select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
A study published in JAMA Internal Medicine looks at CT scans and lifetime cancer risk in the USA.
Lynda Johnson, Professional Officer for Clinical Imaging and Radiation Protection, The Society and College of Radiographers, said:
“The Society and College of Radiographers (SoR) welcomes research into the harmful effects of ionising radiation and recognises the importance of balancing benefit and risk information to patients and the public.
“This paper articulates the complexities of large-scale dose estimation and acknowledges the many variables which influence an individual’s likelihood of developing cancer at some point in their lifetime. In the UK, the use of ionising radiation is governed by The Ionising Radiation (Medical Exposure) Regulations 2017 (The Ionising Radiation (Medical Exposure) Regulations (Northern Ireland) 2018). Central to the legislation and UK radiographic practice, as this paper rightly concludes, are the principles of justification and optimisation. Justification means that any exposures to ionising radiation for medical purposes must be demonstrated to provide a greater benefit than risk to the individual. Once justified, the exposure must be optimised, meaning that it is as low as reasonably practicable to provide the intended outcome, or answer the clinical question.
“Computed Tomography (CT) scans are undertaken by highly trained radiographers and nuclear medicine technologists who have met the educational and professional standards required to ensure all CT scans are appropriately justified and optimised. Considering the increased use of CT as an invaluable diagnostic tool, it is imperative that the risk of harm from potential misuse, poor quality referrals, or inappropriate exposure parameters continues to be managed effectively. This is achieved by safeguarding standards of education, training and practical experience, compliance with the regulations, and applying best practice quality standards such as The Quality Standard for Imaging.
“It is particularly important to recognise, as this paper highlights, the increased risk to children from unjustified CT exposures. Staff are trained to give special consideration to the justification and optimisation of CT scans for children and will assess the benefits and risks of using CT against alternative techniques that do not involve ionising radiation such as MRI and Ultrasound.
“Accurate communication around the benefits and risks of CT is essential to protect the public from harm. Focussing on risk alone is not helpful and, in some cases, might prevent a person from attending a scan that could provide early diagnosis of cancer. Anyone undergoing a CT scan must be provided with balanced, accurate and relevant information to enable them to understand what it means to them as an individual in terms of their diagnosis, treatment and potential long-term care.
“The UK Health Security Agency is responsible for undertaking dose audits and producing National Diagnostic Reference levels (NDRLs) for computed tomography. These inform local practices and employers must ensure their organisational doses do not consistently exceed the NDRLs. They are publicly available here alongside helpful dose comparisons here and benefit and risk information for patients here.”
Dr Doreen Lau, Lecturer in Inflammation, Ageing and Cancer Biology at Brunel University of London, said:
“This is a well-conducted modelling study using robust data from US hospitals and established methods for estimating cancer risk from radiation exposure. It provides a timely reminder that while CT scans are often life-saving and essential for diagnosis, they do come with a small but real potential risk of contributing to cancer over a lifetime, especially when used repeatedly, in younger patients, or when not clinically necessary.
“The findings don’t mean that people should avoid CT scans when recommended by a doctor. In most cases, the benefit of detecting or ruling out serious illness far outweighs the very small risk of harm. What this research highlights is the need to minimise unnecessary imaging and use the lowest dose possible, particularly in settings where CT usage is high. Where appropriate, clinicians may also consider alternative imaging methods that do not involve ionising radiation, such as MRI or ultrasound—especially for younger patients or when repeat imaging is anticipated.
“CT scan rates are much higher in the US than in the UK, where imaging is used more conservatively and with stricter clinical justification. That means the estimated risks in this study are likely to be much lower in the UK context, though the message about appropriate use still holds.
“Importantly, this study models estimated cancer risk from radiation exposure. It does not show a direct causal link between specific CT scans and individual cancer cases. These are projections based on population-level data and assumptions about radiation risk, not observed cancer rates. Although the model estimates a small increased risk with each scan, it does not prove that any one scan causes cancer. Other factors such as underlying health issues and clinical decision-making, may also influence who gets scanned and how often.”
Prof Stephen Duffy, Emeritus Professor of Cancer Screening, Centre for Cancer Screening, Prevention and Early Diagnosis, Queen Mary University of London, said:
“This paper reports on a very high quality numerical modelling exercise, estimating the likely number of cancers occurring in the USA as a result of 93 million CT examinations. The authors estimate that just over 100,000 cancers are predicted to occur as a result of radiation from these CT examinations. This amounts to around a 0.1% increase in cancer risk over the patients lifetime per CT examination. When we consider that the lifetime risk of cancer in the general population is around 50%, the additional risk is small. Doctors do not order CT examinations unless they are necessary, and it seems to me that the likely benefit in diagnosis and subsequent treatment of disease outweighs the very small increase in cancer risk.
“I would also remark that the estimates, while based on the best models available to the authors, are indirect, so there is considerable uncertainty about the estimates.
