An analysis of which part of the body was exposed to radiation by ct scan and whether the cancer occured in the same area should be studied before concluding that ct scan increased the risk of cancer in these patients.

Nevertheless radiation is better avoided in children unless necessary.

My reaction is that we need a few more studies in this area to be 100% sure of this. But the advice on CT scan use being limited to situations with a definite clinical indication sounds very reasonable to me and should be de facto clinical practice.

Medical Exposures Rise 7-Fold

I do believe that the extreme increase in medical exposures is an alarm that confirms the need for efficient efforts to ensure appropriate use of medical imaging procedures! According to NCRP report No. 160, people in the US were exposed to more than SEVEN TIMES as much ionizing radiation from diagnostic medical procedures in 2006 than they were in the early 1980s. Needless to say, the largest part of this increase was from CT scans and nuclear cardiac scans. Interestingly, NCRP 160 DOES NOT attempt to address the benefit-risk analysis.

Exposure of the population of the United States to ionizing radiation in early 1980s and in 2006 (Source: NCRP report no. 160)

Kiran and Kalpana emphasise that only necessary examinations should be done in children. This is part of the problem, ie unnecessary examinations are probably still being done in children worldwide. These were estimated to be as high as one in three a few years ago. A contributor to this is scepticism about the linear no threshold model by many practitioners and some physicists. I think the value of the Australian study is that it provides data that lend some support to the conclusion that we have to be careful.

Smj and Lucien concentrate on issues around dose. I agree with Smj that the NCRP figures were at the time alarming and with Tan Lucien that in some practices now the doses are much lower. However for many years adult scanning factors were used for paediatric CT, so generations of children were subjected to much higher doses than were needed. Therefore we have an obligation to follow these historical cohorts given problematically high doses. I think this is one of the values of the study. The other is that it is real data (even with some problems in it). It is derived from CT scans and not atomic bomb survivors and is closer to the dose range of interest for radiology. Real data is better than extrapolation, and in radiation protection we have been over dependent on extrapolation. Sometimes we even seem to think that extrapolations from the LNT model is real data. Obviously it is not.

Well, I recall that warnings had been flagged already a few years ago, especially looking at the overall number of CT scans per year (which soared) for example in the US, and most of which was judged unnecessary.

Yes, Lorenzo, from about 2007 on there was increaasingly strong comment on both the extent of the growth in the use of radiology and on the extent to which its utilization was unjustified or unnecessary in individual cases. In Europe the law requires that each medical exposure be justified medically (or dentally). The available data show that this often isn't the case and the drive in departments where practice is good is two fold. A lot of work is going into low dose techniques both by manufacturers and by using good departmental protocols that are dose conscious. In addition in many places, there is now a much greater awareness of the need to reduce unjustified examinations by using appropriateness or acceptability guidelines for examinations. But there is still a long way to go with both.

This epidemiological study although provide indirect evidence of radiation effect, suffers from serious methodological problems. It is quite clear that most of the CT scans are done due to the presence of disease or symptoms and signs that patients had at that time. These patients may have immune system malfunction, congenital abnormality, autoimmune disease or any other disease. Presence of a disease or even a chemotherapy in childhood, increase the risk of other malignacies more than 20 times. So the authors conclusion would be correct if the two groups were clinically in a similar condition or the physicians in Australia ordered unnecessary CT scan in most of patients! Anyhow that data may be valid for brain tumors as significant number of brain CT scans in traumatic patients may be unnecessary. These CT images would have been saved many lives and the benefits of the CT in these patients are not considered as well. It would be valuable if the authors repeat analysis comparing patients who have done CT scan with patients who underwent MRI (as both group may had some diseae).

Seyed, I agree with most of your comments, but they emphasise the problems with the study. On the other hand, even if it is defective, I think it is a big improvement to be looking at data derived from a CT scanned population rather than atomic bomb survivors, and most of the information used in making risk estimates to date come from the latter. Radiology now operates on such a scale that studies based on follow up of radiological patients are probably now possible and it seems, relatively inexcusable that they have not been done long ago. When one considers the scale of the spend on staff, buildings, equipment and research in radiology, the fact that this fundamental question is not yet addressed seems to be something of an elephant in the room. The option of comparing with MRI may be good if the populations turn out to be really similar.

Big numbers does not make better numbers. Let's keep things on the science level. Agree mostly with Seyed. There is this immense selection bias in the study. There is no reliable way to get ou of it. ALARA remains the only criteria. X rays are not good for health unless the balance is in favour of diagnosis and treatment. Same things as in ICRP.

I with agree with Filippo Cademartiri, we should keep things at the science level wherever possible. But the data is not complete and ICRP publications are not science. They, on the contrary, represent the commission's best consessus of what the risk is and what should be done about it when we don't have real data covering the medical risks of imaging. Essentially ICRP reports are a subtle (and sometimes not so subtle) integration of incomplete data with a set of values. We know all about the data, but we do not know so much about the value system used.

The risk of medical exposures is a “sealed black box”, nobody knows what is inside!

Although we don’t know the accurate magnitude of the risks (and/or benefits) of medical exposures, we should not forget that current radiation safety philosophy is based on the conservative "linearity". However, being conservative is not always a good thing! Issues such as radiophobia and the huge expenses of linearity should be clearly addressed. On the other hand, non-linear phenomena such as adaptive response may occur after any medical exposure.

A reminder of the Image Gently project / pledge -

http://www.pedrad.org/associations/5364/ig/

Yes Yueh,

The Image Gently (and the Image Wisely) pledges are in the background to all of this.

Two points. First the idea of a pledge is particular to the US. There are many parts of the world where people will implement the idea but would find the notion of taking a pledge to do so embarassing and OTT.

