Radiation from medical imaging and cancer risk

ResearchBlogging.orgScience-based medicine consists of a balancing of risks and benefits for various interventions. This is sometimes a difficult topic for the lay public to understand, and sometimes physicians even forget it. My anecdotal experience suggests that probably surgeons are usually more aware of this basic fact because our interventions generally involve taking sharp objects to people’s bodies and using steel to remove or rearrange parts of people’s anatomy for (hopefully) therapeutic effect. Ditto oncologists, who prescribe highly toxic substances to treat cancer, the idea being that these substances are more toxic to the cancer than they are to the patient. Often they are only marginally more toxic to the cancer than to the patient. However, if there’s one area where even physicians tend to forget that there is potential risk involved, it’s the area of diagnostic tests, in particular radiological diagnostic tests, such as X-rays, fluoroscopy, computed tomography (CT) scans, and the variety of ever more powerful diagnostic studies that have proliferated over since CT scans first entered medical practice in the 1970s. Since then, the crude images that the first CT scans produced have evolved, thanks to technology and ever greater computing power, to breathtaking three dimensional-views of the internal organs. Indeed, just since I finished medical school back in the late 1980s, I’m continually amazed at what these new imaging modalities can accomplish.

The downside of these imaging modalities is that most of them require the use of X-rays to produce their images. True, over the last 15 years or so MRI, which uses very strong magnetic fields and radiofrequency radiation rather than ionizing radiation to produce its images, has become increasingly prevalent. MRI is great because it produces more contrast between different kinds of soft tissue than CT scans do. However, CT tends to be superior for examining calcified organs, such as bone. (The breast surgeon in me notes that breast MRI is pretty much useless for detecting microcalcifications, an important possible indicator for cancer.) Also, MRI scans require a prolonged period of laying still in a very tight tube, which is a problem for patients with any degree of claustrophobia, although “open” MRIs are becoming increasingly available. More importantly for the quality of images, because they require a patient to lie more still than a CT, MRIs tend to be prone to more motion artifacts, which is perhaps why CT is more frequently used to image the abdomen other than large solid organs such as the liver. The point is that, although MRI is becoming more prevalent, CT scans aren’t going away any time soon. They have different strengths and weaknesses as imaging modalities and are therefore best suited for different, albeit overlapping, sets of indications.

Even so, it’s pretty amazing to consider how much these imaging modalities have changed medical practice in the last three decades. Before CT, surgeons often did exploratory surgery to diagnose a problem, often not knowing what they would find. They had to be ready for almost anything, and there were frequent surprises. (Some older surgeons lament that this has taken some of the excitement out of surgery, but there’s little doubt it’s better for the patients.) Another area where surgery used to be done routinely was in the staging of Hodgkin’s lymphoma. Patients underwent staging laparotomy, where the surgeon in essence carefully explored the abdomen, removed the spleen, and took biopsies of multiple areas in order to define precisely the extent of intraabdominal disease. Based on the results, the stage would be determined and therapy chosen. However, over the last 25 years or so, fewer and fewer of these have been done, thanks to better CT imaging and evolving practice in which more and more Hodgkin’s lymphoma patients receive chemotherapy. Indeed, during my residency I can only recall doing one or two staging laparotomies.

While CT imaging has revolutionized surgery and medicine, it is not entirely benign. Often, it requires the injection of intravenous contrast agents that can damage the kidneys and cause allergic reactions, occasionally life-threatening. Pretty much every physician is aware of these risks. Less acknowledged is the risk from the ionizing radiation from such tests, and physicians tend to downplay the risks from radiation. One exception is pediatrics, because it’s long been known that children are more sensitive to the effects of radiation than adults are, and they have much more time left in their lives for potential radiation-induced cancers to make themselves known. That is why pediatricians tend to be more judicious about the use of CT scans. In any case, by and large, CT scans require far more radiation than other imaging modalities. This concern was again brought to the fore last week by two studies recently published in the Archives of Internal Medicine, along with an accompanying editorial1,2,3. Together, these studies suggest that far more radiation is used in some CTs than is necessary and that there may be far more radiation-induced cancers due to medical tests than we would like to acknowledge. Taken together with another review article in the New England Journal of Medicine from a couple of years ago, they should make us as physicians think more carefully about how we use diagnostic imaging studies.

