Childhood CT and X-ray radiation dose tracks with higher brain tumor risk in a 3.7 million child cohort
A JAMA cohort of over 3.7 million children links cumulative imaging radiation to dose-related brain tumor risk. Editorialists say the finding must be weighed against diagnostic benefit.
Summary
Medical imaging that uses ionizing radiation, such as CT scans, was linked to a dose-related rise in brain tumor risk in children and adolescents. The retrospective cohort study followed more than 3.7 million young people and was published in JAMA. Compared with no exposure, hazard ratios rose from 1.12 at the lowest cumulative dose to 3.85 at 150 mGy or more. By age 21, brain tumors occurred in 5.4 per 10,000 unexposed children versus 22.1 per 10,000 at the highest dose. At a dose similar to one head CT, the estimated excess was 7.9 tumors per 10,000 children. This is an observational association, not proof of cause. Editorialists stress that the small long-term risk must be weighed against the often larger, immediate benefit of a correct diagnosis. The practical message is to image only when justified and to use the lowest dose that gives adequate quality.
Detailed Summary
Radiation exposure is a well-established driver of DNA damage and cancer, and children are thought to be more vulnerable because their tissues are still developing and they have decades ahead in which a tumor could emerge. This new study puts numbers on that concern for brain tumors, which makes it relevant to anyone thinking about lifetime cancer risk and the cumulative cost of medical procedures.
In a retrospective cohort study of more than 3.7 million children and adolescents, researchers led by Diana Miglioretti of UC Davis examined imaging with ionizing radiation and later brain tumors. Compared with no exposure, hazard ratios climbed from 1.12 (95% CI 1.01-1.26) for the lowest cumulative dose category (above 0 to under 1 mGy) to 3.85 (95% CI 2.43-5.75) for 150 mGy or more. The excess relative risk was 1.27 per 100 mGy (95% CI 0.83-1.77), and 1.51 per 100 mGy for aggressive histologies.
In absolute terms, cumulative incidence by age 21 rose from 5.4 per 10,000 with no exposure to 22.1 per 10,000 at 150 mGy or more, an excess of 16.7 per 10,000. For 50 to under 75 mGy, a range comparable to one head CT, incidence was 13.3 per 10,000, an excess of 7.9 per 10,000. The estimated population-attributable fraction was 10.1% (95% CI 5.8-14.2).
The results fit with earlier studies and reinforce a familiar principle: order radiation-based imaging only when clinically justified, and optimize protocols to minimize dose. Editorialists from Rome called the signal strengthened but said it should not be used alone when counseling families. For an individual child, the small long-term risk has to be set against the often much larger, immediate benefit of the right diagnosis.
Caveats matter. This is observational, so residual confounding is possible, including reverse causation if scans were ordered because of early tumor symptoms. The news report gives limited detail on how that was handled, so check the JAMA paper. Brain tumors remain rare, and these data do not support refusing needed imaging.
Key Findings
- Cohort of over 3.7 million children: cumulative imaging radiation dose was associated with higher brain tumor risk in a dose-related pattern.
- Hazard ratios rose from 1.12 at under 1 mGy to 3.85 at 150 mGy or more versus no exposure.
- Excess relative risk was 1.27 per 100 mGy, and 1.51 for aggressive histologies.
- At a dose near one head CT, excess risk was about 7.9 tumors per 10,000 exposed children.
- Editorialists say the small risk must be weighed against diagnostic benefit and not used alone in family counseling.
Methodology
This is a news report on a retrospective cohort study (over 3.7 million children and adolescents) published in JAMA, with an accompanying editorial. The endpoint is incident brain tumors, an observational outcome rather than a surrogate marker. MedPage Today is a reputable trade outlet, but the article is brief and truncated.
Study Limitations
Observational design cannot prove causation; confounding and reverse causation (scans ordered for early tumor symptoms) are possible, and the report does not detail how these were handled. Dose estimates and the excess-risk figures are modeled, and the article text cuts off before comparing with other studies. Verify methods, sensitivity analyses and lag periods in the JAMA paper.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe:
