Earlier quoted context omitted.
The only issue I see is, how could you tell which are those books?
We manage to do this for endangered wildlife too without anyone counting every single specimen; why shouldn’t we be able to estimate how rare a book is?
The same techniques roughly apply. You can look at all the book marketplaces and see how many of each title are available and it'll give you a rough estimate. If you look across all the marketplaces, all the library catalogs, plus eBay etc and zero copies surface, even going into Worthpoint to look at the last several years of eBay sales, then you know you have a problem. That's how I normally estimate it.
After that you are diving into forum posts etc to see if you can find anyone who has even mentioned owning a copy or having seen a copy.
I don't know what happens to some works. Supposedly thousands, tens of thousands, or sometimes apparently a million or more copies published and yet not a single copy surfaces for years.