Scaremongering, Cancer and Medical Imaging
March 12th, 2010
|
| Share |
In recent times, medical imaging procedures that utilize x-rays have come under increased scrutiny as numerous media outlets and even peer-reviewed studies have reported that these procedures are putting patients at risk of cancer and causing tens of thousands of new cases of cancer per year, with many of those resulting in death. These reports are leading to not only increased fear of medical imaging, but also to calls for policy changes and reduced use of x-ray medical imaging.
The CT scan has become on of the biggest targets. CT scans are rapidly becoming one of the most common forms of medical imaging. They’re relatively quick, painless and provide excellent diagnostic data for a wide number of conditions. The technology of CT scanning and imaging processing has improved dramatically in recent years, making the procedures both more economical and increasing data quality and resolution. There’s no doubt that these procedures save lives and improve the quality of life by giving doctors a look inside the human body at both bone and soft tissue.
CT scans do subject patients to some radiation exposure, however. The levels of radiation which a patient receives from a CT scan (or even several of them) has never been shown to actually cause any long term harm, but the long-standing and many times debunked linear non-threshold model presumes that it will. LNT was conceived as a “worst case scenario” for the effect that ionizing radiation might have on the human body in the earliest days of radiation research, when data was scarce.
LNT holds that radiation effect is directly proportional to dose, all the way down to zero and therefore assumes that the human body is incapable of repairing damage caused by radiation, recovering from radiation-induced cellular damage or in any way mitigating the effects of radiation. It further assumes that the effect of a few atoms in a human cell being ionized is smaller, but otherwise no different than an enormous amount of ionization of the cells.
There’s no doubt that cancer risk is increased by radiation if the dose is high enough. Radiation can damage the body on both a macroscopic and cellular level. A high enough dose of radiation can cause tissue damage, internal bleeding and even death. Those who do recover from multiple incidence of near-fatal acute radiation poisoning may face life-long complications and health issues relating to the damage caused by radiation. At high enough levels, the dangers appear to be relatively linear in relation to radiation dose. In other words, a person exposed to 1000 rems has approximately twice the risk of complications like cancer as someone exposed to 500.
By this logic, one rem should result in .1% the likelihood of developing cancer as 1000 rems. This has been likened to observing that when a person is thrown off a ten story building they die 100% of the time, when thrown off of a 5 story building, they die 50% of the time and therefore assuming that 10% will die from a one story fall and 1% will die after falling 12 inches. Furthermore, the implication is that the optimal living situation is zero radiation exposure (something which is effectively impossible) and the higher the exposure is over zero, the worse off you are for cancer risk.
Needless to say, many have pointed out that there’s a flaw in this logic and as knowledge of radiation has increased, LNT has begun to look less and less likely. Unfortunately for the promoters of LNT, examination of the cancer rates in populations living in high radiation background areas shows no increase in cancer and in fact, has produced some evidence of a decreased risk of cancer.
Yet what these reports on the dangers of medical imaging really are is nothing more than a combination of the LNT hypothysis and fifth grade level mathematics.
In the media:
SF Gate: CT scan cancer warnings worry patients
WEB MD: Multiple CT Scans Raise Cancer Risk
Daily Mail: CAT scan cancer fear: Radiation ‘could trigger the disease in one in 80 patients’
US News and World Report: Too Many CT Scans Pose Cancer Risk, Studies Say
Wall Street Journal: Radiation Risks Prompt Push to Curb CT Scans
The math:
Radiation Exposure From Procedure * Presumed Increased Cancer Risk Per Rem = Increased Risk Per Procedure
Increased Use Per Procedure * Number of Patients Who Receive Procedure = Presumed Cancer Incidence Caused By Procedure
Assumed lifetime risk increase per rem is .08% ( .0008) – this is based on extrapolation of observational data with the presumption that the risk is a linear function. (It has not been proven through direct observational data at such low levels.)
