What are these checks that you need to do?
My perhaps unknown hero of pdf processing is https://mupdf.com/
If it's in the data somewhere it should be detectable. They're an open source project, maybe bring your issue to them and they can make something?