The NEJM review4 is useful because it gives the background in terms of typical doses of radiation for various imaging studies:


And for the huge increase in the number of CT scans being done in the U.S. every year:


That’s over 60 million CT scans performed in the U.S. in 2006. More recent data shows that 70 million scans were performed in 20072. and, for example, a typical CT scan of the chest results in an absorbed radiation dose 100-fold higher than a typical two-view PA and lateral chest X-ray. Moreover, as Smith-Bindman et al2 point out:

Exposure to ionizing radiation is of concern because evidence has linked exposure to low-level ionizing radiation at doses used in medical imaging to the development of cancer. The National Academy of Sciences’ National Research Council comprehensively reviewed biological and epidemiological data related to health risks from exposure to ionizing radiation, recently published as the Biological Effects of Ionizing Radiation (BEIR) VII Phase 2 report.7 The epidemiologic data described atomic bomb survivors, populations who lived near nuclear facilities during accidental releases of radioactive materials such as Chernobyl, workers with occupational exposures, and populations who received exposures from diagnostic and therapeutic medical studies. Radiation doses associated with commonly used CT examinations resemble doses received by individuals in whom an increased risk of cancer was documented. For example, an increased risk of cancer has been identified among long-term survivors of the Hiroshima and Nagasaki atomic bombs, who received exposures of 10 to 100 milli-sieverts (mSv).8-11 A single CT scan can deliver an equivalent radiation exposure,12 and patients may receive multiple CT scans over time.13

They then observe that few studies have tried to quantify rigorously the typical real-world doses of radiation received in hospitals due to CT scanning. Most studies, other than for CT coronary angiography, have used phantoms rather than patients. So Smith-Bindman et al2 looked at imaging studies in four San Francisco-area hospitals, one of which was at UCSF, and used a method called the “effective dose” to quantify radiation exposure in consecutive studies. What was most shocking is what they found regarding the variability in radiation exposures both within and between institutions even for the same test. Indeed, they noted a mean 13-fold variation between the highest and lowest radiation dose for each study type:


This particular plot type, known as box-and-whiskers, shows the 25th to 75th percentile range in the boxes, while the bars show the range between the highest and lowest values, with the median value being represented by the dots. The investigators then estimated the increased excess risk of cancer for these radiation doses and concluded:

Among 40-year-old women, 1 cancer would occur among 8105 patients who underwent a routine head CT scan (IQR, 1 in 6110 to 1 in 9500). For a 60-year-old woman, the risks were substantially lower and varied from approximately 1 in 420 examinations for CT coronary angiography (IQR, 1 in 370 to 1 in 640) to 1 in 12 250 examinations for a routine head CT scan (IQR, 1 in 9230 to 1 in 14 360). For a 20-year-old woman, the risks were substantially higher and varied from approximately 1 in 150 examinations for CT coronary angiography (IQR, 1 in 130 to 1 in 230) to 1 in 4360 examinations for a routine head CT scan (IQR, 1 in 3290 to 1 in 5110).

These are not insignificant risks. It should be noted, however, that this study has several weaknesses. The biggest weakness is that the cohort studied (1,119 patients) was not large enough to identify possible reasons why the dose of radiation varied so much for even the same tests, including experience of the technologist, physician availability to check the studies and determine the need for additional imaging, geographic variation, imaging algorithms available or used, and patient factors (such as weight). The authors point out that far more standardization is required and that studies are needed to figure out why there may be such variation in radiation dose.

The second study drives home the point that radiation from CT scans can increase cancer risk by using different methodology. Berrington et al1 started with risk models based on National Research Council’s “Biological Effects of Ionizing Radiation” report and organ-specific radiation doses derived from a national survey were used to estimate age-specific cancer risks for each scan type and then combined these models with age- and sex-specific scan frequencies obtained from insurance claims data and surveys. Using a Monte Carlo simulation, they then estimated the number of excess cancers due to radiation from CT scanning. Their conclusions:

Overall, we estimated that approximately 29,000 (95% UL, 15 000-45 000) future cancers could be related to CT scans performed in the US in 2007. The largest contributions were from scans of the abdomen and pelvis (n = 14,000) (95% UL, 6,900-25,000), chest (n = 4100) (95% UL, 1,900-8,100), and head (n = 4000) (95% UL, 1,100-8,700), as well as from chest CT angiography (n = 2,700) (95% UL, 1,300-5,000). One-third of the projected cancers were due to scans performed at the ages of 35 to 54 years compared with 15% due to scans performed at ages younger than 18 years, and 66% were in females.

This graph tells the tale:


The black bars are for men; the white for women. Women tend to have a higher sensitivity to the effects of radiation in cancer production.