An average CT Scan will produce an exposure of roughly 600 mrem pr .60 rem (A full body CT Scan is roughly one rem* A local scan of the chest Abdomen pelvis is roughly 450 600 mrem)
There are at least 62 million CT scans preformed per year, in the United States alone. Hundreds of millions are preformed world wide
*depending on the exact nature of the equipment and procedure. Nominal exposure from Chest, Abdomen and Pelvis averages 990 mrem, slightly more if additional local imaging is included.
Therefore, using this simple formula:
.60 * .0008 = 0.00048 or .048% increase in lifetime cancer risk
.00048 x 62,000,000 = 29,760
Thus we have come to the conclusion that in the United States alone, there will be about twenty nine to thirty thousand new cases of cancer per year due to CT scans.
Of course, we could go a bit further with this, we could break down CT scan dose further and rather than just using 600 mrem as a good approximation, we could break down the number of CT scans by type and the independently multiply each by the exposure it results in. We could do the same with everything from simple two-dimensional bone x-rays to mammograms, dental x-rays, angiograms and all the other procedures that use x-rays and then add them all up.
In the end, however, all we’re really doing is fourth grade math – multiply the exposure level from the procedure times the presumed added risk per rem (or sievert if you prefer to be metric about it) and then multiply that by the number of procedures preformed. Thus, we have our theoretical number of cancer incidence produced by the use of the procedure.
Numbers which, I will stress again, don’t come from going out there and collecting data or studying the results of medical imaging on the population, but simply using a hypothetical constant whose validity is never actually addressed by the study itself.
Is this actually any revelation? Is this news?
There’s absolutely no observational data here and there’s no empirical evidence to support the numbers at all. It’s just multiplying out the numbers based on a pre-conceived notion with no actual validation that this notion is even true. Furthermore, it’s math that you could do on the back of a cocktail napkin.
So how is this really news? It’s not. Yet somehow it gets reported as if it were.
Conclusion:
What these studies indicate is really fairly obvious: the LNT hypothesis indicates that any level of radiation exposure will increase cancer risk. CT scans use ionizing radiation. Therefore if CT scans are used, LNT predicts that there will be an increase in cancer proportional to the number of CT scans preformed. This is obvious, even self-evident, but it should be stressed that this is only based on the LNT model – it does not actually mean that there will be more incidence of cancer.
This entry was posted on Friday, March 12th, 2010 at 11:54 am and is filed under Bad Science, Good Science, Misc, Obfuscation, Quackery, media. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
View blog reactions
March 12th, 2010 at 1:43 pm
The linear no-threshold (LNT) model is not supported by scientific data at doses less than about 100 mSv or at chronic dose rates up to at least 200 mSv yr−1. Radiation protection practices based on the LNT model yield no demonstrable benefits to health when applied at lower annual doses. The assumption that such exposures are harmful may not even be conservative and has helped to foster an unwarranted fear of low-level radiation.
After the atomic bomb explosions in Hiroshima and Nagasaki, studies concerning life span of atomic bomb survivors showed a linear relationship between cancer mortality and high doses of radiation. The United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), then proposed the linear no-threshold (LNT) theory in 1958. Chernobyl, where a large population of people received radiation doses in the 50 to 200 millisieverts range, has been the first large-scale opportunity to test whether the LNT assumption is true. The evidence suggests it is not.
Quote Comment
March 12th, 2010 at 6:14 pm
Sorry I’m just a little confused. You’re saying that these studies and published reports are meaningless because they just do some simple multiplication and don’t validate the results I guess? Where does the .0008 come from? Why do they use that number for anything and how does this relate to whether or not the results are meaingful?
Quote Comment
March 13th, 2010 at 2:02 pm
Biff Henderson said:
Basically, yeah. The people who wrote the reports in question simply took a mathematical model which has never been experimentally verified to be correct for the circumstances to which they are applying it and ran the numbers. In fact the majority of experimental data suggests that the model is NOT accurate in those circumstances.
