Breast cancer risk 55+ years after irradiation for an enlarged thymus and its implications for early childhood medical irradiation today

Abstract

Background: Radiotherapy during childhood increases long-term cancer risk, but the risk from radiation as a result of relatively higher dose diagnostic procedures remains less well known. This study, which evaluates breast cancer incidence in a cohort treated with "lower dose" chest radiotherapy over 50 years ago, can assist with estimating lifetime breast cancer risk in young children exposed to radiation from procedures such as chest computed tomography (CT) or treatment with recent "lower dose" chest radiotherapy protocols.

Methods: A population-based, longitudinal cohort of subjects exposed to thymic irradiation during infancy from 1926 to 1957 and of their unexposed siblings was re-established. Previously followed until 1987, we resurveyed cohort members from 2004 to 2008. Poisson regression models compared breast cancer incidence rates between women in the cohort by treatment and dose category groups.

Results: Breast cancer occurred in 96 treated (mean breast dose, 0.71 Gy) and 57 untreated women during 159,459 person-years of follow-up. After adjusting for attained age and treatment/birth cohort, the rate ratio was 3.01 (2.18-4.21). The adjusted excess relative risk per Gy was 1.10 (95% confidence interval, 0.61-1.86). Traditional breast cancer risk factors did not contribute significantly to multivariate model fit.

Conclusion: Our results show that at radiation doses between those received by the breast from chest CT and cancer therapy during early childhood, breast cancer incidence rates remain elevated >50 years after exposure. This implies that increased breast cancer risk will remain a lifelong concern in females treated during childhood with currently reduced radiotherapy doses and for infants receiving multiple chest CTs.

Figure 1

Radiation dose response for breast cancer incidence among 3449 women in the Rochester, NY Thymus Irradiation Cohort, with known thymic irradiation dose. The analysis is adjusted for attained age (which is highly correlated with time since radiation exposure) and treatment/birth cohort. Dose intervals correspond with categories in Table 4 with data points at mean dose of each category and cross bars representing 95% confidence intervals. Solid line represents best fit linear regression line with dotted lines representing upper and lower 95% CI of regression line