Computed tomography scan radiation and brain cancer incidence.

BACKGROUND: Computed tomography (CT) scans make substantial contributions to low dose ionizing radiation (IR) exposures, raising concerns about excess cancers caused by diagnostic radiation.

METHODS: Deidentified Medicare records for all Australians aged 0-19 years between 1985-2005 were linked to national death and cancer registrations to 2012. The National Cancer Institute CT program was used to estimate radiation doses to brain from CT exposures in 1985-2005, Poisson regression was used to model the dependence of brain cancer incidence on brain radiation dose, lagged by two years to minimize reverse causation bias.

RESULTS: Of 10,524,842 young Australians, 611,544 were CT-exposed before age 20, with a mean cumulative brain dose of 44 milligray (mGy) at an average follow-up of 13.5 years after the two-year lag period. 4,472 were diagnosed with brain cancer, of whom only 237 had been CT-exposed. Brain cancer incidence increased with radiation dose to the brain, with an excess relative risk of 0.8 (95% CI 0.57-1.06) per 100 mGy. Approximately 6,391 (95%CI 5255, 8155) persons would need to be exposed to cause one extra brain cancer.

CONCLUSIONS: For brain tumours that follow CT exposures in childhood by more than two years, we estimate that 40% (95%CI 29-50%) are attributable to CT radiation and not due to reverse causation. However, because of relatively low rates of CT exposure in Australia, only 3.7% (95% CI 2.3-5.4%) of all brain cancers are attributable to CT scans. The population attributable fraction will be greater in countries with higher rates of pediatric scanning.