Abstract
In the letter to the editor written by Brito et al. (1), the authors weigh in on the debate as to whether thyroid cancer incidence trends reflect a true increase in thyroid tumors or an artifact of improved detection modalities and a subsequent detection bias. They cite two recently published articles in Thyroid (2,3) as evidence that thyroid cancer is being overdiagnosed and overtreated. We agree that improvements in technology, as well as increased access to better imaging, has likely been partly responsible for an increase in thyroid cancer detection. However, we would also argue that there is a substantial amount of evidence in the literature supporting the idea that something more than just improved diagnosis is impacting thyroid cancer trends.
Brito et al. address the increasing papillary thyroid cancer incidence trends in their introduction by stating, “neither new effective treatments nor the emergence of a subtype of nonlethal thyroid cancer can explain this phenomenon” (1). However, it is unclear why the emergence of a new subtype would be necessary to cause an increase in papillary thyroid cancer incidence. It is also possible that an unexpected or unidentified increase in exposure to an environmental risk factor such as radiation from medical imaging, endocrine disruptors in the environment, or even an unidentified environmental exposure could also lead to such an increase (4). Currently, there are no data directly linking increased environmental exposures to increased thyroid cancer rates. However, when one critically reviews the literature, there are also no studies that directly link the increased use of imaging and FNA to the higher thyroid cancer rates. All the studies on this subject have inferred such a link. However, no large-scale investigation of the direct cause–effect of FNA–thyroid cancer has been accomplished or a reliable estimate of the proportion of new thyroid cancers attributable to FNA reported.
In a recently published article in Thyroid (5), we reported on the papillary:follicular thyroid cancer incidence rate ratios. In this study, we illustrated a significant increase in the proportion of papillary thyroid tumors over time compared to follicular cancers. In fact, the rate of follicular cancers has barely increased over time. Because imaging studies cannot discriminate pathologic type, why would they not also increase the rate of follicular thyroid cancer detection as well? One argument is that there is not as large a “pool” of follicular cancers to be diagnosed. However, it is unlikely that, prior to these imaging modalities being available, 100% of follicular thyroid cancers were diagnosed in the past. Thus, a pool of undiagnosed follicular cancers likely also exists, yet they are not being discovered by the increased use of imaging. In fact, we would argue that the lack of an increase in follicular thyroid cancer rates along with the significant change in the relative distribution of the known thyroid tumor types over time is a very important part of the conversation that argues against overdiagnosis in this case.
The authors also suggest that better access to healthcare leads to higher rates of papillary thyroid cancer diagnosis by pointing out, “Morris et al. showed that the incidence of thyroid cancer increased faster in patients who had healthcare insurance through Medicare and thus, had more access to healthcare” (1). We would argue that age itself is a confounder in an analysis such as this because thyroid cancer is known to grow very slowly, and even though thyroid cancer does occur in young people, on average it is a cancer that is diagnosed later in life. It should also be noted that when evaluating the age-specific trends, a peak in thyroid incidence is not observed at the age of 65 (6). In the National Cancer Institute's Surveillance, Epidemiology, and End Results database, healthcare access and socioeconomic variables are available at the census-track level, whereas race/ethnicity and age is available for individuals. For these reasons, we believe race/ethnicity is a better surrogate for access to healthcare than being Medicare eligible, particularly in studies relating to thyroid cancer. In a recent article examining racial and ethnic differences in thyroid cancer diagnosis rates (7), we showed that the greatest annual percent change in thyroid cancer in the United States is observed among black women, a group traditionally known to have lower access to healthcare, particularly when compared to white and Asian women.
Brito et al. also state, “the findings of Morris et al. strengthen the hypothesis that the current epidemic of thyroid cancer is a consequence of the aggressive detection of small indolent forms of papillary thyroid cancer” (1). We again would argue that there is convincing evidence to the contrary in the literature (8). While microcarcinomas have increased (144.9% in women and 114.4% in men), an even greater percent increase is observed in large (≥4 cm) papillary thyroid cancers (150.7% in women and 107.9% in men). Most concerning, in the United States, there have been striking increases in the rate of metastatic thyroid cancer, both in the regional neck lymph node and distant metastasis (111.9% in women and 82.7% in men for regional disease and 64.3% in women and 60.9% in men for distant metastasis) (8). These are vast increases in metastatic cancer incidence that should not be ignored.
Finally, the surveillance argument against Graves' disease as a thyroid cancer risk factor is not convincing. We would suggest the opposite—that the Chen et al. article (3) presents evidence that the direct relationship between thyrotropin (TSH) and thyroid cancer should be investigated further using a longitudinal design and using molecular methods of hormone quantification. Although the chronic stimulation of the thyroid gland by TSH is a hypothesized pathway of thyroid carcinogenesis frequently referred to in the literature, a large prospective or nested case-control study has not been used to assess the relationship directly.
In a recent study, we reported that papillary thyroid cancer will be the third most common cancer in American women within the next decade following cancer of the breast and lung (9). We also reported that despite the substantial clinical and economic burden, thyroid cancer research remains significantly underfunded by the National Cancer Institute in comparison to many other cancer sites (9). The argument that the dramatic increase in thyroid cancer is due to overdiagnosis and its indolent nature are often used as a justification for why thyroid cancer research is so poorly funded. However, to date, the actual proportion of thyroid cancer cases that are due to detection bias has not been quantified. We are concerned that perpetuation of the idea that increasing thyroid cancer rates are solely related to overdiagnosis, without acknowledging that other reasons for the increase might exist, may cause harm by hindering the investigation into the causes of the “real” tumors. Research on better diagnostic accuracy to distinguish “real” from indolent tumors may also be hindered. We would suggest that the literature supports evidence that overdiagnosis exists concomitantly with a true increase in thyroid cancer rates, and that much more work needs to be done to truly understand these roles in thyroid cancer trends.