To actually manipulate the SAR RADAR data and project onto a sphere, you'll need to use ISIS or you can use the released data and figure small scale features won't be too distorted.
Most of the channel networks have been manually traced out from SAR RADAR data. I think there's been some recent efforts by some folks to use the RADAR data and automated processing with ARC-GIS with some specialized programming.
Check out the work by Devon Burr (mentioned in Emily's post above) and some other work by Ralf Jaumann and colleagues and some published traces done by Ralph Lorenz et al. (I used the manual method when tracing out and assigning the valley and channel networks in the Sikun Labyrinthus region, I initially used the publicly available image PIA08399)
A really nice description of the method is in a thesis done by Richard Cartwright "Analysis of Channel Networks and the Potential for Sediment Transport in the Vicinity of the North Polar Seas of Titan" freely available here:
http://digitalarchive.gsu.edu/geosciences_theses/20/The thesis compared some Titan north polar networks to terrestrial networks in Oman and in central Nevada using Google Earth imagery.