The relationship between radiation and thyroid carcinoma was first recognized in 1950, and thyroid carcinoma was the first solid malignant tumour found to be increased among Japanese atomic bomb survivors. This relationship was confirmed subsequently by many epidemiological studies.
Two major limitations should be taken into account in studies of the relationship between radiation and thyroid carcinoma. One is due to the fact that many patients are unaware of, or uncertain about, prior radiation exposure, especially when therapeutic irradiation was administered at a young age (recall bias), taking into account that radiation- induced thyroid carcinoma occurs several years later. The second and perhaps more relevant limitation is related to the frequent occurrence of thyroid nodules in the general population (4– 7% by palpation and up to 50% by ultrasonography in people over 60 years). Moreover, most thyroid tumours are indolent and frequently not recognized clinically. Thus, the diagnosis of thyroid tumours depends on the extent of the diagnostic procedures used (diagnostic bias).
In case- control studies, the cases are patients with thyroid cancer identified by entry into a tumour registry. The controls are matched subjects free from thyroid carcinoma. Information on risk factors, such as radiation exposure, is obtained and the distribution in the two groups is compared. In such studies diagnostic bias is minimized, but recall bias may be important. In cohort studies, exposure to radiation is generally well documented, and recall bias is minimized. The frequency of thyroid carcinoma in the radiation- exposed group is compared with a group of similar subjects not ex posed to radiation. In this case, diagnostic bias may be important. A final additional caveat is due to the fact that retrospective estimates of doses delivered to the thyroid are necessary to prove the aetiological weight of radiation in thyroid cancer (dose– effect relationship). These estimates may be difficult to obtain and are subject to error.
Most epidemiological studies dealing with the risk of developing radiation- induced thyroid cancer use the relative risk (RR) as an index, i.e. the ratio between the observed (O) number of cancers in the radiation- exposed group and the expected (E) number of cancers in the non- exposed group (RR = O/ E). When the expected number is obtained from a registry, the RR is called the standardized incidence ratio. The most frequently used indices of risk estimates are reported in Box 1.
Box1. Most common indices of cancer risk from radiation exposure
Risk estimates for radiation- induced thyroid cancer have been calculated in people exposed to external radiation. According to the National Council of Radiation Protection (NCRP), the excess absolute risk is 2.5 × 10– 4/ Gy per year for persons exposed under the age of 18. For adults, the risk per year is assumed to be half this value. Because of their smaller number of years at risk, the lifetime risk for adults is about one- quarter the risk for children.
In a pooled analysis the excess absolute risk was 4.4 × 10– 4/ Gy per year for persons exposed before the age of 15, confirming that the RR is largely dependent upon age at exposure with young children carrying the highest risk. As shown in Tables 1 and 2, little risk is carried after the age of 20 and almost none after the age of 40, as demonstrated in the study of atomic bomb survivors in Hiroshima and Nagasaki.
Table1. Thyroid cancer excess relative risk (ERR) from exposure to external radiation in adults
Table2. Thyroid cancer excess relative risk (ERR) from exposure to external radiation before the age of 20 years
