Radiation Exposure and Benign Thyroid Nodules

The 1986 Chornobyl accident was a catastrophe, but in the subsequent decades much has been learned from its consequences. Sienko, et al. has comprehensively reviewed radiation-induced thyroid cancer including an excellent account of Chornobyl findings, putting them into the context of the findings from other studies. ¹

In this issue of thyroid, Cahoon, et al. confirm previous results about Chornobyl neoplasms and extends them by asking if benign thyroid nodules (other than follicular adenomas) are associated with radiation exposure. ² Presently, the association between radiation exposure and benign thyroid nodules is generally accepted, while the results of Cahoon, et al., creates uncertainty about whether this is really the case. It is relevant to note that, in their landmark article Morton et al. found that the thyroid cancers in Ukrainian children exposed to Chornobyl radiation had “…radiation dose-dependent increases in clonal (emphasis added) DNA double-strand breaks.” ³ Thus, it is counterintuitive to suspect that radiation could be associated with, and presumably cause, benign (non-clonal) thyroid nodules. The uncertainty raised by Cahoon, et al. emphasizes the need to consider the changing factors affecting the ascertainment of benign thyroid nodules.

The 358 individuals in the Cahoon, et al. study were people with thyroid nodules diagnosed among 13,078 Ukrainian individuals <18 years old at the time of the accident who were screened, including with ultrasound, during 1998–2000. Based on the findings, some were then referred for fine needle aspiration. They were retrospectively divided into three groups designated by the authors as neoplastic (n = 65), suspicious (n = 115), and non-neoplastic, excluding follicular adenomas (n = 178). Dose-response EOR (excess odds ratios) relationships were determined for each of the three categories.

For each of the categories, there was a significant dose-response relationship between radiation exposure and the presence of thyroid nodules. However, when cases with doses ≥5 Gy were excluded from the non-neoplastic group (4 of 178 individuals) the trend was no longer significant. Specifically, for doses <5 Gy. EORs were 5.35 (2.19, 15.5) for the neoplastic, 1.44 (0.57, 3.10) for the suspicious, and 0.24 (−0.07, 0.74) for the benign nodules. Thus, the authors’ main conclusion is clear, that in the Chornobyl cohort “the risk for neoplastic nodules is higher than for non-neoplastic nodules.”

But is there a radiation risk for non-neoplastic nodules at all? The authors imply, but do not emphasize, that there is such a risk, “An elevated EOR/Gy was found for all case groups in the full dose range.” However, there are reasons to be cautious about using the authors’ findings to determine whether Chornobyl radiation doses are associated with or even caused the non-neoplastic nodules.

The main reason for caution is the heterogeneity in the three groups. Each group is composed of two or more subgroups. For example, the non-neoplastic group includes (a) 136 cases defined by their fine needle biopsies, (b) 12 cases defined by fine needle biopsies followed by surgery, and (c) 30 cases who had nodules with “no suspicious characteristics” and so were not referred for fine needle biopsies. It is quite likely that this heterogeneity has resulted in misclassification of some degree, especially between the suspicious and non-neoplastic groups. The second reason is that the EOR for the non-neoplastic group is small, especially in comparison with the neoplastic group.

A series of previous studies of children who were irradiated in Chicago for benign conditions in the head and neck area from 1939 through the early 1960’s and examined during 1974–1975 to address radiation associations were performed. The dose-response relationships for malignant and benign nodules (with histological confirmation) were nearly identical. ⁴ These studies were performed before thyroid ultrasound existed. The use of ultrasound in the standardized screening in Cahoon, et al. would have captured many more small thyroid nodules compared with the Chicago cohort, likely explaining the difference. Similarly, the exams in Cahoon, et al. were performed before diagnostic molecular tests were available. Currently, the methods for detecting thyroid nodules and determining their nature include improved ultrasound interpretation and the use of genomic testing. These changes limit the clinical significance of retrospective data following the Chornobyl accident. The authors look forward to extending the study by including the results of subsequently completed rounds of screening. This will refine determinations of the magnitudes of the associations reported in the current report. However, the expanded data would remain retrospective.

While it is certain that radiation causes thyroid cancer, because of the findings of Cahoon et al., we conclude that the generally accepted view that radiation also causes benign thyroid nodules is no longer clear.

What about the future? The authors suggest pooling data from the NCI-supported Ukraine–American (UkrAm) and Belarus–American (BelAm) projects. This has been an aim from the outset of these two projects. However, given the current geopolitical situation, it seems increasingly unlikely that this will occur any time soon. Also, extending the clinical follow-up of the Ukrainian study population using current diagnostic methods is probably not realistic while the country is at war. Genetic testing of non-neoplastic nodules from Chornobyl or other radiation-related studies, if available, would be revealing.