Epidemiology

Thyroid nodules are approximately 4 times more common in women than in men. Palpable nodules increase in frequency throughout life, reaching a prevalence of approximately 5% in the United States population aged 50 years and older.^1–3 Nodules are even more prevalent when the thyroid gland is examined at autopsy or surgery, or when using ultrasonography, and 50% of these have nodules, which are almost always benign.^2,4 New nodules develop at a rate of approximately 0.1% per year beginning in early life, but at a much higher rate (∼2% per year) after exposure to head and neck irradiation.^5,6

By contrast, thyroid carcinoma is uncommon. For the United States population, the lifetime risk of being diagnosed with thyroid carcinoma is less than 1% (0.83% for women and 0.33% for men).⁷ Approximately 37,200 new cases of thyroid carcinoma were diagnosed in the United States in 2009.⁸

As with thyroid nodules, thyroid carcinoma occurs 2 to 3 times more often in women than in men. With the incidence increasing by 6.2% per year, thyroid carcinoma is currently the sixth most common malignancy diagnosed in women.⁸ Among persons age 15 to 24 years, thyroid carcinoma accounts for 7.5% to 10% of all diagnosed malignancies.^9–11 The disease is also diagnosed more often in white North Americans than in African Americans. Although thyroid carcinoma can occur at any age, the peak incidence from 2004 to 2006 was near age 45 to 49 years in women and 65 to 69 years in men.⁷

Thyroid carcinoma has 3 main histologic types: differentiated (including papillary, follicular, and Hürthle cell), medullary, and anaplastic (aggressive undifferentiated tumor). Of 53,856 patients treated for thyroid carcinoma between 1985 and 1995, 80% had papillary carcinoma, 11% had follicular carcinoma, 3% had Hürthle cell carcinoma, 4% had medullary carcinoma, and 2% had anaplastic thyroid carcinoma.¹² The 10-year relative survival rates for patients with papillary, follicular, and Hürthle cell carcinomas were 93%, 85%, and 76%, respectively.¹² This discussion focuses on papillary, follicular, Hürthle cell, and anaplastic thyroid carcinoma. Medullary thyroid carcinoma was previously published in this journal (May 2010). A complete discussion of thyroid carcinoma, including medullary thyroid carcinoma, is available online on the NCCN web site (www.NCCN.org).

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

View Full Size

NCCN Clinical Practice Guidelines in Oncology: Thyroid Carcinoma

Version 1.2010, 01-14-10 ©2010 National Comprehensive Cancer Network, Inc. All rights reserved. These guidelines and this illustration may not be reproduced in any form without the express written permission of NCCN.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 11; 10.6004/jnccn.2010.0093

In 2009, approximately 1630 cancer deaths occured among persons living with thyroid carcinoma in the United States.⁸ Anaplastic thyroid carcinoma is almost uniformly lethal; however, most thyroid carcinoma deaths are from papillary, follicular, and Hürthle cell carcinomas, which account for nearly 95% of all thyroid carcinoma cases. Although thyroid carcinoma occurs more often in women, mortality rates are higher for men, probably because men are usually older at diagnosis.^7,13

The incidence of thyroid carcinoma increased almost 310% between 1950 and 2004, but mortality rates decreased more than 44%.⁷ From 1975 to 2004, thyroid cancer rates in the United States doubled.¹⁴ Because overall mortality has remained stable since 1975, the increasing incidence probably partially reflects earlier detection of subclinical disease (i.e., small papillary cancers), although even microcarcinomas can metastasize regionally, thereby increasing eventual recurrence risk.^14,15 However, recent data show the incidence has increased across all tumor sizes.^16,17 The stable age- and gender-adjusted mortality rate for thyroid carcinoma contrasts distinctly with the declining rates for other solid tumors in adults.⁸

The Challenge of Managing Differentiated Thyroid Carcinoma

Managing differentiated (i.e., papillary, follicular, and Hürthle cell) thyroid carcinoma can be a challenge, because no prospective randomized trials of treatment have been performed. Results from ongoing randomized trials will not be available for many years, given the typically prolonged course and relative infrequency of these tumors. Most of the information about treatment comes from studies of large patient cohorts in which therapy has not been randomly assigned. This accounts for much of the disagreement about managing differentiated carcinoma.

