Back-Scatter X-rays
This is a followup to my last post about the back-scatter xrays. Several people have pointed me to this letter by John Sedat et al, expressing concerns that the TSA's back-scatter xray machines are more harmful than we've been led to believe. I'd already cited it in my original post but it keeps coming up. I've gone through it and the FDA's response in greater detail. Here's a summary of the new things I've learned. To quote Mary Poppins Maria Von Trapp née Rainer, I'll start at the very beginning, which is a very good place to start.
Units
Radiation output is measured in Joules. One joule is equivalent to putting out 1 watt in for 1 second. A 100 watt lightbulb puts out 100 joules of radiation, not all of which is visible light.
Radiation dosage is measured in gray. One gray is the equivalent of one joule being absorbed by one kilogram of matter. A 100 kilogram Bill Bixby receiving 100 joules of gamma radiation has received a radiation dosage of 1 gray. Rads are an outdated unit corresponding to 1/100 gray.
But joules are not joules in biology. Some radiation is more damaging than other radiation. Radiation dosage impact is measured in Sievert units which are the absolute dose (in gray) scaled by the relative biological effectiveness of the particular radiation (Q value) and the susceptibility of the thing being irradiated (N value) to give an effective dose unit. Rems are a similar but outdated and discouraged unit.
The good thing about Sieverts is that we're comparing apples to apples in terms of dosage and harm. The bad thing about Sieverts is that it's not an objectively empirical unit like joules and watts, which means it's fair to question whether the appropriate fudge factors have been used.
X-rays
X rays are a band on the electromagnetic spectrum between 0.01-10 nm with energies from 120 eV to 120 keV. "Soft" x-rays are low power (0.12 to 12 keV), easily absorbed by many materials. "Hard" x-rays have enough energy (12 to 120 keV) to penetrate many materials. Both soft and hard x-rays cause ionizing damage. Medical x-ray machines filter out the soft x-rays because they cause the same damage, but they don't penetrate deep enough to be useful for imaging. Soft x-rays aren't any more risky than hard x-rays joule for joule, but in transmission x-rays they're a risk without benefit so doctors avoid using them. Higher energy x-rays don't just penetrate further; they have a smaller cross section which means they're less likely to interact with matter. They can also hit more things before they stop. In 1 joule of low-energy x-rays there are more photons, because you need more of them to add up to 1 joule, and those photons are more likely to be absorbed.
Back-scatter X-rays
Back-scatter works by a process conceptually (but not physically) similar to fluorescence. An X-ray hits an atom, which absorbs some energy. The atom ejects an electron and a lower power X-ray, a process called compton scattering. Since momentum must be conserved, scattering angle can be determined by the energy, which lets you determine where it came from. Damage is actually caused by several products of the interaction: the energetic electron, the ionized free radical left behind, and a reduced-strength X-ray that continues forward.
Ordinary medical x-rays are high energy (50-200 keV) to penetrate through the body and expose the plate on the other side. Back-scatter x-rays (28keV) don't need to pass through the body, they just need to be powerful enough to produce the scattering effect. This page to correlates beam strength and penetration, with x-rays penetrating about 3.75x further as their strength doubles. According to your humble lay-author's back of the envelope calculation a medical x-ray beam will penetrate between 3.5-50x further since they're in the 50-200keV range, be absorbed by that much more material. Due to the shallower penetration depth, the dose from a back-scatter beam ought to be that much greater joule-for-joule.
Radiation Damage
We are surrounded by very small levels of natural ionizing radiation. Our bodies are capable of repairing small amounts of damage that natural background radiation causes. (Some bacteria are much better at it than we are.) The body has difficulty repairing large amounts of damage: too much damage at once is irreparable. Practicing radiologist drjohn pointed out that a theory called radiation hormesis predicts that small amounts of radiation are beneficial, since the body's systems are mobilized to repair everywhere when it detects the damage somewhere. Some evidence supports this, but it is far from conclusive. There's an opposite theory saying that any radiation is bad radiation, even low levels that the body can repair. There is evidence to indicate that this is not true, but this evidence is not completely conclusive. My take-away from this point is that if you want to accept tentative, unproven fringe claims one can just as easily say "maybe small doses of radiation are healthy" as "maybe they're harmful".
The Letter of Concern
John Sedat's primary concern is that the lower beam energies of soft X-rays will be preferentially absorbed in the shallow skin, rather than be absorbed by the entire body like hard X-rays. This seems like a reasonable, well-founded concern. 1 joule of 28keV x-rays will contain more photons than 1 joule of 200keV x-rays, and they will be absorbed by less mass.
