Abstract 16096: The Relationship between Global Longitudinal Strain from Novel Feature Tracking MRI Strain and Echo Strain: Comparison against Ejection Fraction and Scar

Abstract

Background: Global longitudinal strain (GLS) has prognostic value in pts with suspected LV impairment. MRI strain has traditionally been assessed by tagging, but recent feature tracking software has allowed use of cine MRI in DICOM format. We sought whether this approach allows quantification of cardiac function similar to echo.

Methods: After exclusion of left bundle branch block, post-valvular surgery, or moderate or greater valvular heart disease, we enrolled 67 consecutive pts (63±10yrs, 55 men, EF 24±10%) who had cardiac MRI (for viability assessment) and echocardiography (separation 1±4 day). MRI strain by feature tracking was measured using 2D cardiac performance analysis MR software (Tomtec, Germany) and 2D speckle echo strain with Velocity Vector Imaging (Siemens, Germany). GLSs (obtained from averaging longitudinal strains of apical 4-, 2- and 3-chamber views) were compared by Echo and MRI as well as with LVEF by MRI and scar (%) and peri-infarct area (%) by MRI late Gadolinium enhancement.

Results: Scar and peri-infarct area were 28.5±17.1% and 5.4±3.6%. The feasibilities of MRI and Echo strain were 90% and 87%, respectively. The correlation between MRI and Echo GLS was moderate (r=0.69, p<0.001), with small mean difference (0.11%) but wide limits of agreement (5.0%). The correlation between Echo GLS or MRI GLS and EF were modest (r=-0.66, p<0.001 and r=-0.58, p<0.001) and not significantly different (p=0.177). Scar burden were not associated with Echo GLS (r= 0.04, p=0.75); MRI GLS (r=0.22, p=0.09) or EF (r=0.03, p=0.82). Peri-infarct area was also unrelated to the respective functional parameters (r= 0.03, p=0.83; r=0.13, p=0.31; and r=-0.001, p=0.99)

Conclusion: Feature tracking MRI strain is feasible and correlates modestly with Echo-derived strain but not scar extent.