Managing Small Thyroid Cancers

The management of thyroid nodules poses several diagnostic dilemmas. The vast majority of thyroid nodules are asymptomatic and impalpable, such that only 4% to 7% of the adult US population has palpable thyroid nodules.¹ However, neck ultrasonography reveals a large reservoir of clinically silent impalpable nodules, termed incidentalomas, the prevalence of which (both benign and malignant) ranges from 30% to 60% in autopsy studies and from 19% to 67% in prospective clinical studies.² Thyroid nodules increase in prevalence with advancing age and are more prevalent in women. An ultrasound screening study of 96 278 German employees aged 18 to 65 years found that the presence of a thyroid nodule was 13% in women and 9% in men aged 26 to 35 years, increasing to almost 45% in women and 32% in men aged 55 years and older.³ Moreover, thyroid cancer, which is relatively rare, is found in approximately 4% of fine-needle aspiration biopsy (FNAB) cytology specimens, will account for 30 000 new cases in the United States in 2006,² and, while usually associated with a good prognosis, will cause about 1500 deaths this year.⁴

This diagnostic scenario unfolds in a background of myriad small (often microscopic) asymptomatic thyroid cancers that lie dormant in the general population, never becoming manifest and usually best left undiagnosed. Still, some microcarcinomas are destined to become aggressive cancers.

Ultrasound-guided FNAB permits the identification of thyroid cancers smaller than 1 cm, and this increased level of detection likely accounts for the increasing incidence of thyroid cancer observed over the past few decades in several countries around the world.^{5 - 6} Many of these cancers are papillary thyroid microcarcinomas (PTMCs),^{7 - 8} defined by the World Health Organization as tumors measuring 1 cm or smaller⁹ that have especially good prognoses.

In this issue of JAMA, Davies and Welch¹⁰ report that from 1973 to 2002, the overall incidence of thyroid cancer in the United States increased 2.4-fold, a change virtually entirely attributable to an increase in incidence of papillary thyroid cancer (PTC); 49% of these cancers measured 1 cm or less (ie, were PTMCs), and 87% measured 2 cm or less. Moreover, during this time, the mortality from thyroid cancer has been stable at about 0.5 deaths per 100 000 population. The authors also found that 75% of persons with PTMC underwent total thyroidectomy and suggest that little is known about both the natural history of thyroid nodules identified by ultrasound and which patients need aggressive management. The authors also suggest that the increasing incidence of cancer reflects an increased detection of subclinical disease, not an increase in the true occurrence of thyroid cancer, and argue that, based on the known prevalence of asymptomatic small PTCs at autopsy, it is most likely that the observed increase is apparent, rather than real.

Davies and Welch attribute the increasing incidence of PTC to overdiagnosis or “increased diagnostic scrutiny,” which makes it difficult to identify which patients need treatment, but they also caution that this problem is only likely to get worse, suggesting that small asymptomatic thyroid nodules should be followed up for a period of time without immediately initiating diagnostic investigation. These are all important issues; however, the current literature provides some disturbing evidence about the natural history of PTMC and reason to believe that not all small PTC tumors are indolent bystanders caught up in a frenzy of excessive diagnosis and unnecessary surgery on small malignant nodules.

In the study by Davies and Welch, mortality rates, which are derived from death certificate data and are expressed as deaths per 100 000 persons, do not reflect the clinical changes among subgroups of patients. The National Cancer Institute's Surveillance, Epidemiology, and End Results database¹¹ shows that among women, the 5-year relative survival rates for thyroid cancer increased significantly, from 92.7% in 1974 to 97.4% in 2001 (P<.05). During this same time, the rates of distant metastases in men at the time of diagnosis of thyroid cancer were more than 2-fold those in women (9% vs 4%), and from 1992 to 2000 the annual percentage change in thyroid cancer mortality significantly increased in men by 2.4% (P<.05), the largest increase of any type of cancer.¹¹

The size of thyroid tumors is a powerful independent predictor of outcome.¹² The risk estimate is usually made according to a fixed cut point, generally around 1 cm for PTC and 4 cm for follicular thyroid carcinoma (FTC),⁹ above which adverse events increase statistically. This approach, however, does not provide a clear picture concerning the smallest thyroid tumor that can produce metastases, which is of fundamental importance to the clinician and patient.

