This is an interesting problem, very crisply stated. Variants of it occur in other life contexts as well. Thus truthful reporting of information can be counterproductive, for example, in the work of Human Rights Watch or Amnesty International in situations where such information may undermine the reputation of a political party or faction that, all things considered, is better than its rivals. This case is different from yours in that the responsibility for being truthful is primarily to third parties: to those who rely on the NGO to report the whole truth and nothing but the truth about the various governments, parties, or factions and their conduct.
In your case, where the responsibility is basically to the patient, the practice you and your colleagues have developed seems entirely right to me. Maybe your slight doubts about this practice can be mitigated by a brief further thought about autonomy. Contrary to what you suggest, I do not think that respecting autonomy requires being entirely candid. Rather, respecting autonomy requires deferring to the patient's judgments about how she ought to be treated and what she should be told. If she wants health professionals to be entirely candid with her -- and many patients do want this and will tell you so -- then respecting autonomy means being entirely candid. But in the absence of such an insistence on hearing the whole truth, you can make your own judgment about what the patient would want. (Directly asking her how much honesty she wants is tricky because it deprives her of what may be for her the best option: being uncertain about how honest you are with her while trusting that you will make decisions about what to tell her in a way that you judge is best for her.) And you may well conclude -- correctly -- that the patient is scared and would rather be firmly reassured than hear the full litany of risks and disclaimers.
When a patient really presses you for the truth, then you can -- in addition to the obfuscations you mention in your last two sentences -- also say truthfully that the condition does have an important psychological component and is most likely (ambiguity intended) to remain latent if one worries about it as little as possible.