Search This Blog

Sunday, December 25, 2011

Linear No Threshold Model & LNT Deniers

I had some discussions with LNT deniers yesterday on a couple of blogs.  So, I thought I'd do a quick write-up so I'd could produce a link in the future.


A dose response curve is just a set of data points.  The abscissa is in units of dose and the ordinate is in terms of observed effect.  One collects data and plots it.  There is a statistical spread with each data point because the data originates from epidemiological studies which have sources of error.  There isn't anything to "prove" here.  It's just a description of the observations.

Now let's look at the dose response curve for radiation.  The abscissa is in units of Sv's and the ordinate is in terms of excessive relative risk of solid cancer.  Open a second browser with this link:

http://www.nap.edu/openbook.php?record_id=11340&page=16

Ignore the "Leukemia" insert.  Notice the points (and error bars) on the graph.  Notice that two lines are drawn to fit the observed points.  The dotted line (with slight curvature) represents a linear-quadratic (l-q) fit of the data.  The solid line (no curvature) represents a linear fit of the data. 

Both curves fit the data and since a linear fit is easier to work and nothing is gained by using a l-q fit, we go with linear.  This model fitting doesn't mean anyone thinks each unit of dose produces an exactly corresponding increase in risk.  It's like half-life.  On average, we can determine the rate of decay of a large number of atoms.  That doesn't mean if we look at two atoms that one of them will exactly decay after one half life. So again, LNT is just a generalized description of what we see.

We know, before undertaking any epidemiological study, that there will be a point in which we can't statistically discern the effect from background.  We know this detection threshold will exist with 100% certainty.  It's an "amazing coincidence" that this coincides with LNT deniers assumed effect threshold!

Look at the graph again.  If, instead of excess risk, the ordinate was actual cancer incidence, the line would be much higher on the graph and at zero dose would cross the ordinate at about 42%, because people get cancer with zero excess dose (above background). 

The LNT deniers need to provide evidence to explain how such a linear curve (or linear-like, if you prefer) magically changes in the exact same window where epidemiology can't discern background radiation cancers from elevated radiation cancers.  But magically becomes linear where epidemiology provides data.

The reason they won't be able to do so is because even one photon has the potential (ie, chance or risk) to contribute to the chain of events associated with carcinogenesis.  There is no threshold in the energy region of ionizing radiation.  In order for there to be a threshold either:

1.  100% of cells exposed die.  In that way cancer cannot evolve.  We know this doesn't happen.

2.  100% of cells exposed don't get damaged.  In that way cancer cannot evolve.  We know this doesn't happen.

3.  100% of cells which are damaged and don't die repair themselves with 100% accuracy.  In that way cancer cannot evolve.  We know this doesn't happen.

Since there is a <100% chance of those items mentioned above, there must be a >0% chance of causing cancer. Therefore there is no threshold in regards to dose. 

But there is a threshold.  It is an energy threshold.  Go below the ionization range of photon energies, into the lowest ultraviolet range and the radiation doesn't have the energy to cause cancer.

One commenter got into a discussion about linearity on the scale of one photon followed by another.   No one knows what the actual dose response model is on that scale.   But that really isn't the scale we are concerned with.  At the low dose end we are concerned with effects around background levels.   For a typical person this is about 1 mrem per day.  If one does the math, this works out to be on the order of tens of thousands of photon interactions per second (of course background radiation doesn't only include photons and there are dozens of photon energies associated with photon emitters).

No comments:

Post a Comment