Second, to be able to do what the pledge(s) aspire to we must know the real risks confidently, and this we don't at the moment. The ICRP and BEIR documents provide the best prudent estimate the scientific community can make. This is why data like the Australian Survey are welcome. Even if the survey is not perfect and has flaws it adds to our knowledge and lets us approach statements about risk a bit more confidently.

jIM

How do we react? Well.... we don't. There is nothing special to react to.

The paper is interesting, as it provides some results on influence of "low" doses, in the range for which we do not really have good data. And it shows rise of cancer risk, that is still statistically significant even if we only take the cancers detected 10+ years after a CT scan - this seems quite convincing for me. It would be certainly nice if the authors would analyze each case (dose, precise localization of scan, why the scan was ordered, histological type of tumor, etc), but that is simply impossible. What they had was data from Medicare records + national cancer database, and they made quite a good use of it. And maybe if they would really analyze each case, and had full data (such as data for ALL of the CT scans, not only on this funded by Medicare), the results would be even more significant. Anyway, the paper is *MUCH* better than calculation of number of induced cancers based on "linear non-threshold" model based on Hiroshima...

But the paper simply says: do not use radiation, if you do not have a good reason for that. Well, we already know that. It also says that in un-scanned population some 3.5 person per 1000 (in some age group, during several years of observation, etc) will get a cancer, and in the scanned group the number will by raised to 4.5. That makes a ONE induced cancer per 1000 of examinations. Examinations that were made to diagnose a patient, and to save her/his live. So there is a risk, but it is quite small compared to the benefit. Well, I guess we know that to. So this is nothing new.

The only problem is with so-called "journalists" (if you can call a student working part-time for an Internet portal a journalist), who always manage to convert such findings into a "CT scans kill people!" news. And with web users, that will comment the news. You know, comments like "yeah, I know someone who had a terrible car accident, but was still alive, but they took him to hospital, made a CT scan, and he died soon after that!". We should try to fight with such news (how??), but not necessarily with scientific papers. Because we should not let people be scared of justified use of radiation in medicine. But on the other hand, we can not react hysterically (or allergically) to all the papers, that suggest that radiation may actually have some bad effect on health. Because sometimes their authors may be right.

It is an example of "sensationalising" a medical procedure, without clear cut guidance. The study suffers from problems, we have put in place laws on radiation exposure, including justification for examination. This is what we should point at. We are seeing a lot of referrals from clinicians who see CT as a wonderful test to ensure their physical examination/history didn't miss anything, which is the wrong way to go about it - "excludograms" tend to be useless.

Working on dose reduction in CT is paramount, and has come a long way since the data of the study were generated.

Lastly - I agree that radiation can have a negative side effect for developing cancer, but at the same time, how many patients would have died due to misdiagnosis? That is the balance we have to work with, and we should most certainly question seemingly inappropriate requests for CT.

Very good point from Edwin Beek. These are the limitations I was referring to in my previous post.

A similar paper was published in the US in the year 2000 - predicting increased cancer risk for children receiving CT scans on machines using adult settings. The problem is that when this report was publicized we had parents bringing children into our hospital and denying permission to perform medically necessary CT scans - in many cases stating a preference for exploratory surgery. In every case the surgery was clearly a higher-risk procedure. However well-intentioned the paper, it increased the risk to the children of these overly worried parents.

We should also keep in mind the absolute risk in addition to the relative risk. If the risk of developing a childhood cancer is (to make up a number) 0.1% then increasing this number fo 25% sounds substantial, but only raises it to 0.125%. In other words, increasing the odds of a very low-probability risk does not affect very many people at all - especially not compared to those who are helped by having accurate diagnostic information. Of course we should minimize dose to everyone - but we need to keep the risks in perspective.

Svi se slažemo da je zračenje štetno. Djeca su posebno osjetljiva skupna jer zbog rasta i brzog metabolizma stanice su osjetljiije nego kod odraslih. Također kod djece postoji i thimyus koji odrasli nemaju. Osnovni problem s kojim se srećem u svojem poslu jer radim na CT skeniranju pacijenata svakodnevno je problem liječnika koji prekomjerno propisuju CT skeniranja, ne odvagnu korist i štetu od zračenja kod pacijenta. Nemojmo zaboraviti genetsku predispoziciju, prehranu, mjesto stanovanja i ostale faktore koji utječu na pojavu karcinoma. Poanta priče; potrebni su iskusni i dobro educirani liječnici koji znaju što žele dokazati propisivanjem CT pretrage. Također prema vrsti dijagnoze na CT uređaju se odabire pojedini protokol za skeniranje, te se može desiti ako ne znamo na što se sumnja, može se desiti da se karcinom u pojedinoj fazi protoka kontrasta (arterijska ili venska faza) ne prikaže. CT nije pretraga koja se radi da se nešto napravilo i da se zadovolji forma.

Vrlo interesantno!

Perhaps we can keep this in English? Sorry, speak Dutch, German and French too - none help me with the latest answers.....

Alan wrote about problem of prescribing CT exams without proper justification, just to make doctors "feel good" - this is what he observes in his job. I just answered that it was "very interesting".

At least this is what I think, or what Google Translate thinks - i don't know Croatian ;-)

Exactly. We need to know at least roughly what we want to prove with examination. Then we can do a search in the patient's best interests. Optimal protocol with less radiation.

Thanks

Discussions in the following two articles will be helpful.

http://www.bmj.com/content/346/bmj.f2360/rr/648506

https://www.ncbi.nlm.nih.gov/pubmed/25816281   (PDF freely downloadable)

Thanks Nobuyuki.  I had seen the first one but not the second.  Jim

My pleasure, Dr. Jim F Malone