One thing that is very important is to put these figures in perspective. 29,000 is a huge number, but compared to the number of new cancer cases every year (estimated to be 1.5 million in 2009, down from earlier years). Indeed, Berrington et al1 estimate that their study suggests that approximately 1% to 3% of cancers in any given year can be attributed to past CT use. Another thing that is very important is that these results are due to a simulation, which is very dependent on the values inputted and the assumptions made in constructing the simulation. The estimates of the number of CT scans. For example, for solid tumors the assumption was a five-year lag period and a linear dose-response model. I’m not sure how valid that assumption for lag time is, given that there are quite a few tumors with longer lag periods after radiation exposure. Still, overall, this study likely represents a fairly good estimate of how many additional cancers there are due to CT scanning, but it is just that, an estimate. It also does not provide any information to tell us which cancers were actually caused by radiation from a CT scan. Neither of these studies do; they’re both population-based and look at aggregate statistics. Even so, the possibility that as many as 3% of adult cancers might be due to radiation from medical imaging studies is a problem that should sober even the most gung ho advocate of using such studies, particularly considering that the risk tends to be higher in younger people.

All of this brings us back to what I started this post with: All of medicine is a balancing of risks versus benefits. One reason I was so disturbed by the proliferation of whole-body imaging studies being marketed by unscrupulous companies on a cash basis is because, in an asymptomatic patient, the risks from radiation from such studies on average probably outweigh any conceivable benefit, especially if we take the risks of false positives leading to invasive tests such as biopsies into account. Still, there is no doubt that CT scans are highly beneficial when it comes to diagnosing disease and, these days, to guiding physicians in doing less invasive needle biopsies for diagnosis where before a surgical biopsy might have been required. That leaves the question: What to do with these results?

One approach to reducing radiation exposure from medical imaging would be to be to try to standardize imaging studies more, so that the dose of radiation for each one varies less, and, even more importantly, to find ways to decrease the dose of radiation for each test without sacrificing image quality or diagnostic sensitivity or specificity. The authors of both studies agreed on these tactics. However, far more difficult will be tactics designed to change physician behavior.

Clearly, the first thing we as a profession should do is to make ourselves very aware that a CT scan (or any scan involving a significant radiation dose) is not an entirely benign thing. We sometimes do treat them that way, and this must stop. There are several strategies to reduce the risk from these imaging studies. One obvious one, of course, is to order fewer studies and to stop ordering them for questionable indications. Critical to this approach would be better data and studies that help us clearly define when such tests are appropriate and indicated; i.e., a more rigorous application of science-based medicine to medical imaging. Sadly, this is not as much the case now as it should be, as the author of the accompanying editorial3, Dr. Rita Redberg, points out:

In addition, it is certain that a significant number of CT scans are not appropriate. A recent Government Accountability Office report on medical imaging, for example, found an 8-fold variation between states on expenditures for in-office medical imaging; given the lack of data indicating that patients do better in states with more imaging and given the highly profitable nature of diagnostic imaging, the wide variation suggests that there may be significant overuse in parts of the country.4 For example, a pilot study found that only 66% of nuclear scans were appropriate using American College of Cardiology criteria—the remainder were inappropriate or uncertain.5

Indeed, medical imaging is highly profitable. Moreover, sometimes laziness rules. It is easier just to order a CT scan than to use more mundane methods of trying to figure out what’s wrong with a patient, and the current malpractice climate often leads to physicians practicing “defensive” medicine, part of which may involve, for example, ordering a CT scan for a patient with abdominal pain “just in case” even when it’s known that the diagnostic yield is likely to be very low. At the risk of getting myself in trouble, I’ll point out that patients, too, bear some of the blame, just as they do for the overuse of antibiotics that leads to resistant organisms. Some just won’t be reassured without an imaging study; although they might be if they were more carefully informed of the increased risk of cancer from medical imaging studies. In any case, all of these factors combine to drive the explosion in CT imaging, which has increased faster than evidence of its benefit.

Finally, it is critical to remember that, for individual patients, the risk of any single imaging study is pretty low, and the potential benefit, when the study is ordered appropriately according to science- and evidence-based guidelines, will almost certainly far outweigh the slightly increased risk of cancer. For example, if you’re in the emergency room with severe chest pain, the last thing you should be worrying about is the radiation you’ll receive from a cardiac catheterization and angioplasty. Even if your chance of developing cancer from the radiation is increased by 1 in 100, that pales in comparison to your chance of dying now if your blocked coronary artery isn’t identified and opened up. If you’ve been in a car crash and might have a lacerated spleen or liver that needs repair or might have a subdural hematoma that could squish your brain against the inside of your skull, the risk from the radiation due to the CT scans that would diagnose these problems is nothing compared to your risk of death or serious disability now.