In a longitudinal study of 480 patients, Machens et al¹³ estimated the cumulative risk of developing metastases as thyroid tumors grow larger, in effect demonstrating the smallest tumor size that begins to affect prognosis. The threshold for developing extrathyroidal tumor growth and lymph node metastases was 5 mm for PTC and 20 mm for FTC. The slope of the curves for predicting metastatic disease was significantly steeper for PTC than for FTC, indicating earlier extrathyroidal extension and lymph node metastases for patients with PTC compared with those with FTC. The threshold tumor size for developing pulmonary metastases was about 20 mm for both PTC and FTC and was 30 to 40 mm for developing bone metastases. The slopes for lung metastases were steeper than those for bone metastases, implying that lung metastases occur with smaller primary tumors at an earlier stage compared with bone metastases. However, in one large study¹⁴ of PTMC, 26% of patients with tumors 5 mm or smaller developed lymph node metastases, 1.4% had extrathyroidal extension, and 26% were multifocal, but none died of thyroid cancer.

Extrathyroidal tumor extension also predicts prognosis, as reflected in the sixth-edition TNM staging system.⁹ In a study¹² of 1501 patients with median follow-up of 16.6 years, extrathyroidal tumor extension was significantly associated with increased rate of tumor recurrence (hazard ratio, 1.4; 95% confidence interval, 1.1-2.2; P = .02) and with increased cancer-specific mortality (hazard ratio, 1.1; 95% confidence interval, 1.0-1.2; P = .002).

Multifocal PTC, which has been associated with distant metastases in some but not all studies,^{14 - 15} is found in about 20% of patients with PTC when the gland is examined routinely and in up to 80% if the thyroid is examined with great care.^{12 ,15} Long held to be intrathyroidal metastases, distinct tumor foci of multifocal PTC tumors recently have been shown to have independent clonal origins.¹⁵ New tumors arise independently, supporting the notion that most patients with PTC, even those with microscopic tumors, require total or near-total thyroidectomy. One large study¹⁴ of PTMC found that the risk of cervical lymph node recurrence increased 6.2-fold (P = .01) and 5.6-fold (P = .02) when lymph node metastases or multifocal disease, respectively, were present at the time of initial diagnosis. Also, the presence of lymph node metastasis was associated with increased risk of distant metastasis by 11.2-fold in this study¹⁴ and by 6-fold in another.¹²

The natural history of PTC plays out over decades.¹² A delay in the diagnosis often leads to larger tumors, higher rates of distant metastasis, and earlier cause-specific deaths. In one study,¹² cancer mortality was 4% when patients underwent initial therapy within a year of discovery of a palpable nodule and 10% when the delay was longer; 30-year cancer mortality rates in these 2 groups, which were 6% and 13%, respectively (P<.001), increased linearly over time of delay. Another study¹⁶ found that a single false-negative FNAB result delayed surgical treatment by 28 months, causing higher rates of vascular and capsular invasion and persistent disease (hazard ratio, 2.28). Ten-year thyroid cancer mortality rates in men are about twice those in women, largely because men present at an older age with more advanced tumors.⁹ This is likely due to late diagnosis. In a study from France,⁶ the proportion of women referred for evaluation of a thyroid nodule increased over the past 2 decades, but this increase did not occur in men; this finding was attributed to the high frequency of thyroid disorders in women and the way in which men use the health care system.

