Maybe your first task is to check if that "not exist" files already exist on
the website.

If they do, you could have a try with --near option (-n). As stated in the
help this option should "get non-html files 'near' an html file", that is,
those external files directly linked in an html file (as your desired pdf
files could be).

Just in case, use this filter: "+*.pdf" (will download any file with this
extension)

And last but not least, it seems your mirror is not complete: maybe it has
been interrupted for one reason or another (there are local links, but the
linked file is not there). If you haven't cancelled it, check your logs
(hts-log.txt) to look for any error.

Hope this helps