I will start by saying that I know very little about Oracle. However, I would say the answer is probably no, especially for more recent versions of AtoM.
As you know, the PHP framework used to build AtoM (Symfony 1) is very outdated - and it included an ORM
(Propel 1) for abstracting the interactions with the database that, as AtoM has grown, has become equally problematic and too slow to scale as needed. To address some performance issues in later versions of AtoM, we have in places bypassed the ORM entirely, and used raw SQL in the code to access the database directly.
Since query languages often vary between different databases (as do database default behaviors, collation formats, etc) I suspect that Oracle would not work out of the box. In fact, with the 2.6 release, we are seeing some cases of community-maintained MySQL forks (such as MariaDB) also not working, because of the changes in MySQL8 that are not yet supported in MariaDB's latest versions, or where the MariaDB defaults differ from those in MySQL. I would guess the same to be true for Oracle here.
Long term (AtoM 3 and on), we can hopefully return to a level of abstraction where it's possible to swap in different data stores. For AtoM2 as it currently exists, I believe that the database and the code have become too tightly coupled for substitutions at this point, without analysis and significant code changes.