Biff Henderson said:
It’s the assumed increased cancer risk from 1 rem of exposure, extrapolated from high dosage experiments.
Biff Henderson said:
When radiation safety protocols were first being developed there simply hadn’t been an opportunity to run experiments with low dosages. So the people responsable simply extrapolated from the high dose rate information they did have, even though they knew that would almost certainly overstate the risk. Of course when you’re designing safety regulations, erring on the side of safety seems like a wise choice.
The other issue is that the the safety regulations are designed for people expected to have high opportunities for chronic exposure, nuclear plant workers, radiolgy technicians, etc. If you get an x-ray maybe once every 5 years or so, like a normal person, you’re far more likely to die from driving to and from the hospital than you are to die from anything that happens in the x-ray room. Of course if you’re an x-ray technician you’re involved in taking x-rays 20 times a day for years or even decades on end so even if each one carries negligible risk, over time it can add up. In that circumstance being really careful about each individual exposure makes sense, even though the actual additional risk from a single slightly higher individual exposure is essentially zero.
Quote Comment
March 14th, 2010 at 11:03 pm
Dr. Buzzo — Am wondering what your thoughts are about thermography (using infrared camera and software) to identify areas of increased heat as an adjunct to mammograms. If cancerous tumors require increased blood flow to feed it and that increased blood flow would be indicated by an increased “heat print” compared to surrounding tissue, would it not follow that a change in physiology would precede a change in structure (tissue)? If that is true, then a thermogram could alert a health professional to biopsy a tissue before a tumor formed (which is only what a mammogram is capable of doing – finding it after the fact).
Not to mention, the thermogram is non-iradiative and causes no discomfort to the patient.
Here is one site for your review: http://www.thermology.com/
Quote Comment
March 14th, 2010 at 11:28 pm
DocForesight said:
To be honest, I do not feel qualified to really comment on the validity of that. I do know that the use of thermography has at least some medical value, but I don’t really know enough to know how much or for what applications.
Quote Comment
March 15th, 2010 at 3:56 am
There is one point that is too easy to forget: medecine is never an absolute science, but rather a risk/benefit ratio.
When I worked for medical imaging, we had a presentation that basically said:
-Assume LNT (worst case), about 0.01% of our patients may suffer radiation-induced complications (but nobody will ever know which ones because such a small number is statistically unsignificant)
-Our system allows to switch from high mortality open chest surgery (up to 1/3rd of patients used to die on the table during emergency coronary bypass) to lighter intra-vascular surgery, saving hundred of lifes now for every person that may die in the future because of our tool.
-Conclusion: Using as little radiation as needed is a noble goal, but never to the point of reducing the quality of the information provided to the surgeon.
Quote Comment
March 15th, 2010 at 1:35 pm
Shouldn’t it follow: The linear non-threshold is a reasonable inference? So, those who doubt it are the ones who have the burden of proof, not those who are its proponent. For example, according to classic physics, the momentum is the square of the velocity. It’s not necessary to prove that this property is true at 5 to 10 mph, we can assume it is true because we know it is true at 20, 30 and 100 mph. If I came along and said, “well, we don’t know, maybe momentum is not the square of speed above 100 mph or below 5 mph”, I would be laughed off.
But somehow, people make exactly this assertion with radiation!
By the way, green energy continues progress…
http://greeninc.blogs.nytimes.com/2010/03/15/nevada-wind-turbine-factory-to-create-1000-jobs-backers-say/
Quote Comment
March 15th, 2010 at 2:03 pm
Bruce said:
The carcinogenic risks of radiation that are observed at high doses be extrapolated to low doses, is what is being asserted with the LNT model. Unlike the laws of motion from classical physics, it has never been proven so your analogy falls apart right there.
What we do know is that there is not a linear dose response for other types of radiation, the best example being sunlight. The human body evolved being exposed to sunlight, and while too much exposure can and does cause tissue trauma, and death can occur in extreme exposures, low doses are not only provably harmless, but are in fact needed to maintaine good health.
