A left bundle-branch block (LBBB) contraction pattern identified from longitudinal-strain analysis predicts outcomes following cardiac resynchronization therapy (CRT). We investigated the use of an LBBB-contraction pattern identified from radial- or circumferential-strain analysis in the prediction of CRT benefits. Eighty CRT candidates were prospectively enrolled. Before CRT implantation, speckle-tracking analyses in three deformation directions were performed to determine whether an LBBB-contraction pattern was present. The study endpoints were reverse remodeling at 6 months, and adverse outcomes including death or heart failure hospitalization. At 6 months, 49 (61%) patients had reverse remodeling. An LBBB-contraction pattern identified from the radial strain in the mid-ventricular short-axis view or longitudinal strain in the four-chamber view provided excellent true-positive (86%) and false-negative (8%) rates for predicting reverse remodeling. During a median follow-up of 30 months, 31 patients (39%) had adverse outcomes. Absence of an LBBB-contraction pattern in radial (hazard ratio 3.74; 95% confidence interval 1.83–7.62) or longitudinal strain (hazard ratio 3.49; 95% confidence interval 1.71–7.13) was significantly associated with an increased risk of adverse outcomes. Adding the LBBB-pattern assessment by radial-(model χ² from 8.2 to 18.5, p = 0.005), or longitudinal-strain analysis (model χ² from 8.2 to 16.9, p = 0.011) to a risk model significantly improved the model, including QRS duration and ischemic etiology. In conclusion, an LBBB-contraction pattern identified from radial-strain analysis in the mid-ventricular short-axis view predicted reverse remodeling and outcome following CRT, similarly to the longitudinal-strain analysis.