Nonetheless, most patients can be cured of this disease when properly treated by experienced physicians and surgeons.¹⁸ The preferred treatment is surgery, whenever possible, followed by radioiodine (¹³¹I) and thyroxine therapy in many patients. Externalbeam radiation therapy (RT) and chemotherapy have less prominent roles in managing these tumors.

Radiation-Induced Thyroid Carcinoma

Exposure to ionizing radiation is the only known environmental cause of thyroid carcinoma, usually causing papillary carcinoma. The thyroid glands of children are especially vulnerable to the carcinogenic action of ionizing radiation. A child's thyroid gland has one of the highest risks for developing cancer of any organ. In fact, the thyroid gland is the only organ linked to risk at approximately 0.10 Gy by convincing evidence.⁵ The risk for radiation-induced thyroid carcinoma is greater in women, certain Jewish populations, and patients with a family history of thyroid carcinoma,¹⁹ suggesting that genetic factors are also important in its development. Beginning within 5 years of irradiation, new nodules develop at a rate of approximately 2% annually, reaching a peak incidence within 30 years of irradiation but remaining high at 40 years.^5,6

In adults, the risk of developing thyroid carcinoma after exposure to ¹³¹I seems to be small or nonexistent.²⁰ After the Chernobyl nuclear reactor accident in 1986, many children developed papillary thyroid carcinoma after exposure to radioiodine fallout. It became evident that ¹³¹I and other short-lived radioiodines were potent thyroid carcinogens in children, particularly those younger than 10 years at exposure.²¹ Although radiation-induced papillary thyroid cancer usually appears more aggressive histologically and has high recurrence rates, the prognosis for survival is not clearly different from that of spontaneously occurring tumors.^22,23 Iodine deficiency is associated with follicular and anaplastic thyroid carcinoma.

Clinical Presentation and Diagnosis

Differentiated (i.e., papillary, follicular, or Hürthle cell) thyroid carcinoma is usually asymptomatic for long periods and commonly presents as a solitary thyroid nodule. However, evaluating all nodules for malignancy is difficult, because benign nodules are so prevalent and because thyroid carcinoma is so uncommon.^1,24,25 Moreover, both benign and malignant thyroid nodules are usually asymptomatic, giving no clinical clue to their diagnosis. Approximately 50% of the malignant nodules are discovered during a routine physical examination, by serendipity on imaging studies, or during surgery for benign disease. The other 50% are usually noticed first by the patient, usually as asymptomatic nodules.^1,24 Regrettably, the typically indolent nature of differentiated thyroid carcinoma often leads to long delays in diagnosis that may substantially worsen the course of the disease.¹³

Factors Affecting Risk for Malignancy

Nodule size affects the risk for malignancy and the clinical evaluation. Thyroid nodules smaller than 1 cm occur with such frequency in the asymptomatic general population that they are often found serendipitously during imaging studies for other head or neck problems. Often termed incidentalomas, nodules smaller than 1 cm are typically clinically benign lesions and usually do not require biopsy, unless there are suspicious findings (see page 1231).^4,26,27 In select cases, following up these nodules with serial ultrasounds may be reasonable. By contrast, nodules larger than 4 cm in diameter pose a somewhat higher risk for malignancy. Fine-needle aspiration (FNA) is the preferred procedure for evaluating suspicious thyroid nodules.^3,25

The Society of Radiologists in Ultrasound wrote a consensus statement about managing thyroid nodules identified at thyroid ultrasonography. Their recommendations describe which nodules should undergo FNA based on nodule size and ultrasound characteristics, and on clinical features that might predict risk for morbidity from an undiagnosed malignancy.²⁸ Suspicious criteria found using ultrasound include central hypervascularity, microcalcifications, and irregular borders.

