The risks of radiation-induced cancer do not generally decrease with increasing age at exposure, according to an article published online Oct. 25 in the Journal of the National Cancer Institute. Although the findings contradict earlier data, an accompanying editorial characterizes the analysis as “premature.”

Studies of Japanese atomic bomb survivors and children exposed to medical radiation had estimated that excess relative risks (ERRs) are higher for individuals exposed to radiation during childhood than individuals exposed at older ages. However, “[t]he relationship between radiation-induced cancer risk and age at exposure in adulthood is less clear,” wrote Igor Shuryak, MD, lead author; David J. Brenner, PhD, of the Center for Radiological Research at Columbia University Medical Center in New York City, and colleagues.

Existing models focus on radiation-induced initiation, or induction of irreversibly altered premalignant cells, and predict continuous decreases in ERRs with increasing age. However, more recent models also consider radiation as a promoter of preexisting premalignant damage. The promotion model may play a larger role with middle age exposure because the adult body contains more pre-malignant cells than the pediatric body.

“A primary goal of this analysis was to assess whether an observed pattern of radiation-induced cancer risks that do not decrease monotonically with increasing age at exposure is biologically plausible,” according to Brenner and colleagues. The authors sought to develop a model that reflects initiation and promotion.

The model incorporated three parameters that characterize background cancer risk and three that describe short-and long-term radiation-induced modulations of cancer risks. Background risk parameters are spontaneous stem cell initiation and subsequent malignant transformation, premalignant niche replication and effects of age on premalignant niches. Radiation-related parameters include dose dependence of initiation, promotion processes and regulation of the number of premalignant stem cells per niche.

ERRs of cancer incidence as a function of age at exposure of Japanese atomic bomb survivors provided the dataset for analysis. Brenner and colleagues analyzed ERR data for all solid cancers combined and individually for liver, colon, lung, breast, stomach and bladder cancers. The datasets and methodologies were largely the same as BEIR-VII risk estimates, according to the authors.

After fitting the model to age-dependent cancer risks in atomic bomb survivors, the authors developed a formula that characterizes dose dependence of the radiation-induced initiation process and dose dependence of radiation-related promotional processes to measure the relative importance of initiation versus promotion for tumor sites. Although radiation-induced breast cancer risks do monotonically decrease with age at exposure, radiation-induced lung cancer risks reflect the increased importance of promotion and increase with middles age, noted Brenner and colleagues.

The next step entailed applying model-fitted estimated age dependent ERR results and sporadic cancer incidence data in the U.S. population to estimate absolute lifetime radiation-induced cancer risks per radiation dose in the US population as a function of age at exposure, explained the researchers.

The researchers estimated that excess lifetime cancer risks for all solid cancers and five of the six most radiogenic cancers (breast cancer excluded) slowly increases for older ages at exposure from approximately age 20 years through approximately age 60 years.

Brenner and colleagues referred to the sixfold increase in medical exposure to radiation in the last three decades and noted that the most common age range for CT studies is approximately 35 to 50 years. “When a CT scan is medically warranted, its benefits far outweigh any radiation risks, so that even increasing the estimated risks by a factor of two would not materially affect the risk-benefit balance,” they wrote.

The authors questioned CT screening of asymptomatic adults, particularly screening of the colon, lung and heart, and emphasized that excess lifetime risks of lung cancer may peak around age 50 years.

Brenner and colleagues acknowledged several limitations to the study. There are statistical uncertainties associated with the underlying atomic bomb survivor data, and there are possible other interpretations of the risk patterns.

Questioning lifetime risk models
In a related editorial, John D. Boice Jr., ScD, of International Epidemiology Institute in Rockville, Md., questioned the role of lifetime risk models. “[R]adiation is a relatively weak carcinogen when compared with certain chemical agents. … It remains important that the benefits accrued from radiation in our society be balanced with the costs in terms of any adverse health effects that might result following exposure,” he wrote.

Boice cited uncertainties related to generalizing Japanese incidence data to the U.S. population as well as contradictory evidence from medically exposed populations and called the conclusions “premature.” In contrast to Brenner’s results, he wrote, national and international committees (evaluating the same data) have all chosen alternative models that do not incorporate an increase in ERR at older ages.

Boice recommended a cautious approach to interpreting Japanese atomic bomb survivor incidence data, noting that age at exposure and birth cohort are confounded, so birth cohort factors could affect the risk pattern with age at exposure. In addition, evidence of increasing risk with older exposure age is weaker in Japanese mortality data than it is in the Japanese incidence data used in the risk assessment models.

Ultimately, the modeling “raises provocative hypotheses and conclusions that, although preliminary, draw attention to the continued importance of low-dose radiation exposures in our society,” wrote Boice.