Comparison of strain parameters in dyssynchronous heart failure between speckle tracking echocardiography vendor systems

Background: Although mechanical dyssynchrony parameters derived by speckle tracking echocardiography (STE) may predict response to cardiac resynchronization therapy (CRT), comparability of parameters derived with different STE vendors is unknown.

Methods: In the MARC study, echocardiographic images of heart failure patients obtained before CRT implantation were prospectively analysed with vendor specific STE software (GE EchoPac and Philips QLAB) and vendor-independent software (TomTec 2DCPA). Response was defined as change in left ventricular (LV) end-systolic volume between examination before and six-months after CRT implantation. Basic longitudinal strain and mechanical dyssynchrony parameters (septal to lateral wall delay (SL-delay), septal systolic rebound stretch (SRSsept), and systolic stretch index (SSI)) were obtained from either separate septal and lateral walls, or total LV apical four chamber. Septal strain patterns were categorized in three types. The coefficient of variation and intra-class correlation coefficient (ICC) were analysed. Dyssynchrony parameters were associated with CRT response using univariate regression analysis and C-statistics.

Results: Two-hundred eleven patients were analysed. GE-cohort (n = 123): age 68 years (interquartile range (IQR): 61-73), 67% male, QRS-duration 177ms (IQR: 160-192), LV ejection fraction: 26 +/- 7%. Philips-cohort (n = 88): age 67 years (IQR: 59-74), 60% male, QRS-duration: 179 ms (IQR: 166-193), LV ejection fraction: 27 +/- 8. LV derived peak strain was comparable in the GE-(GE: -7.3 +/- 3.1%, TomTec: -6.4 +/- 2.8%, ICC: 0.723) and Philips-cohort (Philips: -7.7 +/- 2.7%, TomTec: -7.7 +/- 3.3%, ICC: 0.749). SL-delay showed low ICC values (GE vs. TomTec: 0.078 and Philips vs. TomTec: 0.025). ICC's of SRSsept and SSI were higher but only weak (GE vs. TomTec: SRSsept: 0.470, SSI: 0.467) (Philips vs. QLAB: SRSsept: 0.419, SSI: 0.421). Comparability of septal strain patterns was low (Cohen's kappa, GE vs. TomTec: 0.221 and Philips vs. TomTec: 0.279). Septal strain patterns, SRSsept and SSI were associated with changes in LV end-systolic volume for all vendors. SRSsept and SSI had relative varying C-statistic values (range: 0.530-0.705) and different cut-off values between vendors.

Conclusions: Although global longitudinal strain analysis showed fair comparability, assessment of dyssynchrony parameters was vendor specific and not applicable outside the context of the implemented platform. While the standardization taskforce took an important step for global peak strain, further standardization of STE is still warranted.