“Thus I would say to patients that if you are recommended to have a CT scan, it would be wise to do so.”
Dr Giles Roditi, Consultant Cardiovascular Radiologist and Honorary Clinical Associate Professor of Radiology, University of Glasgow, said:
“CT scanning is a powerful diagnostic tool and has become a bedrock of modern radiology departments, particularly for emergency department imaging. However, the paper by Smith-Bindman et al. is a timely reminder that with great power comes great responsibility. The paper makes the case that the rise in the utilisation of CT scanning is now at such a scale that its projected use could lead to scenario in which CT-associated cancer eventually accounts for 5% of all new cancer diagnoses annually in the USA. What should we do with this information and how does this translate to and inform practise in the UK ?
“Firstly, the evidence base is sound and there is little new as regards the basic assumptions that the paper is based upon but the authors have updated this with more modern dose estimates and data on the utilisation of CT scanning not only across different age groups but also stratified by gender and the exposure of different organs that have different sensitivities to ionising radiation induced damage. The authors are to be congratulated in the detailed breakdown of CT utilisation across these categories and how lifetime risk of cancer impacts across age and gender etc. as well as the modern dosimetric approach used plus accounting for multiphase CT examinations that inevitably entail higher dose.
“With all medical endeavours there is an element of risk. Risk is generally defined as a situation involving exposure to danger or the possibility that something unpleasant will occur. Furthermore, the use of the word risk often implies an element of chance, uncertainty or unpredictability. However, risk can often be well defined in any particular context as – Risk = (probability of an event) x (impact of event)
“Risk is thus different for ‘well’ versus ‘sick’ patients with the latter deriving greater benefit. This paper helps us better define risk at a population level by updating knowledge on the probable incidence of later CT-associated cancer. A potential limitation that could be levelled at the paper is that not all the risks associated with CT are included, only those related to later development of cancer diagnoses. For example, other relevant factors as a demerit to CT scanning could include the very small risks of anaphylaxis related to the use of contrast medium, used now in a large proportion of scans in Western medicine. Similarly, the small but potential other risks such as cataract acceleration are not mentioned.
“On the other hand, while the authors mention that ‘CT is frequently lifesaving’ they have not in my opinion really put the information in full relevant context. The authors context is that this is approximately 5% of new cancer diagnoses could be attributable to CT i.e. a figure of 100,000 cancers in the USA is where there were 1,777,566 new cancer cases reported in 2021 and 608,366 people died of cancer in 2022 (the latest CDC data available). This is because the natural incidence of cancer induction is 1 in 2 for adults. Hence, an alternative way of looking at this would be that although the figure of 100,000 cancers is alarming this is only a small additional risk over and above an individual’s lifetime risk of developing cancer i.e. a risk rising from about 50% to 52.5%. The authors also do not address how many of these cancer will be fatal although we presume based upon CD data it would be approximately one third.
“The main issue, however, is that the benefits of CT scanning are not more explicitly stated. This is likely because the benefits of most medical imaging in terms of morbidity & mortality have been very difficult to quantify with surprisingly little published in the literature. This is mainly because imaging has too often only been part of an overall therapeutic strategy where the main treatment outcomes depend critically upon the imaging but the imaging itself is not tested (e.g. treatments for stroke and cancer). However, there have been recent trials that provide some context, for example SCOT-HEART was probably the first major trial in which diagnostic CT was shown to save lives. In SCOT-THEART the patients were randomised to a conventional treatment pathway without CT scan or an investigative arm in which the standard care pathway was simply supplemented by a CT scan of the coronary arteries. This trial showed clear benefit for those patients that had CT with a significantly lower mortality rate and this has been shown to persist now up to 10 years following the end of the trial. Similarly trials of lung cancer screening have now shown positive benefit from CT scanning in the detection of early, treatable stage lung cancer in high risk patients.
“So how does this translate into the situation in the UK ? Firstly, there are significant differences in practise due to both cultural and legislative environments. In the UK we operate under the precepts of the Ionising Radiation (Medical Exposure) Regulations last updated in 2017 which mandates that we apply the ALARA/ALARP principles and should opt for diagnostic imaging tests with the lowest radiation dose, or preferably an imaging test with no ionising radiation exposure (e.g. ultrasound or MRI) where this answers the clinical question. Culturally in the UK we also regard all requests for imaging as just that, requests that can be questioned through discussion. In the USA clinicians order scans and radiology departments have little room to manoeuvre when it comes to not performing or changing these orders, particularly since the imaging fees that accompany the scanning activity are the lifeblood of the department. Another issue in the USA in addition to the overuse of CT mentioned in the paper is the repeat imaging that is often performed in a fragmented healthcare system where it is easier (and more profitable) for an institution to simply repeat a scan on a patient referred in from elsewhere rather than seek out and transfer the original scans.
