Americans underwent 93 million CT scans in 2023, and researchers now estimate that radiation from those exams could eventually cause roughly 100,000 new cancers each year if current imaging patterns persist. That projection, derived from National Cancer Institute modeling tools applied to updated scan volumes, would mean CT-related cancers account for about 5 percent of all new U.S. cancer diagnoses. The gap between the clinical value of CT imaging and its cumulative radiation cost has become one of the most pressing unresolved questions in diagnostic medicine.
Rising scan volumes and the 100,000-cancer projection
The scale of the problem tracks directly to how many scans are performed and how much radiation each delivers. With 93 million CT scans performed in 2023, the United States now conducts far more imaging per capita than any prior decade. Earlier National Cancer Institute analyses, based on lower annual scan counts, had placed the projected cancer toll at approximately 29,000 future cases attributable to a single year of scanning. Updated scan volumes and refined risk coefficients have pushed the estimate sharply higher, with NCI researchers now suggesting that CT-related radiation may contribute to roughly 5 percent of all new cancers in the United States.
The newer figure draws on the NCI’s Radiation Risk Assessment Tool, known as RadRAT, which converts low-dose radiation exposures into lifetime cancer incidence risks. RadRAT applies established dose-response models, accounts for patient age and sex, and produces probability distributions rather than single-point estimates. Its peer-reviewed methodology, published in the journal Radiation Research, documents how uncertainty distributions and dose-rate adjustments feed into each projection. The tool was not originally designed as a population-level forecasting engine, but researchers have applied its per-scan risk estimates across the full volume of annual imaging to arrive at aggregate numbers.
The critical arithmetic works like this: if each CT scan carries a small but nonzero probability of inducing cancer over a patient’s remaining lifetime, multiplying that probability across tens of millions of scans per year produces a large absolute number. The resulting estimate, that CT radiation could account for roughly 100,000 future cancers annually, depends on the assumption that current scan rates and dose levels continue unchanged. As the Division of Cancer Epidemiology and Genetics has noted in its overview of cancer risks from CT imaging, even modest changes in dose per exam or total scan volume can significantly shift long-term projections.
Pediatric patients face the steepest documented risks
Children absorb more radiation per scan relative to body size, and their developing tissues are more sensitive to DNA damage. A retrospective cohort study published in The Lancet tracked young patients who received CT scans and found increased risks of both leukemia and brain tumors at higher cumulative doses. That study, one of the largest to directly link pediatric CT exposure to later cancer diagnoses, provided observational evidence that complemented the statistical models. Its findings gave clinical weight to what radiation biologists had long predicted: that younger patients face disproportionate long-term consequences from repeated imaging.
The National Cancer Institute’s Radiation Epidemiology Branch has continued to study the relationship between medical radiation and cancer outcomes. Their work forms the scientific backbone for tools like RadRAT and for public health recommendations about limiting unnecessary exposure. The pediatric data, however, still rest heavily on a single large UK cohort. Updated U.S. incidence records broken down by age group and scan type have not yet been published in the same detail, leaving a gap between the modeled risk and confirmed outcomes in American children.
That uncertainty does not erase the signal already seen. In children, the same cumulative dose can represent a larger fraction of lifetime exposure, and the longer remaining lifespan gives more time for radiation-induced mutations to progress into clinically detectable cancers. As a result, pediatric imaging guidelines increasingly emphasize the principle of “as low as reasonably achievable” (ALARA), urging clinicians to tailor scan parameters to body size, avoid multiphase studies when single-phase imaging will suffice, and consider non-ionizing alternatives such as ultrasound or MRI whenever they can answer the clinical question.
Dose reduction technology and the open questions
One path to shrinking the projected cancer burden runs through technology. Modern CT scanners can use iterative reconstruction algorithms and automated exposure-control software to deliver diagnostic-quality images at lower radiation doses. Decision-support tools built into electronic health records can also flag orders for scans that may not be clinically necessary, prompting physicians to reconsider before the patient enters the scanner. If these systems achieved widespread adoption and cut the average effective dose per scan by a meaningful margin, the lifetime risk calculated by RadRAT for each patient would drop accordingly.
Whether that reduction will happen fast enough to change the trajectory is an open question. Hospitals adopt new protocols at different speeds, and the financial incentives in fee-for-service medicine can favor more imaging rather than less. No primary institutional source has published data on how many of the 93 million scans performed in 2023 were deemed avoidable or low-value. Without that baseline, measuring progress toward fewer unnecessary exams is difficult, and policymakers lack a clear target for how much imaging could be safely reduced without compromising diagnostic performance.
The modeling itself carries important limitations. RadRAT extrapolates cancer risk from studies of populations exposed to higher doses, such as atomic bomb survivors, and applies adjustment factors to estimate effects at the lower doses typical of CT scans. That extrapolation remains scientifically debated. Some radiation biologists argue that very low doses may carry even less risk than linear models predict, while others contend that cumulative exposure from repeated scans pushes many patients into dose ranges where the epidemiologic evidence is more direct. No recent real-world validation cohort has linked current U.S. scan volumes to observed cancer incidence in a way that would confirm or refute the 100,000-case projection.
Balancing clinical benefit and radiation risk
For patients, the practical takeaway is straightforward but not simple. CT scans remain among the most valuable diagnostic tools in medicine. They detect cancers, guide surgeries, and identify life-threatening injuries within minutes. The risk from any single scan is small, and in emergencies or for clearly indicated tests, the benefit overwhelmingly outweighs the potential long-term harm. The concern arises from cumulative exposure over many years, particularly when imaging is ordered reflexively rather than selectively.
Clinicians and patients can jointly lower that cumulative risk by asking a few key questions before each exam: Will the result change management? Is there a non-radiation alternative that can answer the same question? Has the patient had similar scans recently that might make a repeat unnecessary? When the answer to these questions supports proceeding, dose-optimized protocols and modern scanners can keep exposure as low as possible while preserving image quality.
Public understanding also plays a role. Educational efforts from federal agencies and academic centers have tried to convey both the benefits and risks of medical imaging in accessible terms. Initiatives linked through the National Institutes of Health’s science education resources aim to help students, patients, and families grasp concepts like radiation dose, stochastic cancer risk, and the trade-offs inherent in diagnostic testing. Better-informed patients may be more likely to question marginal scans and to keep personal records of prior imaging, which can prevent redundant exams.
Ultimately, the projection that CT radiation could contribute to 100,000 cancers per year is not a verdict against the technology but a warning about how it is used. The same tools that save lives daily can, in aggregate, nudge population-level cancer rates upward if deployed indiscriminately. Narrowing that gap will require coordinated efforts: manufacturers continuing to lower dose without sacrificing clarity, health systems aligning incentives with appropriateness rather than volume, regulators and professional societies updating guidelines as new data emerge, and patients engaging in informed conversations about when a scan is truly necessary.
If those changes take hold, the United States could maintain the life-saving advantages of CT imaging while bending the projected cancer curve downward. If they do not, the country may find that a portion of tomorrow’s cancer burden was quietly written into today’s radiology orders-an unintended consequence of a technology whose benefits were never meant to come with such a hidden cost.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
