Does CT cause cancer? Assumptions questioned by new evidence

One of the biggest medical controversies in recent years concerns claims about radiation risks from CT (computed tomography) imaging. Although experts have questioned certain facets of such claims, an especially powerful riposte was published in an article late last year in the ‘American Journal of Clinical Oncology’. The authors took great pains to explain that the “widespread belief” about a link between medical imaging and cancer was founded on “an unproven” and “illegitimate” theoretical model, dating to the 1940s.

Media fuels emotion, fear
The emotive nature of the CT and cancer debate may be best illustrated by an Op-Ed in the ‘New York Times’ on January 31, 2014. Its authors are two Professors at the University of California, San Francisco Medical Center – radiologist Rebecca Smith-Bindman and cardiologist Rita Redberg.
The headline of the article, in one of the world’s most influential publications, clearly seeks to draw maximum attention. “We Are Giving Ourselves Cancer,” it says. The opening paragraph continues in the same vein, closing with the observation that “we” are “silently irradiating ourselves to death.” So too does the final sentence of the Op-Ed – that “we” must find ways to use CTs “without killing people in the process.”

Expert oversights
Professors Redberg and Smith-Bindman acknowledge that “medical imaging can be lifesaving.” Their principal argument is that once rare, CTs have now become routine and that the “current high rate of scans” do not correlate to “better health outcomes.” The reasons for growth in CT use are both good and bad: “desire for early diagnoses, higher quality imaging technology,” as well as the “financial interests of doctors and imaging centers.” However, the authors do not even attempt an informed guess about whether good motives outweigh the bad.
In contrast, they state conclusively that there is “evidence of its harms.” This evidence consists of two clinical studies in Britain and Australia where the risks of CT, they say, were “directly demonstrated,” especially in children.
One of their biggest oversights was to avoid mentioning the conclusions of the two studies. Both, in fact, took great care to qualify their verdicts.

Key studies in Britain and Australia qualify judgements
The British study, published in ‘Lancet’ in August 2012, was titled ‘Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study’. It used data on 175,000 children and young adults and found a three-fold increase in the risk of brain tumoUrs and leukemia. However, the cumulative 10-year risk was one excess case of leukemia and one excess case of brain tumour per 10,000 head CT scans.

On its part, the Australian study was published in the ‘British Medical Journal’ in May 2013 and titled ‘Cancer risk in 680,000 people exposed to computed tomography scans in childhood or adolescence: data linkage study of 11 million Australians.’ The authors found a 24% increase in childhood cancer risk over a 10-year period – from 39 per 10,000 young people to 45, after a CT scan. These findings were, like the British study, put in perspective by an accompanying editorial: the incidence of cancer in children “is extremely small and so a 24% increase makes this risk just slightly less small.” In addition, the authors observed that almost 60% of CT scans were of the brain and “in some cases the brain cancer may have led to the scan rather than vice versa.”

Indeed, unlike the ‘New York Times’ Op-Ed, the two studies provided a balanced view. Above all, they took great care to avoid drawing alarmist conclusions.

The need for balance
The British study stated that “immediate benefits of CT outweigh the long-term risks in many settings and because of CT’s diagnostic accuracy and speed of scanning …, it will remain in widespread practice for the foreseeable future.” Lead author Mark S. Pearce of Newcastle University’s Institute of Health and Society echoed this forcefully, noting that “CT can be highly beneficial for early diagnosis, for clinical decision-making, and for saving lives. However, greater efforts should be made to ensure clinical justification and to keep doses as low as reasonably achievable.”
The Australian study, too, concluded that practitioners “will increasingly need to weigh the undoubted benefits of CT scans in clinical practice against the potential risks to justify each CT scan decision.”
Dosage: wide margins for error
Meanwhile, one of the biggest issues of concern with CT is a lack of clarity – and some uncertainty – about dosage and exposure. The British study, for example, underscored that the increase in risk followed “two or three CT scans of the head” under “current scanner settings” for brain tumours and “five to 10 head CT scans” for leukemia.
Radiologists have in fact not reached a consensus on how to define a dose, according to Michael McNitt-Gray, an associate professor of radiological sciences at the University of California, Los Angeles. Doses per indication vary by institution and by patient size. As a result, no national average is available.
An additional problem is that effective, organ-specific doses relate to how much radiation the body absorbs. However, CT scanners do not report the absorbed fraction, says McNitt-Gray. Instead, they report only what the machine emits, which is less than what is absorbed by the body.

