Hoisted from Dave Farber's mailing list:
From: Ethan Ackerman
Date: January 12, 2010 1:22:00 PM EST
Subject: Re: [IP] Stop the panic on air security – err, no, irradiate it
Since the Schneier editorial brings up the subject of thinking rationally about small risks…
IPers following the debate around TSA's whole body scanning might have noticed that not too much ink has been spilled over the fact that these imagers are a source of x-rays – ionizing radiation. Ionizing radiation (at the right dose and probability) can cause or increase the likelihood of cancer and other ailments.
But one reason there's not been a big hullabaloo is because the risks from these machines are rather small, though not zero. How small a risk? About as (un)likely as a terrorist attack, it turns out.
The risk of being on a plane subject to a terrorist attack is ~1 in 10 million. 
Similarly, a single backscatter scan corresponds to a 5% increased risk of fatal cancer in ~1 in 10 million cases. ( While reliable studies suggest that a scan-level dose would result in a statistically verifiable increase in fatal cancer risk in about 1 scan in 100,000, the “5% increased risk at 1 in 10 million” conclusion is supported with more studies than the former, and more statistically sound.) 
So how many additional cases of fatal cancer (or just debilitating cancer, or just cataracts) is it worth for us as a society to cause an innocent traveler in order to possibly detect a drug smuggler or would-be-bomber?
In how many people are we ok with just increasing the likelihood of cancer for this kind of security?
Can you give a number? The TSA and FDA already have.
 http://www.fivethirtyeight.com/2009/12/odds-of-airborne-terror.html [www.fivethirtyeight.com]
 http://www.fda.gov/ohrms/dockets/AC/03/briefing/3987b1_pres-report.pdf [www.fda.gov]
-The dose-adjusted nominal risk estimate of fatal cancer associated with exposure from a single backscatter x-ray scan is 0.0000005% for a member of the general public, at a 5% increased risk of fatal cancer per Sievert dosed and a single scan dose of 0.1 microSieverts.
Although, I suppose, someone might argue that the risks are distributed slightly differently in that all of the cancer risk goes to the traveling public, while some of the terrorism risk is shared by people in the flightpath if a plane lands on them.