Systems and methods are provided for generating a model relating parameters generated via a first molecular imaging modality to parameters generated via a second molecular imaging modality. First and second feature extractors extract, from images of a region of interest obtained via respective first and second molecular imaging modalities, respective sets of parameters for respective first and second sets of locations. A mapping component associates respective locations of the first and second sets of locations according to their spatial relationship within the region of interest to produce a training set. Each example in the training set comprises a set of parameters associated with a location in the first set of locations and a set of parameters associated with a location in the second set. A modeling component generates a predictive model relating the parameters associated with the first modality with at least one parameter associated with the second modality.
展开▼