The problem is that the indications for CT scans have expanded to the point where they are often done even when they don’t provide information that will change the course of management for a patient. For instance, it used to be that general surgeons (of which I still count myself one) could diagnose acute appendicitis in a young male (who doesn’t have female reproductive organs, disorders of which can be confused with appendicitis) by history and physical exam alone and be highly accurate doing so. Yet these days, even young men with right lower abdominal pain get a CT scan that tells the surgeon that, yes, they have acute appendicitis before going to the OR. Many patients with acute peritonitis don’t need a CT scan for a surgeon to know that they need an operation. A very ill-appearing febrile patient lying perfectly still because the slightest movement causes him intense abdominal pain doesn’t need a CT scan; most of the time, he needs a trip to the OR as soon as possible to fix whatever intraabdominal catastrophe is going on. (This reminds me of a surgical aphorism that some attendings used to use to tweak residents examining a patient with peritonitis, which went, “What are you waiting for? Even the janitor can see that this patient needs an operation!”) Unfortunately, these days it seems that virtually all patients presenting to the ER with abdominal pain get a CT scan. As Dr. Redberg points out, “more and more often patients go directly from the emergency department to the CT scanner even before they are seen by a physician or brought to their hospital room.” This approach is all too easy and seductive, and all too often even general surgeons have allowed it to become the rule rather than the exception because it’s far easier to wait for the CT scan than to get out of bed to determine if a patient really needs a CT scan. Indeed, back when I still did general surgery call and chastised an ER doc for ordering a CT scan that I didn’t consider indicated, the response was that all the surgeons there wanted a CT before he even called them about a patient with abdominal pain and would get irate if he didn’t have one. This happened over ten years ago.

The bottom line is that, when the test is indicated based on guidelines constructed using science and evidence, the benefits of doing a CT scan or other medical imaging procedures requiring similar amounts of radiation outweigh the risks. The problem is that all too often these scans are not ordered using science-based guidelines and in all too many cases the evidence is not clear that doing a CT scan will improve patient outcomes. Clearly, we require more and better studies that define when the benefit of doing a CT scan outweighs the risk from the radiation. In the meantime, physicians and patients need to be aware of data like these regarding the risk of cancer due to radiation from CT scans, and physicians need to exercise some restraint and–dare I say?–clinical judgment when deciding to order these tests.


The NCI Factsheet on Computed Tomography: Questions and Answers


1. Berrington de Gonzalez, A., Mahesh, M., Kim, K., Bhargavan, M., Lewis, R., Mettler, F., & Land, C. (2009). Projected Cancer Risks From Computed Tomographic Scans Performed in the United States in 2007 Archives of Internal Medicine, 169 (22), 2071-2077 DOI: 10.1001/archinternmed.2009.440
2. Smith-Bindman, R., Lipson, J., Marcus, R., Kim, K., Mahesh, M., Gould, R., Berrington de Gonzalez, A., & Miglioretti, D. (2009). Radiation Dose Associated With Common Computed Tomography Examinations and the Associated Lifetime Attributable Risk of Cancer Archives of Internal Medicine, 169 (22), 2078-2086 DOI: 10.1001/archinternmed.2009.427
3. Redberg RF (2009). Cancer risks and radiation exposure from computed tomographic scans: how can we be sure that the benefits outweigh the risks? Archives of internal medicine, 169 (22), 2049-50 PMID: 20008685
4. Brenner DJ, & Hall EJ (2007). Computed tomography–an increasing source of radiation exposure. The New England journal of medicine, 357 (22), 2277-84 PMID: 18046031

Posted in: Cancer, Public Health

Leave a Comment (37) ↓

37 thoughts on “Radiation from medical imaging and cancer risk

  1. kirkmc says:

    What I don’t understand, and would like to know, is what is the real risk of a CT scan? It’s not clear to me, from both the original article and your presentation, what the risk is. The article you link to says:

    “radiation from the more than 70 million CT scans performed in the United States in 2007 will ultimately cause some 29,000 cases of cancer. Researchers said that could lead to 15,000 deaths.”

    What does that mean? Does that mean 29,000 people will get cancer _each year_ from _a single_ CT scan? Or are those cancers all (or mostly) from multiple CT scans, or accumulated radiation over time?

    Because if it’s the former, then CT scanners should most likely be eliminated. If they are going to kill 15,000 people every year, they’re obviously extremely dangerous.

  2. wales says:,0,406924.story

    Now accidental overdosing of radiation from CT brain scans is a growing problem (150,000 performed in the US annually). If this is a problem at a “good” hospital like Cedars-Sinai, what about other hospitals? This article suggests the problem may be nationwide.

    Wonder what this does to the cancer risk numbers from the studies cited……and also what about “under the radar” overdosing that might not be extreme enough to cause the obvious symptoms such as hair loss mentioned in the article but extreme enough to increase cancer risk even more than a correct dosage of radiation?

  3. windriven says:


    There are on the order of 70 million CT scans performed in this country each year. Many of these scans are critically important and life saving.

    29,000 is an aggregate approximation. Dr. Gorski listed the actual predicted rates for several different CT targets. It isn’t a case of one size fitting all.