Australian study urges validation of risk model
Nevertheless, in general, there is widespread agreement that CT scans should be limited to situations with a definite clinical indication and that scans should be optimiZed to provide a diagnostic CT image at the lowest possible radiation dose. The editorial accompanying the Australian study in the ‘British Medical Journal’ called for further validation of risk models, and stated that “more accurate risk assessment “can be performed to “better inform imaging decisions.”

Missing nuances
Such nuances seem to have been missed out by the Op-Ed in the ‘New York Times’.
A “single CT scan,” the authors wrote, “exposes a patient to the amount of radiation that epidemiologic evidence shows can be cancer-causing.” They also asserted that CT radiation was “100 to 1,000 times higher than conventional X-rays.”
Their certainty stands in some contrast to a quote from the British study used to justify the authors’ views: The “amount of radiation delivered during a single CT scan,” it said, “can still vary greatly and is often up to 10 times higher than that delivered in a conventional X-ray procedure.”

French study in 2014 raises new doubts
In October 2014, nine months after the ‘New York Times’ Op-Ed, a French study of 67,274 children raised further doubts about the strength of the correlation between CT and childhood cancer.  The objective of the study, published in the ‘British Journal of Cancer’, was to estimate how cancer-predisposing factors (PFs) affected assessment of radiation-related risk in CTs.
The authors found that adjusting for PF “reduced the excess risk estimates related to cumulative doses from CT scans” and that “no significant excess risk was observed in relation to CT exposures.” The study concluded  that there was a need “to avoid overestimation of the cancer risks associated with CT scans.”

The 2009 NCI study
Aside from the British and Australian studies, an authoritative and oft-cited source for most of the alarms about CT have their origins in a 2009 report from the US National Cancer Institute (NCI). This report was referenced by the authors of the ‘New York Times’ Op Ed, in their statement that  “CT scans conducted in 2007 will cause a projected 29,000 excess cancer cases and 14,500 excess deaths over the lifetime of those exposed.”

BEIR and Lifetime Attributable Risk
The NCI figures were based on the US National Research Council’s Biological Effects of Ionizing Radiation (BEIR) report in 2006 and estimated the mean number of radiation-related incident cancers with 95% uncertainty limits (UL). 
The so-called BEIR VII model generates what it calls lifetime attributable risk (LAR) factors. These estimate the likelihood of cancer in hypothetical individuals as a function of dose. Multiplying LAR by the number of people exposed to a given dose yields an estimate of expected cancers from that exposure in the population.
BEIR VII was also used by Smith-Bindman, one of the authors of the ‘New York Times’ Op-Ed, who used it in a study of four San Francisco facilities to estimate that one cancer might appear for every 270 middle-aged women undergoing CT coronary angiography, and that women aged 20 who underwent the procedure had twice the risk as middle-aged women.

BEIR and linear no-threshold: assumptions challenged
The so-called BEIR VII model postulates that there is no safe level of ionizing radiation exposure. As a result, it is assumed that carcinogenic effects follow a linear dose response – which means that even the smallest exposure carries some level of cancer risk. It is precisely this baseline assumption which has been challenged by a recent article in the ‘American Journal of Clinical Oncology’. The article, published in November 2015,  is titled ‘The Birth of the Illegitimate Linear No-Threshold Model: An Invalid Paradigm for Estimating Risk Following Low-dose Radiation Exposure.’
BEIR VII is based on the linear no-threshold (LNT) model. However, risk estimates based on LNT “are only theoretical” and “have never been conclusively demonstrated by empirical evidence,” according to lead author James Welsh, Professor in the Department of Radiation Oncology of Loyola University Chicago Stritch School of Medicine.

Dosing fruit flies after 70 years
The article painstakingly re-examines the original studies, dating back over 70 years, which led to adoption of LNT. The experiments involved exposure of fruit flies to various doses of radiation, and concluding there was no ‘safe’ level of radiation – the basis of the LNT model which is used to this day. However, such a conclusion was unwarranted as the experiments had not been done at truly low doses and there is growing evidence that the human body has evolved the ability to repair damage from low-dose radiation – such as that which occurs naturally in the environment.
Indeed, as the authors argue, the first study to expose fruit flies to low-dose radiation was conducted only in 2009 and its findings did not support the LNT model. Other sources too suggest that the dose-response relationship between radiation and somatic mutation has a threshold, and that biological defence mechanisms come into play at low radiation levels.

Abandoning LNT
Use of the LNT model, according to the authors of the article in the ‘American Journal of Clinical Oncology’, dissuades many physicians from using appropriate imaging techniques and “discourages many in the public from getting proper and needed imaging, all in the name of avoiding any radiation exposure.” They conclude that the LNT model “should finally and decisively be abandoned.”