On the other hand the units I've found are Sievert units, whose Q value is already scaled by this value. That's the whole point of Sievert units: they let you compare apples to apples. The again the people providing these numbers may not have taken this into account. ikkyu2 thinks they might not have, and this question is worth investigating.
John Sedat's letter warns of dosage increases "possibly by one to two orders of magnitude", which is similar to my estimation that hard x-rays penetrate 3.5-50x further. But at 25μrem total per scan that's still a practical dose of 0875-1.25mrem per scan. The plane ride itself will give you 0.24-0.66 mrem/hr. If that extrapolation is correct (it's not - see below) the dose increase we're talking about would still be only equivalent to an extra 8 - 300 minutes of flight time. The standard for annual effective dose is 25 mrem - thats 1000 screenings if the TSA is right, 20-280 screenings if Sedat is right, and 37-104 hours of flight time if they scrap the screening systems entirely.
If you're that worried about ionizing radiation, don't get screened but more importantly don't fly. Complaining about ionizing radiation on your way to a plane flight is like complaining about a TSA pat-down on your way to Caligula's palace. Especially if you're flying to Reno and getting a week's worth of sun exposure at Burning Man, 4000 feet above sea level. Don't make me calculate how many thousands of x-rays are equivalent to one hour of Playa exposure.
The Letter of Response
The above fudge factors are numbers that I've estimated based on being a fairly smart expert in absolutely nothing. The DHHS employs experts who know more than I do, and their response involves significantly better information and less guesswork than my analysis.
They point out that while backscatter x-rays are concentrated in the skin, the skin is as good as bone at resisting/repairing x-ray damage. The general annual effective dose limit for the public is .25 mSv, but the recommended limit for annual dose to the skin is 50 mSv.
They list the skin specific dosage for the RapiScan Secure 1000 at 0.56 μSv, and say that a Johns Hopkins University Applied Physics Laboratory Assessment has experimentally confirmed this dosage. These numbers have been confirmed by Sandia National Laboratory (1991), the FDA (1998), The Folsom State Prison (1999), and the NIST (2006). It would take 89,000 scans (one scan every 6 minutes for a year) to reach the annual dosage limit for skin.
Of course skin isn't the only organ exposed to X-rays. Breasts and testes tissue are under thin skin, more sensitive, and receive an x-ray dose as well. Which is why the FDA uses software called PXMC to estimate organ dosage using the same monte carlo techniques that I use at work to compute light distribution. They found that "the dose to other organs is less than, equal to, or at most approximately three times the effective dose". Worst-case your most fragile organs still need over 300 scans per year before you reach the annual effective dose limit.
Maybe Everything We've Read Is A Lie
There's an upcoming paper by Peter Rez in Radiation Protection and Dosimetry that questions whether the published figures are correct. It's possible that the FDA, back-scatter companies, Sandia, FDA, Folsom Prison, and NIST are all lying to us about the dosages they claim to have measured. It's also possible that everyone I've ever met is lying to me in some sort of Truman Show charade. Unfortunately I can only go on the information I have. Fortunately it ought to be pretty easy for a third party to put a dosimeter in a screening machine, repeat Johns Hopkins' study and figure out what the incident radiation actually is. They could even do this without the TSA's or manufacturer's permission by taping the dosimeter to their chest and going through a screening. Hopefully they will, because "what if all the figures that anyone knows are wrong and everyone's lying" is a frustrating argument to present and refute.
Conclusion
This debate reminds me an awful lot of the debate about Thimerosal in vaccines. On one hand they're right - thimerosal contains mercury, which is toxic. On the other hand the safety standards and margins are designed with this in mind. That's why the doses are so small, and why the safety margins are so enormous.
But both anti-vax groups or the emerging anti-backscatter groups rarely recognize these margins. They rarely contextualize the risks. They talk about toxins and and radiation and elevated risk but not compare it to anything meaningful. Anti-vax people will try to avoid discussing how much mercury is in a vaccine vs a tuna fish sandwich, they just keep repeating a facetiously simplistic party line that toxins are bad. Thimerosal has nonetheless been phased out of vaccines to humor poorly informed, panicky people with poor risk estimation skills. I wouldn't be surprised if the current brouhaha causes the TSA to phase out back-scatter x-ray machines. I'll count that as another victory for poorly informed, panicky people with poor risk estimation, not public safety.