    If the 29,000 is accurate then that implies that about .04% – 4 cancers in 10,000 scans – will occur. Will those 10,000 CT scans save more than 4 lives? Book it. So I don’t think your proscription of CT is defensible.

    The entire point of Dr. Gorski’s post was that CT scans are a powerful diagnostic tool but that use of that tool carried risks that can not and should not be ignored; that the modality should be used with care rather than ordered willy nilly.

  4. David Gorski says:

    If the 29,000 is accurate then that implies that about .04% – 4 cancers in 10,000 scans – will occur. Will those 10,000 CT scans save more than 4 lives? Book it. So I don’t think your proscription of CT is defensible.

    True for adults; however, it should be noted that in young children the risk is approximately 5- to 10-fold higher, if memory serves me correctly.

  5. windriven says:

    Dr. Gorski-
    Even at 10-fold if my child presented with severe symptoms of uncertain etiology and a CT scan was diagnostically appropriate, I for one wouldn’t hesitate. 4 in 1,000 is still pretty good odds.

    But your point is well taken. Is that the sound of radiologists howling that I hear?

  6. kirkmc says:

    It’s a bit hard to understand that a medical test that kills 15,000 people a year is still acceptable.

    I read this:

    They project that 1 in 270 women who undergo a CT scan of the heart’s blood vessels at age 40 will develop cancer from the procedure compared with 1 in 600 men.


    That seems to be a pretty high risk for a specific test.

    I’m just saying… I know that CT scans are useful, and valuable. But with this many cancers and deaths, it makes me wonder the next time a doctor orders a CT scan for me whether I’ll say yes to it.

    BTW, I also read that CT scans of different areas give very large differences in the amount of radiation. I’ve had several CT scans of my brain, and IIRC brain scans use much less radiation than others. Is that correct?

  7. windriven says:


    If the test killed 15,000 people a year and saved no one, I would agree with you wholeheartedly. But that isn’t the case. Unfortunately, I don’t have (and doubt there exists) studies that indicate how many lives are saved by CT scans each year. Let’s say for argument’s sake that only 1 in 100 CT scans provides information that ultimately saves a life. That is 700,000 lives saved per year. 46 lives would theoretically be saved for each life lost. Is that an acceptable ratio? I would argue that it is.

    I am not competent to address your technical question on radiation dosage in brain scans. A patient information site funded by the professional radiologist community has this to say: “The effective radiation dose from this procedure is about 1 to 2 mSv, which is about the same as the average person receives from background radiation in four to eight months.” (

  8. Zoe237 says:

    Thanks for this. 15,000 deaths seems very high from CT scans in one year.

  9. kirkmc says:

    “Let’s say for argument’s sake that only 1 in 100 CT scans provides information that ultimately saves a life. That is 700,000 lives saved per year. 46 lives would theoretically be saved for each life lost. Is that an acceptable ratio? I would argue that it is.”

    Seriously? And if it’s 1 in 200, then 23 lives are saved for one lost? And so on?

    My guess is that only a small number of CT scans actually lead to lives saved; I’d think that a lot of them are for things like arthritis or other musculo-skeletal problems… But I don’t know how often they’re done. I know I’ve had several in my lifetime, none of which provided any info that “saved” my life.

  10. Harriet Hall says:

    The question is not so much how many lives are saved or lost with current practices, but whether we can improve the risk/benefit ratio with better practices. We can.

  11. Harriet Hall says:

    I’d like to see the numbers put into perspective with other radiation risks. We are constantly bombarded with cosmic rays and radioactive isotopes of common elements make up part of our bodies. How many lives are lost from the increased background radiation of living at high altitude, in brick or stone houses, flying in airplanes, older coal-fired power plants, radon, smoking, and other sources of radiation?

    I have read that smoking cigarettes exposes lung tissue to 200 times the background level of radiation (from Polonium 210).

    I have read that 80% of human radiation exposure is from background sources and only 20% is from medical radiation. Is that accurate? If not, what are the true numbers?

  12. windriven says:

    @ Dr. Hall

    Great points, especially the environmental v medical radiation risks.

    I would though point out that lives saved and lost is an inescapable part of the risk/benefit analysis. And of course Gorski’s point was exactly the importance of carefully weighing the risks against potential benefits: “and physicians need to exercise some restraint and–dare I say?–clinical judgment when deciding to order these tests.”

  13. windriven says:

    @ Zoe-

    “Thanks for this. 15,000 deaths seems very high from CT scans in one year.”

    Agreed, but Dr. Gorski’s post did not address hard numbers regarding the benefits of these tests, probably because quality information doesn’t exist. As Dr. Gorski and Dr. Hall point out, the technology must be evaluated not only in terms of risk but in terms of benefit as well.

    @ kirkmc

    “Seriously? And if it’s 1 in 200, then 23 lives are saved for one lost? And so on?”