Although more than 50% of all malignant nodules are asymptomatic, the pretest probability of malignancy in a nodule increases considerably when signs or symptoms are present (see page 1230).²⁹ For example, the likelihood that a nodule is malignant increases approximately 7-fold if it is very firm, is fixed to adjacent structures, is growing rapidly, is associated with enlarged regional lymph nodes, or causes vocal cord paralysis, or if symptoms of invasion into neck structures are present.^29,30 Family history of thyroid cancer is also indicative of malignancy. If 2 or more of these features are present, the likelihood of thyroid cancer is virtually assured; however, this is a rare situation.³⁰

A patient's age and gender also affect the probability of malignancy. The risk of malignancy is higher in patients younger than 15 years and in men. Other factors that increase the suspicion of malignancy include 1) a history of head and neck irradiation; 2) a family history of diseases associated with thyroid carcinoma, such as, familial adenomatous polyposis (formerly called Gardner's syndrome), Carney complex, Cowden's syndrome, and multiple endocrine neoplasia (MEN) types 2A or 2B; 3) evidence of other thyroid cancer–associated diseases or syndromes, such as hyperparathyroidism, pheochromocytoma, marfanoid habitus, and mucosal neuromas (MEN2B), which make the presence of medullary thyroid cancer more likely; or 4) the presence of suspicious findings detected with imaging, such as focal FDG (18-fluorodeoxyglucose) uptake on PET, or central hypervascularity, irregular border, and/or microcalcifications on ultrasound.³¹

Initial Workup

FNA of the nodule and clinically suspicious lymph nodes is recommended as the first diagnostic test in a clinically euthyroid patient before any imaging studies are performed.^1,3 Ideally, the serum thyrotropin (thyroid-stimulating hormone [TSH]) results should be known before FNA is performed. This is often impractical, however, and FNA may be performed during the initial office visit. Recent data show that higher TSH levels are associated with risk for differentiated thyroid cancer.³²

Some clinicians, especially in Europe,³³ recommend obtaining serum calcitonin levels from all patients with thyroid nodules. However, controversy surrounds the cost-effectiveness of this practice in the United States, especially in the absence of confirmatory pentagastrin stimulation testing, and the assumptions used in cost-effective analyses. The American Thyroid Association is equivocal about measuring serum calcitonin.^3,34 A recent study showed that calcitonin screening may be cost-effective in the United States.³⁵ However, false-positive calcitonin readings that can result from minimal elevations can only be ruled out with pentagastrin testing, and pentagastrin is not available in the United States. Ultrasound of the thyroid and central neck is also recommended,³⁶ and can also be performed on the lateral neck (category 2B).

Cytologic examination of an FNA specimen is typically categorized as 1) carcinoma (papillary, medullary, or anaplastic) or suspicious for malignancy; 2) follicular or Hürthle cell neoplasm; 3) follicular lesion of undetermined significance; 4) thyroid lymphoma; 5) benign (i.e., nodular goiter, colloid goiter, hyperplastic/adenomatoid nodule, Hashimoto's thyroiditis); or 6) insufficient biopsy (nondiagnostic) (see page 1231). These diagnostic categories for FNA results reflect the NCI's State-of-the-Science Conference held in 2007.³⁷

Pathology and cytopathology slides should be reviewed at the treating institution by a pathologist with expertise in diagnosing thyroid disorders. Although FNA is a very sensitive test, particularly for papillary carcinoma, false-negative results are sometimes obtained; therefore, a reassuring FNA should not override concerns in the presence of worrisome clinical findings.³⁸ Hürthle cell neoplasms can sometimes mimic medullary carcinoma cytologically and on frozen section. Discriminating between anaplastic thyroid cancer and other primary thyroid malignancies (e.g., medullary carcinoma, thyroid lymphoma) or poorly differentiated cancer metastatic to the thyroid can sometimes be difficult.³⁹ Metastatic renal carcinoma can mimic a follicular neoplasm, melanoma can mimic medullary carcinoma, and metastatic lung cancer can mimic anaplastic carcinoma of the thyroid (http://thyroidfna.cancer.gov/pages/conclusions/).

Pathology synoptic reports (protocols) are useful for reporting results from examinations of surgical specimens. The NCCN Thyroid Carcinoma Panel supports pathology synoptic reports from the College of American Pathologists (CAP) and the Association of Directors of Anatomic and Surgical Pathology (ADASP). Some pathologists currently use a modified format that is believed to comply with both of these synoptic reports. Although no ADASP checklist has been published for thyroid carcinoma, the CAP protocol information and checklists (updated in October 2009) can be accessed on the CAP Web site (http://www.cap.org).

On January 1, 2004, the Commission on Cancer (COC) of the American College of Surgeons mandated the use of specific checklist elements of the protocols as part of its Cancer Program Standards for Approved Cancer Programs. Therefore, pathologists should familiarize themselves with these documents. The CAP protocol checklist complies with the COC requirements.

