LNT is the only data we have but that doesn't make it right. All the data points for figuring radiation risk are based on high level exposures, LNT is the best

LNT is the only data we have but that doesn't make it right. All the data points for figuring radiation risk are based on high level exposures, LNT is the best fit line that passes through 0,0. The problem is that this makes the assumption that the line is straight. The problem with this is we have many data points for lower level exposure that seem to almost all lie well below the line, but none have the statistical power to prove it even with very large sample sizes. Even the hottest of these data points shows nothing--why isn't there a cancer cluster in the people who live the lives in 20 rem/year?? And note that we aren't dismissing the LNT out of hand, we are questioning it in a certain realm. While I question the homeostasis argument it seems completely reasonable that the rate has something to do with it. In chemistry the norm is the dose makes the poison (you can breathe 3 psi of oxygen for a lifetime, 30 psi of oxygen and you'll be in trouble in less than an hour), what's so unreasonable about thinking that it might apply in this case? (And note that I don't think it applies much in your realm--while medical imaging isn't all that much radiation it's delivered quickly. Other than my one encounter with Tc-99m everything I've encountered has been in hold-your-breath type times.) There's also the possibility the true line is a parabola. We have so few data points at the high end that the current data is completely incapable of excluding this case and it would fit the low data points better.