Lung perfusion analysis with sequential contrasted MRI is an important clinical tool. This work is part of larger research in which the objective of fusing lung perfusion analysis and lung anatomical structures. In this work, it is proposed a multimodality MR/CT lung fusion algorithm based on lung vessel determination that analyzes lung perfusion. 3D contrasted MR and 3D CT images are first normalized (slice thickness, resolution and pixel intensity), then lung features are determined and elastically registered. The registration algorithm was checked by mapping in both directions, from MR to CT and vice-versa. A MR perfusion analysis result is fused with a 3D CT segmented lung vessel tree.