values of two-dimensional speckle tracking-derived strain and strain rate. These occur even when authors use the same software. To measure strain, the operator creates a region of interest (ROI) to define the myocardium to be analyzed. The purpose of this study was to test the hypothesis that measurements vary significantly with the chosen ROI width. METHODS AND
RESULTS: In 20 healthy subjects (11 males, mean age 17.6 ± 6.18 years) an apical four-chamber view (4CH) and parasternal short-axis view (SAX) were analyzed. Initially ROI width was set automatically by the software. Two subsequent measurements were obtained from each cine loop by choosing the ROI width one step narrower and one step wider than the automatic ROI width. The mean differences between the measurements of narrower and automatic ROI and between automatic and wider ROI were -1.8 ± 0.7% and -0.9 ± 0.5% for global longitudinal strain (SL), -2.2 ± 0.6% and -1.7 ± 0.7% for global circumferential strain (SC), -0.10 ± 0.06/sec and -0.07 ± 0.06/sec for global longitudinal strain rate (SrL), and -0.15 ± 0.09/sec and -0.12 ± 0.07/sec for global circumferential strain rate (SrC) (all P < 0.000). This corresponds to a relative difference to the mean of both measurements of -4.4 to -11.0%.
CONCLUSION: Layer-specific myocardial deformation and curvature dependency lead to an inverse correlation between the chosen ROI width and strain and strain rate measurements. Just one step of ROI-width change leads to a significant bias. Precise ROI-width definition is essential but technical factors limit its feasibility.