Exactly. So—assuming all results are statistically significant—where's the causal relationship? Radiation-> Cancer? Radiation->Lives longer? Lives longer->Cancer? and if there are two results that are related independently of this study (lives longer/gets more cancer), then how can we presume one of the other results is more "valid" than another (as broadcast in the headlines)
There were 6 times as many test groups as control groups (3 intensities * 2 modulations). It would have been good to have more control groups to reduce statistical effects.