A method for multi-modality registration using virtual cursors including receiving a two-dimensional image dataset for an object at a first position and receiving a three-dimensional image dataset for the object at the first position. The three-dimensional image dataset includes a plurality of image slices. The two-dimensional image dataset is registered with the three-dimensional image dataset without taking into account a magnification factor. A user cursor position for a location in the two-dimensional image dataset is received. A slice of interest in the three-dimensional image dataset is received. The slice of interest is selected from the plurality of image slices in the three-dimensional image dataset. A shadow cursor position for a location in the three dimensional dataset is calculated. The shadow cursor position corresponds to the user cursor position and the calculating includes a correction for the magnification factor corresponding to the shadow cursor position for the slice of interest. The shadow cursor position is output.