Layer-Specific Segmental Longitudinal Strain Measurements: Capability of Detecting Myocardial Scar and Differences in Feasibility, Accuracy, and Reproducibility, Among Four Vendors A Report From the EACVI-ASE Strain Standardization Task Force

BACKGROUND: Segmental longitudinal strain (SLS) is reported to be vendor specific. Despite standardization efforts, vendors still use different myocardial layers for strain measurements. It is unclear, however, which layer is the most favorable for clinical purposes. Therefore, in this study we evaluated the reproducibility, accuracy, and scar detection ability of SLS measurements from different myocardial layers. METHODS: In data sets of 58 patients with prior myocardial infarction and five healthy volunteers, we measured the intervendor bias, the relative test-retest variability, and scar discrimination ability of endocardial and midwall SLS, using software packages from four different companies (GE, Siemens, Toshiba, and TomTec). Cardiac magnetic resonance delayed enhancement images were used as the reference standard of scar definition. RESULTS: Variability of SLS measurements was significant among the vendors for both midwall and endocardium. In addition, relative errors of SLS measurements varied considerably among vendors (P < .001 for both layers). Comparisons of test-retest errors from different layers for individual vendors did not show any significant differences. Regardless of the vendor, both endocardial and midwall strain values were decreased in scarred segments. Endocardial to midwall ratio of strain measurements showed no difference between scar-free and scarred segments. Endocardial and midwall strain parameters showed no significant difference in scar detection capability. CONCLUSIONS: Layer-specific SLS measurements vary significantly among vendors. Endocardial and midwall SLS measurements have a high yet comparable test-retest variability. Combining layer-specific SLS measurements does not provide additional information for detection of regional functional abnormalities. Our results do not provide evidence to favor the use of one myocardial layer over another