I would suggest that retrospective studies can uncover a signal that can be examined in a subsequent prospective study. I think corroborate is too strong a word here.