FNA is far less able to discriminate follicular and Hürthle cell carcinomas from benign adenomas, because diagnosis of these malignancies requires demonstration of vascular or capsular invasion.²⁵ Thus, follicular and Hürthle cell carcinomas are rarely diagnosed on FNA.^18,40 Nodules that yield an abundance of follicular cells with little or no colloid are nearly impossible to categorize as benign or malignant based on FNA. Surgical biopsy is advisable, because approximately 20% of these lesions are follicular carcinomas.²⁹ Male gender, older patient age, and larger nodule size may increase the likelihood of a malignant diagnosis at surgery as high as 80%, whereas female gender, younger age, and smaller nodule size may reduce the risk as low as 5%. Repeat FNA will not resolve the diagnostic dilemma. Before thyroidectomy is performed, however, serum TSH level and thyroid ¹²³I or 99m technetium scanning may identify patients with an autonomously functioning or “hot” nodule who often may be spared surgery, because the diagnosis of follicular adenoma is highly likely.⁴¹

Clinically euthyroid patients with a low TSH and a hot nodule on thyroid imaging should be evaluated and treated for thyrotoxicosis as indicated, even when cytology is suspicious for follicular neoplasm; those with a “cold” nodule should proceed to surgery (see page 1231).^2,3 Patients with a high or normal TSH and cytology suspicious for follicular or Hürthle cell neoplasm should undergo diagnostic lobectomy. Total thyroidectomy should be considered for bilateral disease, unilateral disease greater than 4 cm (especially in men), or if the patient prefers this approach.

An FNA that yields insufficient cellular material for diagnosis and is solid should be repeated, because approximately 50% of subsequent specimens are adequate to assign a diagnosis.²⁹ In patients with serial nondiagnostic aspirates, 5% of women and 30% of men may have malignant nodules.⁴² Nodules yielding benign cytology do not require repeat FNA unless the nodules show evidence of growth.²⁹

When a diagnosis of thyroid carcinoma is promptly established using FNA, the tumor is often confined to the thyroid or has metastasized only to regional nodes, thus providing ample opportunity for cure. However, as many as 5% of patients with papillary carcinoma and up to 10% of those with follicular or Hürthle cell carcinoma have tumors that aggressively invade structures in the neck or have produced distant metastases. These cancers are difficult to cure.

Recurrence of Differentiated Thyroid Carcinoma

Depending on initial therapy and other prognostic variables, approximately 30% of patients with differentiated thyroid carcinoma have tumor recurrences during several decades; 66% of these recurrences occur within the first decade after initial therapy.¹³ Although not usually fatal, a recurrence in the neck is serious and must be regarded as the first sign of a potentially lethal outcome.^43,44 In one large study, central neck recurrences were seen most often in the cervical lymph nodes (74%), followed by the thyroid remnant (20%), and then the trachea or muscle (6%). Of the group with local recurrences, 8% eventually died of cancer.¹³ Distant metastases were the sites of recurrence in 21% of this patient cohort, most often (63%) in the lungs alone. Of the patients with distant metastases, 50% died of cancer.¹³

Age, Stage, and Sex at Diagnosis

Although many factors influence the outcome of patients with papillary and follicular thyroid carcinomas, patient age at the time of initial therapy and tumor stage are important.^13,45–47 Age is the most important prognostic variable for thyroid cancer mortality. However, thyroid cancer is more aggressive in men. Thyroid carcinoma is more lethal in patients older than 40 years, and is increasingly so with each subsequent decade of life. The mortality rate increases dramatically after age 60 years (see Figure 1; available online, in these guidelines, at www.NCCN.org [MS-29]). However, tumor recurrence shows a remarkably different behavior with respect to age. Recurrence frequencies are highest (40%) among patients younger than 20 or older than 60 years, whereas at other ages it occurs in only approximately 20% of patients.^13,45–48 This disparity between cancer-related mortality and the frequency of tumor recurrence probably accounts for most of the profound disparity of opinion among clinicians concerning optimal treatment for patients with differentiated thyroid cancer. How clinicians assess the importance of tumor recurrence (as opposed to cancer-specific survival) accounts for much of the debate surrounding the influence of age on the treatment plan for children and young adults.

