To my way of thinking, the best way to handle this is to simply attempt to delete them all.   It’s wasteful to do one directory-search to see if the file exists, only to turn around and (oblige the target computer to) do another one in order to delete them.

If you have reason to believe that most of the filenames do not exist, and that the number of files which actually do exist is small enough to warrant a scan, then obviously you could request a list of files and cherry-pick from that.   It really depends on what you consider the realistic probability of “file not found” to be.