Inconsistency of 3D strain measurements between vendors

Background: Three-dimensional (3D) speckle-tracking echocardiography provides a comprehensive analysis of left ventricular (LV) myocardial deformation, overcoming the "out-of-plane" flaw of two-dimensional (2D) method. Serious concerns have been raised about the consistency of 2D strain values between vendors. Whether discrepancies may also apply for 3D strain parameters is currently unknown. Aim: To compare two commercially available systems and related softwares for 3D strain quantification. Methods: In 60 consecutive patients (pts, age 58±15, 68% men, LV ejection fraction 17-65%) in sinus rhythm, two LV full-volume data sets were acquired using Vivid E9 with 4V probe (GE Healthcare, Horten, NO) and Artida with PST-25SX probe (Toshiba Medical Systems, Tokyo, JP) immediately one after another 11 pts (15%) were excluded due to poor acoustic window or stitching artifacts. Peak global longitudinal, radial, circumferential and area strain were analyzed using the two vendor-specific softwares: 4D strain EchoPac BT 11 and 3D WMT,respectively. Results: Average time for analysis was similar for both vendors (3:58 vs 3:44min, p=NS). In 2 pts (3%), 4 segments/pt with inadequate tracking were rejected by GE software and global strain values were not provided. Strain results of the remaining 58 pts are shown in Table 1. Except for longitudinal strain, the two tested softwares yielded significantly different values for all parameters, most notably for radial strain. Area strain showed the closest correlation between vendors. Conclusions: Global longitudinal 3D strain seemed reasonably comparable between the two vendors, while all the other strain components were significantly different. The largest discrepancy was found for radial strain values. Our study highlights the critical importance of using the same system to perform follow-up in longitudinal studies and, unless a uniform standard will be developed, of using vendor-specific reference values.