Children typically present with more advanced disease and experience more tumor recurrences after therapy than adults, but their prognosis for survival is good.^49,50 Although the prognosis of children with thyroid carcinoma is favorable for long-term survival (90% at 20 years), the standardized mortality ratio is 8-fold higher than predicted.⁵¹ Some authors believe that young age affects survival so favorably that it overshadows the behavior expected from the characteristics of the tumor. Therefore, they classify most thyroid tumors as low-risk tumors that may be treated with lobectomy alone.^52–54 However, most treating physicians believe that tumor stage and its histologic features should be as significant as patient age in determining management.^13,49,55,56

Prognosis is less favorable in men, but the difference is usually small.^13,54 One study found that gender was an independent prognostic variable for survival and that the risk for death from cancer was approximately twice as high in men as in women.¹³ Because of this risk factor, men with thyroid carcinoma, especially those older than 40 years, may be regarded with special concern.⁵⁷

Familial Syndromes

Familial, nonmedullary thyroid carcinoma accounts for about 5% of papillary carcinomas and, in some cases, may be clinically more aggressive than the sporadic form.^58,59 Microscopic familial papillary thyroid carcinoma tends to be multifocal and bilateral, often with vascular invasion, lymph node metastases, and high rates of recurrence and distant metastases.⁶⁰ Other familial syndromes associated with papillary thyroid carcinoma are familial adenomatous polyposis,⁶¹ Carney complex (multiple neoplasia and lentiginosis syndrome, which affects endocrine glands),⁶² and Cowden's syndrome (multiple hamartomas).⁶³ The prognosis for all of these syndromes is not different from that for spontaneously occurring papillary thyroid carcinoma.

Tumor Variables Affecting Prognosis

Some tumor features have a profound influence on prognosis.^48,64–66 The most important features are tumor histology, primary tumor size, local invasion, necrosis, vascular invasion, BRAF mutation status, and metastases.

Histology: Although survival rates with typical papillary carcinoma are good, cancer-specific mortality rates vary considerably with certain histologic subsets of tumors.¹ A well-defined tumor capsule, which is found in approximately 10% of papillary thyroid carcinomas, is a particularly favorable prognostic indicator. A worse prognosis is associated with 1) anaplastic tumor transformation; 2) tall-cell papillary variants, which have a 10-year mortality of up to 25%; 3) columnar variant papillary carcinoma (a rapidly growing tumor with a high mortality rate); and 4) diffuse sclerosing variants, which infiltrate the entire gland.^18,67 Follicular-variant papillary thyroid carcinoma (FVPTC), which is recognized by its follicular architecture and typical papillary cytology, does not seem to have a worse prognosis than the pure papillary lesions if the FVPTC is encapsulated.^48,67–69

Follicular carcinoma is typically a solitary encapsulated tumor that may be more aggressive than papillary carcinoma. It usually has a microfollicular histologic pattern, and is identified as cancer through follicular cell invasion of the tumor capsule and/or blood vessels. The latter has a worse prognosis than capsular penetration alone.⁷⁰ Many follicular carcinomas are minimally invasive tumors, exhibiting only slight tumor capsular penetration without vascular invasion. They closely resemble follicular adenomas and are less likely to produce distant metastases or cause death.⁷¹ FNA or frozen section study cannot differentiate a minimally invasive follicular carcinoma from a follicular adenoma. Therefore, the tumor is often simply referred to as a follicular neoplasm by the cytopathologist (see page 1231). Cancer may be diagnosed only after thyroidectomy and when analysis of the permanent histologic sections shows tumor capsule invasion by follicular cells.

Highly invasive follicular carcinomas are much less common; they are sometimes recognized at surgery by their invasion of surrounding tissues and extensive invasion of blood vessels. Up to 80% of these cancers metastasize, causing death in as many as 20% of patients, often within a few years of diagnosis.⁴⁸ The poor prognosis is closely related to the patient's older age at diagnosis, advanced tumor stage, and larger tumor size.¹³