We also know that humans evolved being exposed to a certain level of background ionizing radiation, and while it may be premature to say that we need some to stay healthy, claiming that there is a quantifiable risk to DNA down to zero is demonstrably absurd.
Plus there is a vast amount of data accumulating that even moderate exposures are not as harmful as originally thought, thus as consequence the LNT hypothesis is demonstrably wrong.
Quote Comment
March 15th, 2010 at 10:47 pm
Bruce said:
While I am tempted to point out some flaws in this statement and refute the infrence that this represents a net positive impact, I’ve decided not to dignify such comments with a response when they are placed on a post that is entirely off topic when there are plenty of perfectly on-topic posts that it could be made on.
Bruce said:
Actually, a while back there was this guy who proposed that the laws of motion that work well at 5 and 10 mph and even 100 mph don’t work so well around 186,000 miles per second – or for that matter, even half of that speed.
He also proposed that the effect he described also existed at 5 mph and 100 mph but was too small to be measured by instruments of the time and was generally insignificant.
Maybe you’ve heard of him? His name was Albert Einstein.
Oh, there were some other dudes who said that the classical laws don’t work at really really small scales. Maybe you’ve heard of them? Max Planck, Werner Heisenberg, Erwin Schrödinger, Niels Bohr
Oh, also, they weren’t laughed off. They generally were seen as being brilliant.
Quote Comment
March 16th, 2010 at 3:58 am
Bruce, I would totally agree with you if you were considering the effects of radiation on an inorganic target, but people are living things, that have protection (primarily outer skin, already dead tissues) and repair devices.
The LNT is assuming these protections are useless.
Also, if you want to stay on purely mechanical ground, it would mean that I will surely break someone’s arm by shaking it every day for enough years, after all, I will deliver quite a lot of energy (and it’s not only because the person is living, even inorganic materials have to some extend an elesticity range in which they can dissipate absorbed mecanical energy without damage).
Laws of energy are linear, but try to find something else that momentum for which there is no theshold to permanent effect, that would be a very interesting exercise.
Quote Comment
March 16th, 2010 at 4:46 am
Bruce said:
Complex systems do not follow the same linear principles as fundamentally basic ones.
You can analyse two objects colliding based on speed and mass and be confident of the outcome in terms of momentum based on a dozen or so rules (mainly involving infinite stiffness, zero friction etc). Apply this to Buzz0’s analogy of a person jumping off a building and you will find that a linear relationship of collision momentum and fatality rate is clearly a pile of trash. People have survived jumps from planes without parachutes (very rarely, I accept) and no-one dies from stepping off their doorstep (unless they trip over, but based on CoG that’d be a fall of more than a foot and thy’d have to have a pre-existing medical problem).
Oh, and SBuzz0: given you cited the wonderful bastion of British media brilliance that is The Daily Mail, I feel you might appreciate the following website, which provides a summary of their superb analysis of the various factors affecting your risk of contracting a cancer:
http://kill-or-cure.heroku.com/
Next time you’re wondering what to eat for dinner, maybe you should check which meal causes cancer and which cures…
Quote Comment
March 16th, 2010 at 9:59 am
DocForesight said:
I’m not a doctor, but I do recall Orac (Respectful Insolence) commenting a little on thermography. He’s a surgical oncologist specializing in breast cancer. If I recall correctly, he was conservative on the subject, so I think it’s probably premature to tout thermography as superior to mammogram or MRI, but the jury is still out on that one. He’s had some words on the subject of MRI, too, which is becoming more popular as an alternative to mammogram. One of his big concerns is the rate of false positives. Biopsies are not benign processes; they can be painful for quite a while after, and can result in nasty complications. So false positives are a serious problem with respect to medical imaging and breast cancer.
Also, I don’t think it would make any sense whatsoever to biopsy before a tumor formed; if there’s no tumor, then there is nothing for a pathologist to find in the biopsy sample.
