Interesting subject, I think. The "pouring gasoline on a fire" theory is certainly the conventional wisdom and it is a "safe" position for a doctor to take (his malpractice insurance company would certainly want him to follow it; it's hard to get sued for sticking to the conventional wisdom.) What'e more, most urologists have quite a bit of anecdotal clinical experience that tends to support it.
... but ...
On the other hand, when the assumption is subjected to the rigors of controlled research it proves, again and again, to be hard to find. Here are a few full-text reviews of recent research:
The benefits and risks of testosterone replacement therapy: a reviewHypogonadism in the Aging Male Diagnosis, Potential Benefits, and Risks of Testosterone Replacement TherapyAnd here is a link to an interesting article by Dr. Morgantaler in Life Extension magazine:
Destroying the Myth about Testosterone Replacement and Prostate Cancer (By Abraham Morgentaler, MD) Life Extension MagazineJNF:
I was writing this when you posted. I have read quite a bit of the research (see above). I don't think there are any urologists who prescribe T for patients with active metastatic advanced non-castrate-resistant prostate cancer.
The controversial idea is that there are two levels of T that make sense: castrate level (which halts or delays the growth of prostate cancer) and normal (which avoids the effects of hypogonadism). If you aren't going to shoot for castrate levels of T then you might as well adjust it to normal since the difference between low T and normal T, in terms of their effect on PCa, is quite small. No one denies that castrate-level T is an effective treatment when it is needed.
The reason that the T manufacturers say that men at risk for PC shouldn't use their product is not because they have done research that shows a danger, it's because it is not worth their money to do the research to try to show that it is safe and they don't want to take the liability risk.
In my metaphor the crumbs of my single pie are enough to feed all the hypothetical mice you plan to throw at me. I will merely hypothesize a bigger, crumbier pie. That's my point, actually. Once you feed the mice more crumbs don't make a difference.
Edit: Fixed the link to the Life Extension Mag article.