    Yes, seriously. And yes, I made clear earlier in this thread that 1:100 was a plug based on nothing. The point was to underscore Gorski’s point that the benefits of CT come with some serious risks and someone needs to balance those risks and benefits. It seems to me that an informed dialog between patient and physician is where that calculus is best made. But what about emergent care when the patient is unable to participate in the decision?

  14. David Gorski says:

    But what about emergent care when the patient is unable to participate in the decision?

    In cases of emergent care where the patient can’t participate in any decision making and there is no one there to make decisions for the patient, the default assumption is that the patient wants to have everything done, and we do everything. In trauma, we don’t wait for consent or to find a family member, for instance, to whip a bleeding patient off to surgery to save his life. Ditto an unconscious person who appears to be in the middle of a stroke who could benefit from TPA.

  15. Jeff says:

    Could some protection be provided by administering multiple antioxidants before using radiation? Several articles claim lipoic acid, which is both water and fat-soluble provides some protection without impeding the diagnostic process.

  16. windriven says:


    Please include citations when you make a statement like: “Several articles claim lipoic acid, which is both water and fat-soluble provides some protection…”

    Claims were once made for DMSO as being protective against radiation but that didn’t seem to pan out.

  17. wales says:

    NPR also ran articles regarding the CT scan overdosing problem, quoting an FDA official. “A regular CT scan to the brain is the equivalent of about 100 chest X-rays, says Dr. Jeffrey Shuren, acting director of the FDA’s Center for Devices and Radiological Health. By contrast, a perfusion CT scan of the brain is equal to several hundred chest x-rays.

    Patients at the four hospitals who received excess radiation were exposed to the equivalent of several thousand X-rays instead — three to eight times the expected radiation dose.”

  18. windriven says:


    A search on PubMed shows several articles on the radiation protective effects of lipoic acid. Most of these studies in mice were done at very high energy levels more associated with cosmic radiation than diagnostic CTs. Still, interesting stuff.

  19. wales says:

    NPR also ran articles regarding the CT scan overdosing problem. “A regular CT scan to the brain is the equivalent of about 100 chest X-rays, says Dr. Jeffrey Shuren, acting director of the FDA’s Center for Devices and Radiological Health. By contrast, a perfusion CT scan of the brain is equal to several hundred chest x-rays.

    Patients at the four hospitals who received excess radiation were exposed to the equivalent of several thousand X-rays instead — three to eight times the expected radiation dose.”

  20. Scott says:


    In all fairness, Jeff did include links to a pair of journal abstracts (one regarding lipoic acid, one on the multi-antioxidant angle). Neither was anything that seemed anywhere close to showing clinical efficacy, however. “Potentially interesting” does seem to be the best description.

  21. Calli Arcale says:

    wales — the overdosing problem is a separate issue, IMHO, though an extremely serious one. Dr Gorski is talking about the risks from CT even when the machines are working correctly. It goes without saying that when the machines are not working correctly (as happened at Cedars-Sinai) the risk is much greater, and I think that case is a great cautionary tale for doctors who order CT scans too casually. They need to remember that one risk they are asking of their patients is the risk of the equipment malfunctioning.

    As far as how can 15,000 deaths per year be acceptable — well, they’re not, but the thing is, it’s very hard to tie an increase in radiation to a specific cancer. We’re all exposed to radiation every minute of our lives. How do you know which cancers are resulting from which radiation?

    An interesting example: the Apollo astronauts were subjected to fairly high radiation doses as they passed outside the protective envelope of the Earth’s magnetosphere, and especially during their brief transits of the Van Allen Belts. (They actually saw sparkles caused by alpha particles exciting the cells of their retinas.) Out of 24 astronauts who travelled to the Moon (including the 12 who didn’t actually set foot on it), 2 have thus far died of cancer. Alan Shepard died of leukemia in 1998. Jack Swigert died of bone cancer in 1982. It goes without saying that their careers increased their lifetime cancer risk — but they could’ve died of cancer even if they’d taken jobs as accountants. How do we know Apollo actually caused their cancers? We don’t. It just doesn’t work like that unless you get such a massive dose that it’s obvious (as with the malfunctioning CT scanners that wales mentioned).

    I doubt that 15,000 people really die each year specifically because of CT scanners. It really just comes down to aggregate risk. If you get a CT scan, then fly aboard SpaceShipTwo during a solar storm, then get a chest x-ray, then go through a full-body x-ray machine at the airport, then take a vacation to Pripyat, Ukraine (the abandoned city near Chernobyl, where yes, you really can go as a tourist — there are organized tours), and then 30 years later you get cancer . . . . What caused your cancer? Many things raised your risk, but what tipped it over?

