There is no way to be sure if a pdf is corrupted/broken many actually are and PDF apps will do their own sub set of fixes to cope. The worst cases are a file is truncated for one of many reasons like incomplete download or trailer bad edit by annotation.
The way a PDF reader generally works is open bytes at start to check it starts %PDF- (40 bit signature) or similar with version e.g. %PDF-2.0 then most likely read the end of file length for %EOF (but there can be more than one of those) then generally work backwards looking for startxref (if it was not at the beginning of web pdf file) the xref table can be corrupted for many reasons and the reader tries to work through those decimal addresses looking for the catalog of pages (so index after index after index as font can be too) the file is tree like but adobe built it code named Carousel thus goes round and round.
A bad disk sector recovered file is a nightmare as there could easily be two file starts and no endings. For assurance in the days pre self healing smart drives there was crc32 or better MD5 for directory of files but clearly that just confirms the file went bad at some time same as attempting open with a bad result.
For basic recovery try different readers including Acrobat Reader and reprint as a new printout. However SumatraPDF printing converts to image format (not with searchable text fonts)