I feel like a definition of isolated test is needed at the top. I see test isolation as meaning one of two things.
1. We isolate our tests so that the order tests are run doesn't matter, tests don't depend on each other at all, and ultimately they are deterministic
2. We isolate tests so that they aren't traversing code pathways down a chain of dependencies. i.e. we should be using fakes for the dependencies, so we can test just the code we're testing now.
It seems you are using definition #2 here.