For now the whole thing - the infinitesimally small chance that you will get cancer from this infinitesimally small dose of radiation, the infinitesimally small chance that it will prevent someone on your flight from exploding a dick bomb, and the infinitesimally small chance that your TSA agent is actually going to enjoy being in momentary contact with your infinitesimally boring dick - seems like an infinitesimally small tempest in an infinitesimally small teapot.
Update: Ars Technica is a good source.
Units
Radiation output is measured in Joules. One joule is equivalent to putting out 1 watt in for 1 second. A 100 watt lightbulb puts out 100 joules of radiation, not all of which is visible light.
Radiation dosage is measured in gray. One gray is the equivalent of one joule being absorbed by one kilogram of matter. A 100 kilogram Bill Bixby receiving 100 joules of gamma radiation has received a radiation dosage of 1 gray. Rads are an outdated unit corresponding to 1/100 gray.
But joules are not joules in biology. Some radiation is more damaging than other radiation. Radiation dosage impact is measured in Sievert units which are the absolute dose (in gray) scaled by the relative biological effectiveness of the particular radiation (Q value) and the susceptibility of the thing being irradiated (N value) to give an effective dose unit. Rems are a similar but outdated and discouraged unit.
The good thing about Sieverts is that we're comparing apples to apples in terms of dosage and harm. The bad thing about Sieverts is that it's not an objectively empirical unit like joules and watts, which means it's fair to question whether the appropriate fudge factors have been used.
X-rays
X rays are a band on the electromagnetic spectrum between 0.01-10 nm with energies from 120 eV to 120 keV. "Soft" x-rays are low power (0.12 to 12 keV), easily absorbed by many materials. "Hard" x-rays have enough energy (12 to 120 keV) to penetrate many materials. Both soft and hard x-rays cause ionizing damage. Medical x-ray machines filter out the soft x-rays because they cause the same damage, but they don't penetrate deep enough to be useful for imaging. Soft x-rays aren't any more risky than hard x-rays joule for joule, but in transmission x-rays they're a risk without benefit so doctors avoid using them. Higher energy x-rays don't just penetrate further; they have a smaller cross section which means they're less likely to interact with matter. They can also hit more things before they stop. In 1 joule of low-energy x-rays there are more photons, because you need more of them to add up to 1 joule, and those photons are more likely to be absorbed.
Back-scatter X-rays
Back-scatter works by a process conceptually (but not physically) similar to fluorescence. An X-ray hits an atom, which absorbs some energy. The atom ejects an electron and a lower power X-ray, a process called compton scattering. Since momentum must be conserved, scattering angle can be determined by the energy, which lets you determine where it came from. Damage is actually caused by several products of the interaction: the energetic electron, the ionized free radical left behind, and a reduced-strength X-ray that continues forward.
Ordinary medical x-rays are high energy (50-200 keV) to penetrate through the body and expose the plate on the other side. Back-scatter x-rays (28keV) don't need to pass through the body, they just need to be powerful enough to produce the scattering effect. This page to correlates beam strength and penetration, with x-rays penetrating about 3.75x further as their strength doubles. According to your humble lay-author's back of the envelope calculation a medical x-ray beam will penetrate between 3.5-50x further since they're in the 50-200keV range, be absorbed by that much more material. Due to the shallower penetration depth, the dose from a back-scatter beam ought to be that much greater joule-for-joule.
Radiation Damage
We are surrounded by very small levels of natural ionizing radiation. Our bodies are capable of repairing small amounts of damage that natural background radiation causes. (Some bacteria are much better at it than we are.) The body has difficulty repairing large amounts of damage: too much damage at once is irreparable. Practicing radiologist drjohn pointed out that a theory called radiation hormesis predicts that small amounts of radiation are beneficial, since the body's systems are mobilized to repair everywhere when it detects the damage somewhere. Some evidence supports this, but it is far from conclusive. There's an opposite theory saying that any radiation is bad radiation, even low levels that the body can repair. There is evidence to indicate that this is not true, but this evidence is not completely conclusive. My take-away from this point is that if you want to accept tentative, unproven fringe claims one can just as easily say "maybe small doses of radiation are healthy" as "maybe they're harmful".
The Letter of Concern
John Sedat's primary concern is that the lower beam energies of soft X-rays will be preferentially absorbed in the shallow skin, rather than be absorbed by the entire body like hard X-rays. This seems like a reasonable, well-founded concern. 1 joule of 28keV x-rays will contain more photons than 1 joule of 200keV x-rays, and they will be absorbed by less mass.
On the other hand the units I've found are Sievert units, whose Q value is already scaled by this value. That's the whole point of Sievert units: they let you compare apples to apples. The again the people providing these numbers may not have taken this into account. ikkyu2 thinks they might not have, and this question is worth investigating.