  22. antipodean says:

    By far the biggest risk factor for getting cancer is getting older. And we don’t have a risk mitigation stragety for that.

    The Apollo astronauts would also have been serving military personnel and so their exposure to multiple cancer causing agents that have nothing to do with the space programme may also have been high.

    Those are great papers though. Given our recent breast cancer screening discussions it gives another great example of why screening for disease is not usually a good idea. Breast cancer screening gets a great deal of research and thinking but CT scanning for everything doesn’t.

    The US practising clinicians might have a better idea of this than I could, but how much of the routine CT scanning is essentially tort-driven box checking? I wonder whether one might calculate the number of cancers caused by defensive medicine, for instance? Or patients demanding screening?

  23. windriven says:


    Thanks, you are correct. I’m not clever enough to embed links that way when I post to these threads. So I expect others to be similarly inept and I look for the usual URL. I’ll pay closer attention in the future.

  24. wales says:

    Calli, Not to belabor the point, but the overdosing issue is unrelated to equipment malfunction. Equipment malfunction is not what happened at Cedars. “Every time a patient receives a CT scan, a mundane array of numbers appears on a computer screen before a technician. The numbers include the radiation dose. ”

    “The CT machine in question performed several types of scans, each with its own set of computerized instructions, or protocols. To change the instructions for brain perfusion scans, the hospital had to bypass the protocol that came installed on the machine.”

    Here is yet another article on the subject.

  25. vasiln says:

    “Women tend to have a higher sensitivity to the effects of radiation in cancer production.”

    Is it that women have a higher sensitivity, or is it that they receive more imaging studies?

  26. halincoh says:

    Very well done. Thank you.

  27. kirkmc says:

    I agree with Dr Hall about the need to examine this “study” with other radiation figures. The CT study sounds a bit sensationalist, and it doesn’t seem that there’s any way to be sure that these cancers and deaths are fully imputable to CT scans.

  28. Calli Arcale says:

    Calli, Not to belabor the point, but the overdosing issue is unrelated to equipment malfunction.

    That was kind of the point I was trying (inelegantly) to make — that this thread is discussing how excessive CT scans can be dangerous even if the equipment is working perfectly. This thread was about that background risk, not the extra risk of malfunctioning equipment. But now I’ll address the point you’ve raised instead of the one in the OP.

    I think people will have a tendency to blow off what happened at Cedars-Sinai, on the basis that it was a technical problem and thus not reflective of normal situations. But it should give them pause, both because there is a background risk even when everything goes right that they’re disregarding, and because there is no systemic process preventing a repeat of the Cedars-Sinai incident.

    I like space analogies, so I’ll bring up another one.

    On STS-112, the Space Shuttle Atlantis flew to the International Space Station, delivering the S1 integrated truss segment. It seemed to be an entirely normal flight, but mission controllers had a series of heart attacks (metaphorically speaking) over an incident during first stage ascent: a large chunk of foam broke off of the External Tank and impacted one of the Solid Rocket Boosters just forward of the aft skirt. When the SRBs were recovered and inspected, they had quite a shock: a dent was visible where the foam had impacted, and it had missed a critical electronics box by inches. Had the box been hit, it would have crippled the booster’s thrust vectoring system, rendering the vehicle impossible to steer.

    To know how serious this is, you have to know one of the most sobering things about the Shuttle: during first stage ascent, there is no abort mode. Failures before SRB separation are not considered survivable. The range safety officer would have had to make the same horrible decision that was made January 28, 1986, and press the self-destruct button. But this time, it would have been with the Orbiter still attached. The RSO would’ve had to decide, in a split second, to kill the crew in order to save lives on the ground.

    But the box was missed. And now NASA had irrefutable evidence that the known foam-shedding problem wasn’t quite so benign — if large pieces were shed while the Orbiter was in the thick lower atmosphere, pieces could rapidly decelerate and thus have a high impact velocity when the vehicle slams into them.

    This was the only time a large piece of foam had been shed during the first minute *and* hit any part of the Space Shuttle stack. What did NASA do about it?

    In the end, nothing. It was essentially written off as a fluke (though with a lot more discussion than that; I’ve simplified the story a bit). Foam doesn’t usually shed like that; the risk can’t be adequately characterized with such little data; we gotta fly or we can’t get the ISS completed on schedule; etc. etc. Oh, and there’s this non-ISS mission, requiring the EDO pallet currently installed on Columbia, which has been pushed back for several years now. We want to refit Columbia for ISS utilization flights, so let’s hurry up and get that mission out of the way so we can get on with the ISS construction.

    The next flight was STS-113. Endeavour carried the P1 truss segment, sister to the one carried up on STS-112, and no significant foam shedding was observed. NASA officials congratulated themselves for not panicking over the STS-112 incident, decided it was an isolated event, and moved on.

