Are Backscatter x-ray machines at airports safe?

Two competing devices are used in airports to obtain full-body images of passengers: backscatter x-ray scanners and millimeter wave scanners. Today I want to examine those scanners that use x-rays.

Backscatter x-ray scanners work by a different mechanism than ordinary x-ray images used in medicine. Chapter 16 of the 4th edition of Intermediate Physics for Medicine and Biology discusses traditional medical imaging (see Fig. 16.14). X-rays are passed through the body, and the attenuation of the beam provides the signal that produces the image. Backscatter x-ray scanners are different. They record the x-rays scattered backwards toward the incident beam via Compton scattering. This allows the use of very weak x-ray beams, resulting in a lower dose.

The dose (or, more accurately the equivalent dose) from one backscatter x-ray scan is about 0.05 μSv. The unit of a sievert is defined in Chapter 16 of Intermediate Physics for Medicine and Biology as a joule per kilogram (the definition includes a weighting factor for different types of radiation; for x-rays this factor is equal to one). The average annual background dose that we are all exposed to is about 3 mSv, or 3000 μSv, arising mainly from inhalation of the radioactive gas radon. Clearly the dose from a backscatter x-ray scanner is very low, being 60,000 times less than the average yearly background dose.

Nevertheless, the use of x-rays for airport security remains controversial because of our uncertainly about the effect of low doses of radiation. Russ Hobbie and I address this issue in Section 16.13 about the Risk of Radiation.

In dealing with radiation to the population at large, or to populations of radiation workers, the policy of the various regulatory agencies has been to adopt the linear-nonthreshold (LNT) model to extrapolate from what is known about the excess risk of cancer at moderately high doses and high dose rates, to low doses, including those below natural background.

If the excess probability of acquiring a particular disease is αH [where H is the equivalent dose in sieverts] in a population N, the average number of extra persons with the disease is

m = α N H.

The product NH, expressed in person-Sv, is called the collective dose. It is widely used in radiation protection, but it is meaningful only if the LNT assumption is correct [emphasis added].

So, are backscatter x-ray scanners safe? This question was debated in a Point/Counterpoint article appearing in the August issue of Medical Physics, a leading journal published by the American Association of Physicists in Medicine. A Point/Counterpoint article is included in each issue of Medical Physics, providing insight into medical physics topics at a level just right for readers of Intermediate Physics for Medicine and Biology. The format is always the same: two leading medical physicists each defend one side or the other of a controversial proposition. In August, the proposition is “Backscatter X-ray Machines at Airports are Safe.” Elif Hindie of the University of Bordeaux, France argues for the proposition, and David Brenner of Columbia University argues against it.

Now let us see what Drs. Hindie and Brenner have to say about this idea. Hindie writes (references removed)

The LNT model postulates that every dose of radiation, no matter how small, increases the probability of getting cancer. This highly speculative hypothesis was introduced on the basis of flimsy scientific evidence more than 50 years ago, at a time when cellular biology was a largely unexplored field. Over the past decades, an ever-increasing number of scientific studies have consistently shown that the LNT model is incompatible with radiobiological and experimental data, especially for very low doses.

The LNT model was mainly intended as a tool to facilitate radioprotection regulations and, despite its biological implausibility, this may remain its raison d’être. However, the LNT model is now used in a misguided way. Investigators multiply infinitesimal doses by huge numbers of individuals in order to obtain the total number of hypothetical cancers induced in a population. This practice is explicitly condemned as “incorrect” and “not reasonable” by the International Commission on Radiological Protection, among others.

Brenner counters

Of course this individual risk estimate is exceedingly uncertain. Some have argued that the risk at very low doses is zero. Others have argued that phenomena such as tissue/organ microenvironment effects, bystander effects, and “sneaking through” immune surveillance, imply that low-dose radiation risks could be higher than anticipated. The bottom line is that individual risk estimates at very low doses are extremely uncertain.

But when extremely large populations are involved, with up to 10⁹ scans per year in this case, risk should also be viewed from the perspective of the entire exposed population. Population risk quantifies the number of adverse events expected in the exposed population as a result of a proposed practice, and so depends on both the individual risk and on the number of people exposed. Population risk is described by ICRP as “one input to . . . a broad judgment of what is reasonable,” and by NCRP as “one of the means for assessing the acceptability of a facility or practice.” Population risk is considered in many other policy areas where large populations are exposed to very small risks, such as nuclear waste disposal or vaccination.

The debate about the LNT model and the validity of the concept of collective dose is not merely of academic interest. It gets to the heart of how we perceive, assess, and defend ourselves against the risk of radiation. Low doses of radiation are risky to a large population only if there is no threshold below which the risk falls to zero. Until the validity of the linear non-threshold model is confirmed, I suspect we will continue to witness passionate debates—and future point/counterpoint articles—about the safety of ionizing radiation.