Biotech Values

[iwfal]

It certainly means that RFA has improved a lot in the last 5 or so years..

This is incorrect and one of the roots of you error. You are assuming that we (humans) know enough to predict pcs/survival well. We cannot - even in the best of circumstances we stink at it (e.g in a single well controlled trial we get, at best, a C statistic of 0.8 or so). And going across trials is much worse. And trying to go across trials based upon publicly available data is worse yet.