Quote Comment
March 16th, 2010 at 12:28 pm
Franck said:
It’s a facinating subject, and I agree that it’s not simply linear. One of the issues that I remember reading has to do with single strand versus double strand DNA damage. Since DNA has an internal repair mechanism, where each half of the DNA is capable of regenerating the other, it’s believed that a double strand break is worse than twice as bad as a single strand break. Single strand damage is almost always reparable but double strand damage may not be.
This is important, because the probability of a double strand damaging event is not linear. It increases as ionization increases, such that if a cell is hit by multiple photons in a short period of time, there’s a much greater likelihood of double strand damage.
Maybe a good example as an analogy is error correction in other areas. I admit I’m not an expert on this one, but I’ll give it a shot. When you watch satellite TV or something like that, where you’re dealing with long distance signaling, some of the bits are lost in the transmission. Every frame has a small portion missing. This is okay, however, because there’s error correction in the signal. The signal contains some extra data that lets the receiver do some basic mathematics to detect which parts are probably an error and to correct them. It works perfectly up to a point. If you are missing, lets say, 1% of the data, it can all be recovered through the error correction process. Maybe it could be if you’re missing 2%, but it depends on how the error correction is encoded and how much of it there is. At some point too much of the data is missing for the error correction to fill in the gaps and when that happens, you get problems in the picture.
We know our DNA has error correction. Actually, software developers have even looked at DNA research as a guide for developing new error-resistant means of storing data efficiently.
Another point: Based on the simplistic interpretation of the theory behind damage from radiation and the LNT hypothysis, the more opportunity there is for a particle or photon to cause damage and the more cells there are to be damaged, the greater the probability of cancer. More cells mean more places for something to go wrong.
Therefore, in theory, if we expose a mouse, a rat, a cat, a human and an elephant to the same radiative flux, the mouse will have the least probability of cancer, the elephant the most and the rat, cat and human are in between, in order of size. Most of the radiation passes through the mouse without ever causing ionization, but the elephant has many cells and is big, so absorbs more radiation.
The problem: experimental results don’t indicate any greater increase in risk with the size of the organism. The implication from this is that viewing each cell event in isolation is not an accurate way of representing the total damage.
Quote Comment
March 18th, 2010 at 11:54 am
The LNT scare has deprived the human race of at least 4 decades of cheap power & probably of atomic spaceships.
All because it has been politically useful.
Any theory for which proponents attempt to produvce no evidence is, by definition, not testable & therefore not part of science. When the testing has been done & repeatedly proven the opposite theory, hormesis, to be true & LNT [proponents simply refuse to acknowl;edge the evidence that is much worse than simply not being science.
Those US states with high levels of background radiation generally have lower cancer rates; the negative relation between radon in homes & lung cancer has been repeatedly demonstrated, most thoroughly by Professor Cohen http://www.world-nuclear.org/sym/1998/cohen.htm ; the hormesis effect on plants, microbes & lab animals (which unlike human experiment is ethical) have, repeatedly, for over a century shown the hormesis effect & are not now disputed; health records of nuclear reactor workers, nuclear naval dockyard workers & UK radiologists have all shown improved health; a Taiwanese apartment complex, built with steel contaminated by Co60 & which is as close to a controlled human experiment as possible, showed a cancer reduction of 96.4% compared to the population at large http://www.jpands.org/vol9no1/chen.pdf ; there are parts of the world like Kerala in India where natural background radiation is anything up to 200 mSv a year, 13 times the official safe limit yet no adverse effects have been noticed over thousands of year; in the disgraceful & quite unscientific decision in 1964 to quietly euthanase a herd of cows exposed to a bomb test in 1946 having achieved extreme longevity; in the failure of Chernobyl to produce any statistical increase in long term cancers let alone the up to 250,000 predicted.
See also http://a-place-to-stand.blogspot.com/search/label/Hormesis
Quote Comment