In the past few years, some of the natural history of papillary microcarcinomas has unfolded. A study of 1475 patients found that although most malignant incidentalomas were PTMCs with favorable postoperative TNM stages, 20% invaded the thyroid capsule and 12% involved lymph nodes.¹⁷ Another study of 267 patients found a 12% prevalence of malignancy in incidentalomas, 69% of which had either extrathyroidal extension or regional node involvement and 39% of which were multifocal.¹⁸ Another study of 203 patients with PTMC found that 21% had local tumor extension, 25% had lymph node metastases, 2.5% had distant metastases, and 1.0% died of cancer.¹⁴ And still another study found that among 243 patients with PTMC, the tumor was often bilateral (19%), multifocal (32%), and extended extrathyroidally (17%) into muscle or adipose tissue and metastasized to lymph nodes (13%) and/or distant sites (1.6%) without significant differences between incidentalomas and symptomatic tumors.¹⁹ Yet after a median follow-up of approximately 5 years, all patients were free of disease after undergoing total or near-total thyroidectomy and 1 or more iodine 131 (¹³¹I) treatments and in some cases external beam radiotherapy.

In a study of watchful waiting, 732 patients with PTMC were offered the option of forgoing immediate surgery, but most (approximately 78%) chose immediate surgery.²⁰ The other 162 patients, who comprised the study group, underwent surveillance for 1 to 5 years for tumors averaging approximately 7 mm in diameter. Over an average of almost 4 years, more than 70% of the tumors did not change or increase in size, and tumors in 5 patients disappeared completely. Most patients eventually opted for surgery, including some whose tumors increased in size or metastasized to lymph nodes during follow-up. Lymph node metastases were confirmed in almost half of the 626 patients who underwent surgery, and 43% had multifocal tumors. The recurrence rate was almost 3% at 5 years and 5% at 8 years after surgery, but no patients died of thyroid cancer. This study demonstrates that few patients will forgo surgery when they know they have cancer, even a tiny cancer that ordinarily has a good prognosis.

In a similar study of 15 patients, serial ultrasonography was performed over 20 to 85 (mean, 38) months on 27 thyroid nodules initially thought to be benign by FNAB; however, 15 nodules were found to be intrathyroidal PTCs smaller than 15 mm.²¹ The rate of nodule growth was greater in the benign than in the malignant nodules and thus did not prove useful in distinguishing one from the other.

Thus, despite the generally favorable prognosis for patients with PMC, cancer-related mortality rates may be as high as 1.0%, the rate of distant metastases as high as 2.5%, and rates of lymph node recurrence as high as 5%.¹⁴ Moreover, a few patients with aggressive tumors may not respond to standard therapy¹⁹ ; ie, with a PTC no larger than 1.5 cm, as many as 1 in 4 patients may develop relapsing or persisting disease after surgery.⁸ Still, several studies have shown that total thyroidectomy and ¹³¹I therapy can reduce the PTMC recurrence rate to zero.^{7 ,14} It is unlikely that many patients will forgo treatment after receiving this information. Once the decision is made to treat, it appears PTMC should be managed no differently than conventional PTC: complete thyroidectomy with consideration for ¹³¹I therapy if associated with a poor prognosis, such as an aggressive histological variant, multifocal tumor, and metastases; and external beam radiotherapy if extensive local tumor invasion is present.

Even though many patients might not require such aggressive therapy, at present it is not possible to confidently identify who they are. However, it is likely that in the future, measuring tumor BRAF mutations in FNAB specimens, the presence of which predicts a poorer clinical prognosis for PTC, may help resolve this conundrum.²² Until then, clinicians are left with performing FNAB in most patients with thyroid nodules that are larger than 1 cm or smaller nodules that have a suspicious ultrasonographic appearance and operating on those patients with cancer.

The American Thyroid Association evidence-based management guidelines for patients with thyroid nodules and differentiated thyroid cancer²³ have recommendations relevant to this debate. First, thyroid sonography should be performed in all patients with 1 or more suspected thyroid nodules; second, FNAB is the procedure of choice for evaluating thyroid nodules; and third, when several nodules larger than 1 to 1.5 cm are present, those with a suspicious sonographic appearance should be biopsied preferentially. Patients with nodules that are 8 to 9 mm in size and that have suspicious ultrasonographic findings (eg, blurred margins, mixed echogenicity, intranodular calcification, or Doppler flow), suspicious cervical lymph nodes, or a history of radiation exposure or familial thyroid cancer should be considered for ultrasound-guided FNAB. Smaller nodules most likely can be followed up over several years without FNAB if they are not increasing in size.