    The next mission was STS-107, intended to be Columbia’s last flight with the EDO pallet, after which the pallet would be moved to Discovery, and Discovery’s docking system would be moved to Columbia. But we all know how that one ended. Foam shed again, during first-stage ascent. A piece believed to be from the bipod ramp area tore away early enough to be filmed by ground cameras in addition to the “rocketcam” on the ET. Cameras didn’t show exactly where it hit, but later analysis of the wreckage revealed that it punched a hole through the reinforced carbon-carbon carrier panels on the leading edge of the port wing. Ascent was otherwise normal, and the crew completed a highly successful mission. But their thermal protection system was critically compromised, and hot gasses entering the wing caused it to fail structurally over the American southwest. Hypersonic wind blasts finished the job, ripping the vehicle into thousands of pieces.

    No one should casually dismiss radiation overdoses from a CT machine, and there needs to be not only a serious, independent investigation into why that machine did the patients so much harm, but also I believe we need standardization of these things so that the same problem is not repeated anywhere else. It would be inexcusable to find the root cause and correct it only at Cedars-Sinai. That would be perhaps the worst outcome of this, and one that I think is actually quite likely.

  29. wales says:

    Thanks Calli. Again, Cedars was not a fluke. If you read the articles you’ll see it has happened in at least 4 Los Angeles hospitals and the FDA is investigating it as a possible nationwide problem.

  30. DonSelgin says:

    I’m interested in your views regarding CT scans for appendicitis.

    I went to the doctor mainly for some cold symptoms, but also reported some consistent pain in my lower right abdomen. She poked and prodded it and got a bigger reaction from me. However, I did not have a fever, had not been throwing up, etc.

    She sent me downtown for a CT scan to rule out appendicitis. Turns out, I did have it, and they directed to me to the hospital, where I subsequently had a successful appendectomy.

    In that case, was a CT scan justified? Or can a doctor tell by feeling around the appendix whether it is swollen or not, w/o having to resort to the CT?

    Just curious,
    Sean E.

  31. Calli Arcale says:

    Agreed. I think it’s one of the scarier medical stories of 2009. It has haunting shades of the Therac-25 scandal, in which a radiation therapy machine delivered lethal radiation doses to some patients. Perhaps scariest is that the investigation found that technicians were ignoring complaints due to overconfidence in the system. They believed it was failsafe, and so discounted reports that it was, in fact, doing harm. The major problems were all in the software, and so the story is often retold in software engineering coursework. It’s rather sobering to read about it and see how many classic software problems existed in something that was responsible for determining doses of radiation to be delivered to actual live human beings.

  32. David Gorski says:


    I was referring to the classic presentation of appendicitis in a young male. It sounds as though you had a more atypical presentation.

  33. Zoe237 says:

    Calli, you have a gift for the use of analogies in science writing.

    Anyone, there is a book called Flatlined by Guy Cliftord that explores the health care system. He is a neurosurgeon who claims that the overuse of diagnostic tests is due to our “pay per use” system and promotes a pay for outcomes system instead. I can foresee problems with the latter, but the overuse of medical technology seems to be a recurring theme on this site and others. Any thoughts? Anybody up for a book review? What is the solution?

  34. Mark P says:

    “Seriously? And if it’s 1 in 200, then 23 lives are saved for one lost? And so on?”

    Sigh. No, it would not be “23 lives saved for one lost”. It would be 23 lives saved for one extra case of cancer. Not all cancer is fatal.

    And there is a time lag to the start of the cancer. And another time lag to the end of the cancer.

    So a person with an unnecessary CT scan *might* die of cancer as a result. But it might be 25 years later.

  35. JMB says:

    Sorry for the late post. I ran across the article while searching for something else, and wanted to throw in my two cents worth.

    Most (I think probably all) states in the USA have a law requiring inspection of facilities using x ray equipment by a medical physicist. Even though there is plenty of documentation of the variability of exposure of patients that is kept by the state governments, it is pretty much neglected in terms of enforcement (no facility gets stopped from using the equipment when they are consistently overexposing patients to radiation). There are laws that if enforced would significantly reduce the variability in the amount of radiation exposure for a given exam. Some of those overexposures are due to equipment malfunctions. Some are due to errors in the use of the equipment. The published problems at Cedars-Sinai were due to errors in the use of the equipment, not malfunction.

    If a person wonders about their own radiation risk, there are medical physicists qualified to calculate individual radiation exposures. They will charge you. Background radiation does vary by geographic region, and the number of airline trips. For those with no more than 2 CT exams (depends on the type of exam), a few chest xrays or xrays of the extremities, and yearly mammograms, your lifetime background radiation exposure will be much higher than your lifetime medical exposure.

Comments are closed.