I'm not quite sure how this thread has shifted to discussion of androgen replacement, but when a study says "[ADT] makes no difference [in BCR and overall survival after 70]," it may not mean what you think it means. Studies always look for statistical significance in samples. So this statement most likely means the study (as designed and conducted) was not able to measure a statistically different result, not that in the real world there is no difference. This is meaningful not because men live just as long with or without the disease and/or the treatment, but because every treatment (and every man's life) is uncertain, with time and chance happening to us all.
So, if you decide not to treat, or to vary a treatment protocol, you're always taking an [inexactly] calculated risk. In this case, we're dealing with time to death estimates that are difficult to measure in any study. To give an example, a recent treatment for brain cancer "miraculously" increases mean survival from 18 months to 36 months. That's much easier to measure than increasing mean survival rates of 15 years (from PC) in men over 70 whose mean survival rate overall may also be around 15 years....