A framework for combining a motion atlas with non-motion information to learn clinically useful biomarkers: Application to cardiac resynchronisation therapy response prediction.

•Framework to combine cardiac motion and deformation data with non-motion data.•Motion atlas use for comparison of LV displacement, velocity, strain and strain rate.•Random projections and multiple kernel learning are used to extract relevant features.•Application to response prediction in cardiac resynchronisation therapy.•94% and 91% classification accuracy of super-responders and non-responders.