Disclosed is a method of reconstructing a multi-dimensional data set representing a dynamic sample at a series of reconstruction instants. The multi-dimensional data set comprises a static component and a dynamic component. The method comprises acquiring a plurality of projection images of the dynamic sample; reconstructing a static component of the multi-dimensional data set from the acquired projection images; acquiring a further plurality of projection images of the dynamic sample; and reconstructing a dynamic component of the multi-dimensional data set at each reconstruction instant from the further plurality of acquired projection images using a priori information about the sample. The multi-dimensional data set is the sum of the static component and the dynamic component at each reconstruction instant.