A file can be "summarized" into a short string. That string can be compared with the file again to determine if it's the same file. This is often used when downloading a file to check it after the fact. This sort of thing is built into compression software, but using md5sum is way faster/easier because it doesn't need to put the file in a container and uncompress it to use it.
There are various other programs which do similar things, like
sha256sum. These all come with almost all Linux distributions.
I prefer cfv - Tests/creates .sfv, .csv, .crc, .md5(sfv-like), md5sum, bsd md5, sha1sum, and .torrent files.
Checking a file, using a filename ∞
Is your ISO ok? If you don't know, do this:
md5sum -c $MD5_FILENAME