No, I don’t have a specific pipeline I could recommend, but as a general suggestion, you should probably try to register a representative frame from the DSC sequence to the T1w image. You should find a frame in the temporal sequence that has contrast that allows you to see similar structures as in T2w, and use that for registration. You can use MultiVolumeExplorer module to examine the temporal sequence and extract selected frame into a scalar volume that you should use for registration.
For registration, there are many options. Take a look at this FAQ: Documentation/Nightly/FAQ/Registration - Slicer Wiki. You can start with “General registration”, which wraps the BRAINSFit tool, or you can use the SlicerElastix extension, which has a lot of presets that might work for you (see Elastix registration toolbox is now available in Slicer).
That module performs resampling of the image, not registration (i.e., you can sample the values from one image into a voxel lattice of a different geometry).
I do not have any knowledge about how Slicer registration tools compare with FSL FLIRT. If I were you, I would try all tools that are available before deciding on the optimal workflow, unless the first tool you try gives you a satisfactory result.