John Sedat's letter warns of dosage increases "possibly by one to two orders of magnitude", which is similar to my estimation that hard x-rays penetrate 3.5-50x further. But at 25μrem total per scan that's still a practical dose of 0875-1.25mrem per scan. The plane ride itself will give you 0.24-0.66 mrem/hr. If that extrapolation is correct (it's not - see below) the dose increase we're talking about would still be only equivalent to an extra 8 - 300 minutes of flight time. The standard for annual effective dose is 25 mrem - thats 1000 screenings if the TSA is right, 20-280 screenings if Sedat is right, and 37-104 hours of flight time if they scrap the screening systems entirely.
If you're that worried about ionizing radiation, don't get screened but more importantly don't fly. Complaining about ionizing radiation on your way to a plane flight is like complaining about a TSA pat-down on your way to Caligula's palace. Especially if you're flying to Reno and getting a week's worth of sun exposure at Burning Man, 4000 feet above sea level. Don't make me calculate how many thousands of x-rays are equivalent to one hour of Playa exposure.
The Letter of Response
The above fudge factors are numbers that I've estimated based on being a fairly smart expert in absolutely nothing. The DHHS employs experts who know more than I do, and their response involves significantly better information and less guesswork than my analysis.
They point out that while backscatter x-rays are concentrated in the skin, the skin is as good as bone at resisting/repairing x-ray damage. The general annual effective dose limit for the public is .25 mSv, but the recommended limit for annual dose to the skin is 50 mSv.
They list the skin specific dosage for the RapiScan Secure 1000 at 0.56 μSv, and say that a Johns Hopkins University Applied Physics Laboratory Assessment has experimentally confirmed this dosage. These numbers have been confirmed by Sandia National Laboratory (1991), the FDA (1998), The Folsom State Prison (1999), and the NIST (2006). It would take 89,000 scans (one scan every 6 minutes for a year) to reach the annual dosage limit for skin.
Of course skin isn't the only organ exposed to X-rays. Breasts and testes tissue are under thin skin, more sensitive, and receive an x-ray dose as well. Which is why the FDA uses software called PXMC to estimate organ dosage using the same monte carlo techniques that I use at work to compute light distribution. They found that "the dose to other organs is less than, equal to, or at most approximately three times the effective dose". Worst-case your most fragile organs still need over 300 scans per year before you reach the annual effective dose limit.
Maybe Everything We've Read Is A Lie
There's an upcoming paper by Peter Rez in Radiation Protection and Dosimetry that questions whether the published figures are correct. It's possible that the FDA, back-scatter companies, Sandia, FDA, Folsom Prison, and NIST are all lying to us about the dosages they claim to have measured. It's also possible that everyone I've ever met is lying to me in some sort of Truman Show charade. Unfortunately I can only go on the information I have. Fortunately it ought to be pretty easy for a third party to put a dosimeter in a screening machine, repeat Johns Hopkins' study and figure out what the incident radiation actually is. They could even do this without the TSA's or manufacturer's permission by taping the dosimeter to their chest and going through a screening. Hopefully they will, because "what if all the figures that anyone knows are wrong and everyone's lying" is a frustrating argument to present and refute.
Conclusion
This debate reminds me an awful lot of the debate about Thimerosal in vaccines. On one hand they're right - thimerosal contains mercury, which is toxic. On the other hand the safety standards and margins are designed with this in mind. That's why the doses are so small, and why the safety margins are so enormous.
But both anti-vax groups or the emerging anti-backscatter groups rarely recognize these margins. They rarely contextualize the risks. They talk about toxins and and radiation and elevated risk but not compare it to anything meaningful. Anti-vax people will try to avoid discussing how much mercury is in a vaccine vs a tuna fish sandwich, they just keep repeating a facetiously simplistic party line that toxins are bad. Thimerosal has nonetheless been phased out of vaccines to humor poorly informed, panicky people with poor risk estimation skills. I wouldn't be surprised if the current brouhaha causes the TSA to phase out back-scatter x-ray machines. I'll count that as another victory for poorly informed, panicky people with poor risk estimation, not public safety.
For now the whole thing - the infinitesimally small chance that you will get cancer from this infinitesimally small dose of radiation, the infinitesimally small chance that it will prevent someone on your flight from exploding a dick bomb, and the infinitesimally small chance that your TSA agent is actually going to enjoy being in momentary contact with your infinitesimally boring dick - seems like an infinitesimally small tempest in an infinitesimally small teapot.
Update: Ars Technica is a good source.