“In the NHS we have systems that allow image transfer between institutions and of course unlike the USA we are very capacity limited and often have long waiting times for scans. One side effect of this is that it tends to reduce demand such that tests unlikely to influence clinical decision-making are less likely to be requested. On the downside is that the CT scanner base in the UK is aging and we know that older scanners inevitably expose patients to higher radiation doses than modern systems for the same type of scan, often with less good image quality. Indeed, on modern generation systems with advanced iterative reconstruction algorithms and AI enhancements in the imaging chain then CT scans can be acquired at doses similar to (or little more than) conventional x-rays. These advances have largely been spurred by the drive to reduce dose in coronary CT scans but the benefits potentially reduce doses across all CT scanning. The paper by Smith-Bindman et al. reminds us that we must advocate more strongly to upgrade our CT scanners for the benefit of our patients.
“So what would I say to a UK patient scheduled to have a CT scan and worried by this paper ? In general terms I would strongly advise them not to worry as they are highly likely to benefit from a well indicated scan, this is particularly so in those who are unwell and in older patients (those > 55 years). For younger patients, particularly those of child-bearing age where the breasts and/or reproductive organs would be included and for those who are physically well then if concerned they can always ask to discuss the merits of alternative scans such as ultrasound and MRI. For example, in our own practise we image all our altruistic potential living kidney donors with MRI rather than CT since our own (unpublished) estimates indicate that if we used CT then 1 in 526 of these well people would have a fatal induced cancer, a risk eliminated by using MRI.”
Prof Richard Wakeford, Honorary Professor in Epidemiology, Centre for Occupational and Environmental Health (COEH), University of Manchester, said:
“Although it is not unreasonable to reiterate guidance on the potential risks to health arising from exposures to low levels of ionising radiation, such as the x-ray doses received from CT scans, considerable caution is required in providing quantitative estimates of the effects produced by such exposures. This is largely because of the substantial assumptions that must be made in applying risk models derived from epidemiological studies of populations briefly exposed to moderate and high doses, primarily the Japanese survivors of the atomic bombings of Hiroshima and Nagasaki, to low-level exposure circumstances. For example, for the purposes of radiological protection, it is prudent to assume that the size of the additional risk is directly proportional to the dose received, with no threshold dose below which the risk is zero, and this is the assumption made by the International Commission on Radiological Protection (ICRP) in making its recommendations. However, ICRP notes that these assumptions “conceal large biological and statistical uncertainties”, and cautions against risk projections based on large numbers of people receiving low doses.
“The direct epidemiological investigation of cancer incidence among patients who have been examined by CT is a worthwhile exercise, but substantial care is required in the interpretation of results – as with all medical diagnostic procedures, people are examined because they are ill, have been ill, or are suspected of being ill, and such selection for exposure leads to difficulties in obtaining reliable conclusions about the effects of radiation exposure from these studies.
“The “bottom line” of the paper is that ~103,000 cases of cancer (which does not include cases of non-melanoma skin cancer, lymphoma, or multiple myeloma) are estimated to result from CT scans conducted in the USA in 2023, an estimate that must be viewed with circumspection. This estimate of ~103,000 cases of cancer is, on the face of it, rather alarming, but it is also uncertain, to an extent that extends (well) beyond the uncertainty limits presented in the paper. ICRP emphasises that all medical exposures must be justified as doing more good than harm, and the potential risk from radiation exposure during a diagnostic examination clearly needs to be factored into clinical judgement about the need for a specific diagnostic procedure. The level of potential risk posed by exposure to low doses of radiation should be taken into account in reaching a balanced decision on whether or not a CT scan is clinically desirable, but this judgement should not be unduly influenced by large, but uncertain, projected numbers of cancers.”
‘Projected Lifetime Cancer Risks From Current Computed Tomography Imaging’ by Rebecca Smith-Bindman et al. was published in JAMA Internal Medicine at 16:00 UK time on Monday 14 April 2025.
DOI: 10.1001/jamainternmed.2025.0505
Declared interests
Prof Stephen Duffy: I have no conflict of interest.
Dr Giles Roditi: Prof Roditi is a Past-President of the British Society of Cardiovascular Imaging/Cardiovascular CT, a Past President of the Society of Magnetic Resonance Angiography and a member of the SCOT-HEART investigators.
Prof Richard Wakeford: “I am, or was, a member of a number of national and international expert committees addressing radiation risks, such as ICRP, UNSCEAR and (previously) COMARE, SAGE, etc.. Details can be found at: https://research.manchester.ac.uk/en/persons/richard.wakeford
“I am a member of the Technical Working Party of the Compensation Scheme for Radiation-Linked Diseases (http://www.csrld.org.uk/), for which I receive a small consultancy fee. I also receive small payments for lecturing in academic and various professional courses (e.g., https://www.oecd-nea.org/jcms/pl_27505/international-radiological-protection-school-irps-at-stockholm-university). Otherwise, I am formally “retired” from employment, although I seem to be as busy as ever!”
Dr Doreen Lau: no financial or conflicts of interest related to this study.
For all other experts, no reply to our request for DOIs was received.