This article should be posted as a response/comment to Adamn's statement, even though there seems to be no room in those publications for facts.
To your point about Choi, yes. It would seem that # of cells, which would involve both size and density, would be the more valid measure. If a tumor grew and became less dense, or vise versa, the cell count could remain the same. NB: I am not a doctor, though I played doctor as a kid a few times.