Prognostic value of biventricular mechanical parameters assessed using cardiac magnetic resonance feature-tracking analysis to predict future cardiac events

Purpose: To study and compare the prognostic value of cardiac magnetic resonance feature tracking (MR-FT) of biventricular strain parameters with a conventional method. Materials and Methods: We retrospectively enrolled 364 patients undergoing clinically indicated cardiac MR examinations (1.5 or 3T scanner). Standard steady-state free precession (SSFP) images were used for analysis. Left ventricular (LV) and right ventricular (RV) ejection fraction (EF) were measured using conventional disk-area summation methods. Biventricular strain parameters were measured using MR-FT. All patients were followed to record major adverse cardiac events (MACEs). Results: The correlations between LV volumes and LVEF using both methods were excellent (r = 0.87–0.98). RV strain parameters were modestly correlated with RVEF (r = 0.44–0.63). During a median follow-up of 15 months, 36 patients developed MACEs. All MR-FT-derived parameters except for RV global longitudinal strain were significantly associated with future MACEs (P < 0.05) in univariate analysis. In stepwise Cox proportional hazard models, RV global radial strain (RVGRS) provided incremental prognostic value in models adjusted for age, gender, conventional LVEF (hazard ratio 0.93; P = 0.029) or RVEF (hazard ratio 0.93; P = 0.038). LV global transverse strain (LVGTS) also offered additional value over age, gender, conventional LVEF (hazard ratio 0.94; P = 0.041), or RVEF (hazard ratio 0.94; P = 0.004). Kaplan–Meier analysis showed significant survival differences in subgroups stratified by the median value of LVGTS, RVGRS, and LVEF using MR-FT (all log-rank P < 0.05). Conclusion: Deformation analysis of both ventricles using MR-FT provided significant prognostic power similar to parameters obtained using conventional methods. MR-FT is a promising alternative both for ventricular chamber quantification and for providing information of future cardiac events. Level of Evidence: 3. J. Magn. Reson. Imaging 2017;45:1034–1045.