Radioactive Iodine–Related Clonal Hematopoiesis in Thyroid Cancer Is Common and Associated With Decreased Survival

Laura Boucai; John Falcone; Jenny Ukena; Catherine C. Coombs; Ahmet Zehir; Ryan Ptashkin; Michael F. Berger; Ross L. Levine; James A. Fagin


J Clin Endocrinol Metab. 2018;103(11):4216-4223. 

In This Article


In this study, we found that the prevalence of CH in patients with thyroid cancer was 37%, and that of potential drivers of hematologic malignancies (CH-PD) was 5.2%. Age was the strongest predictor of CH and CH-PD. For every year increase in age, there was a 5% and 13% increase in the odds of CH and CH-PD, respectively. Dose of RAI was significantly associated with CH and CH-PD, even after adjustment for age, EBRT, and chemotherapy. For every 10 mCi increase in the dose of RAI administered, there was a 2% and 4% increase in the odds of CH and CH-PD, respectively. Among patients who had received RAI, those with CH-PD had a significantly shorter OS, even when stratified by age.

With a rising incidence of thyroid carcinoma and the consequent large pool of long-term survivors who received RAI therapy, there is an understandable concern about the risk of secondary malignancies in general and hematologic malignancies in particular. Previous studies described the association between RAI and hematologic malignancies.[9,14,23–25] This report demonstrates a dose-dependent association between RAI and clonal events that have a low probability of leading to the development of hematologic malignancies.

Ionizing radiation induces chromosomal aberrations frequently found in secondary leukemias, and its causal contribution to leukemogenesis has generally been established.[26] Nevertheless, the link between RAI and the evolution to leukemia has long been a matter of debate, as early epidemiological studies yielded conflicting results.[27–29] In a comprehensive review and meta-analysis of the literature covering 16,502 patients with thyroid cancer, the relative risk for the development of leukemia increased 2.5-fold in patients treated with radioiodine.[30] In line with this, thyroid cancer was also identified as the second-most common primary neoplasm arising in patients who developed AML following treatment of solid cancers[31] and as the most common primary malignancy associated with CH.[19] Based on current evidence, CH is a precursor clonal state conferring a five- to 10-fold greater risk for development of hematological neoplasms compared with non-CH controls.[15,16,32,33] Therefore, the signal for association of CH with RAI exposure in our retrospective series justifies the development of a more rigorously designed study to ascertain dose response, time course, persistence of CH, and longitudinal risk for development of hematologic malignancies to confirm this association.

Besides assessing for risk, there is intriguing preclinical evidence suggesting that patients with CH harboring mutations of TET2 may benefit from treatment with vitamin C. Loss-of-function heterozygous mutations in TET2 occur frequently in patients with CH, myelodysplastic syndrome, and AML and are associated with DNA hypermethylation. Treatment with vitamin C, a cofactor of Fe2+ and α-ketoglutarate-dependent dioxygenases, can reverse the hypermethylation state by enhancing 5-hydroxymethylcytosine formation. In mouse models, this has been shown to suppress human leukemic colony formation and leukemia progression.[34,35]

CH and particularly CH-PD were associated with poorer survival. In contrast to CH, in a general population without a diagnosis of advanced cancer, where the primary cause of death is attributed to cardiovascular events, the main reason for this cohort's demise was progression of its primary thyroid carcinoma. The association between CH and cancer progression could be a result of cell–cell interactions among mutant myeloid clones and cancer cells,[36] impact of CH on immune surveillance, or other factors that will require further clinical and functional investigation.

The findings of our study are limited by the fact that our cohort was composed of thyroid cancer patients with advanced disease stage and older age. The high prevalence of CH may be related to the increased age of our population compared with that of subjects with other solid tumors. In addition, the low incidence of hematologic malignancies prevented us from exploring the association among RAI, CH, and leukemia, which may have been a result of the short follow-up time after sequencing of their blood, compounded by the fact that mortality was high in this population with advanced thyroid cancer. Furthermore, the association between CH-PD and decreased survival among patients exposed to RAI may be related to the advanced stage of the tumors genotyped. It remains unclear if the association of CH with poor prognosis would also be seen in younger patients with more indolent disease.

In conclusion, the model that has been proposed[19] and further validated here is that age-dependent mutations are constantly acquired in hematopoietic stem cells, which are then selected for by external perturbations, such as RAI, which allow clones to expand. These mutant hematopoietic clones may impact the biology and therapeutic response of primary tumors, besides conferring an increased risk of hematologic malignancies. RAI is associated with a high prevalence of CH, and CH is a precursor state of hematologic malignancies. The extent to which the presence of CH should be considered before administration of RAI is unclear. Furthermore, identification of CH and particularly of CH-PD may inform lifestyle interventions and the use of molecularly targeted therapies to prevent clonal expansion and subsequent hematologic malignancies.