Overdiagnosis of Thyroid Cancer and Graves' Disease

Dear Editor:

The incidence of thyroid cancer has tripled over the last 30 years in the United States from 3.5 in 100,000 person years in 1973 to 11.6 in 100,000 person years in 2009 (1). Yet, the incidence of thyroid cancer death over the same period remained stable at 0.5 in 100,000 person years (1). Neither new effective treatments nor the emergence of a subtype of nonlethal thyroid cancer can explain this phenomenon. The overdiagnosis of small indolent forms of papillary thyroid cancer could potentially explain this paradox. Two articles in the July 2013 issue of Thyroid addressed important topics related to thyroid cancer and overdiagnosis.

The article by Morris et al. (2) presented an ecological analysis of the association of county-level socioeconomic markers of healthcare access and the incidence of thyroid cancer. Chen et al. (3) conducted a national cohort study of 5025 newly diagnosed Graves' disease (GD) patients and 20,100 matched non-GD patients in Taiwan to evaluate the risk of thyroid cancer.

Morris et al. (2) showed that the incidence of thyroid cancer increased faster in patients who had healthcare insurance through Medicare and thus had more access to healthcare. Interestingly, they found an inflection point in the trend of thyroid cancer incidence at the year 1993. This inflection in the incidence corresponds to a similar inflection in the use of magnetic resonance imaging (MRI) and computed tomography (CT) (4). Since 1995, the numbers of available MRI and CT units grew by more than 50%, with a subsequent tripling in the number of MRI and CT procedures conducted among Medicare beneficiaries (4). Moreover, this inflection point also correlates with wide adoption of fine-needle aspiration biopsy in the 1990s, which allowed the cytology evaluation of a thyroid nodule at the bedside (5). The findings of Morris et al. (2) strengthen the hypothesis that the current epidemic of thyroid cancer without a corresponding increase in thyroid cancer death is a consequence of the aggressive detection of small indolent forms of papillary thyroid cancer.

We also read with great attention the report of Chen et al. (3). They conducted a national cohort study of 5025 newly diagnosed GD patients and 20,100 matched non-GD patients in Taiwan. They found that after adjusting for sex, age, and comorbidities, the hazard ratio for developing thyroid cancer was 16-fold higher for patients with GD during the first three years of their GD diagnosis and 10-fold higher at any follow-up time.

We could not avoid making a connection between the findings of these two reports. As opposed to the mechanisms proposed by the authors, is it not also possible that the higher number of thyroid cancers in the GD cohort is an example of overdiagnoses? Patients with GD are exposed to more thyroid examination than patients in the non-GD cohort by means of thyroid palpation, imaging, and/or pathologic examination. This scrutiny grants access to a large subclinical pool of thyroid nodules and thyroid cancers. Approximately 50% of people with a non-nodular thyroid by palpation have thyroid nodules upon macroscopic inspection (6), and approximately 5% to 10% of those nodules might harbor thyroid cancer (7). Have the authors obtained information as to how these cancers were diagnosed and what type of thyroid cancer was found? We hypothesize that the majority of these thyroid cancers were papillary, small, and were found postoperatively or in patients without symptoms related to thyroid nodules, the most indolent type of thyroid cancer. Perhaps GD is not a risk factor for thyroid cancer, as proposed by these authors, but rather a risk factor for finding thyroid cancer.

These two articles illustrate the need to understand better the phenomenon of overdiagnosis of thyroid cancer and its inevitable consequence—overtreatment.