The mortality for papillary and follicular carcinomas is similar in patients of comparable age and disease stage. Both cancers have an excellent prognosis if the tumors are confined to the thyroid, small, and minimally invasive. Papillary and follicular carcinomas have far less favorable outcomes if they are highly invasive or develop distant metastases.^13,72 Staging for patients with papillary and follicular carcinoma who are older than 45 years was revised in the 2002 guidelines (6th edition) from the American Joint Commission on Cancer (AJCC; available online, in these guidelines, at www.NCCN.org [ST-1]).⁷³ Note that new staging guidelines from the AJCC (7th edition) were effective after January 1, 2010, and will be reflected in the 2011 NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for Thyroid Carcinoma (available at www.NCCN.org).⁷⁴ Many studies (including those discussed in this manuscript) have used AJCC-TNM staging from earlier editions, such as the 5th edition,⁷⁵ and not the 6th or 7th editions.^73,74

When Hürthle (oncocytic) cells constitute most or all of a malignant tumor's mass, the disease is often classified as Hürthle cell carcinoma, although the WHO classification considers it a variant of follicular carcinoma.⁷⁶ Molecular studies suggest, however, that this tumor may be more similar to papillary than follicular carcinomas.⁷⁷ Benign and malignant Hürthle cell tumors usually cannot be discriminated by FNA or frozen section examination, although large (> 4 cm) tumors are more likely to be malignant than smaller ones.⁷⁸ Hürthle cell carcinomas may be aggressive, especially when vascular invasion or large tumors occur in older patients.^79,80 Some believe these cancers are not much more aggressive than similarly staged follicular carcinomas without Hürthle cells.⁸¹ In the National Cancer Data Base report, the 10-year relative survival rates were 85% for follicular carcinomas and 76% for Hürthle cell carcinoma.¹²

In 2 large series, pulmonary metastases occurred in 25% and 35% of patients with Hürthle cell carcinoma, approximately twice the frequency of follicular carcinoma metastases.^82,83 Fewer Hürthle cell carcinomas concentrate ¹³¹I compared with papillary or follicular carcinomas. In a series of 100 patients with distant metastases, ¹³¹I uptake by pulmonary metastases was seen in more than 50% of the follicular (64%) and papillary (60%) carcinomas but in only 36% of Hürthle cell carcinomas.⁸⁴

Primary Tumor Size: Papillary carcinomas smaller than 1 cm, termed incidentalomas or microcarcinomas, are typically found incidentally after surgery for benign thyroid conditions. Their recurrence and cancer-specific mortality rates are near zero.⁸⁵

Other small papillary carcinomas become clinically apparent. For example, approximately 20% of microcarcinomas are multifocal tumors that commonly metastasize to cervical lymph nodes. Some researchers report a 60% rate of nodal metastases from multifocal microcarcinomas,⁸⁶ which may be the presenting feature and also may be associated with distant metastases.⁸⁵ Otherwise, small (< 1.5 cm) papillary or follicular carcinomas confined to the thyroid almost never cause distant metastases. Furthermore, recurrence rates after 30 years are one third of those associated with larger tumors; 30-year cancer-specific mortality is 0.4% compared with 7% (P < .001) for tumors 1.5 cm or larger.¹³ In fact, the prognosis for papillary and follicular carcinomas is incrementally poorer as tumors increase in size.^72,87 A linear relationship exists between tumor size and recurrence or cancer-specific mortality for both papillary and follicular carcinomas (Figure 2; available online, in these guidelines, at www.NCCN.org [MS-29]).¹³

Local Tumor Invasion: Up to 10% of differentiated thyroid carcinomas invade through the outer border of the gland and grow directly into surrounding tissues, increasing both morbidity and mortality. Local invasion may be microscopic or gross, and can occur with both papillary and follicular carcinomas.^13,88 Recurrence rates are 2 times higher with locally invasive tumors, and as many as 33% of patients with these tumors die of cancer within a decade.^13,89

Lymph Node Metastases: In one review, nodal metastases were found in 36% of 8029 adults with papillary carcinoma, 17% of 1540 patients with follicular carcinoma, and up to 80% of children with papillary carcinoma.⁴⁸ An enlarged cervical lymph node may be the only sign of thyroid carcinoma. In these patients, multiple nodal metastases are usually found at surgery.⁹⁰

The prognostic importance of regional lymph node metastases is controversial. Some studies show that the presence of regional lymph node metastases has no effect on recurrence or survival.^52–54 Other studies find that nodal metastases are a risk factor for local tumor recurrence and cancer-specific mortality and that nodal metastases correlate with distant metastases, especially if bilateral cervical or mediastinal lymph node metastases are present or if the tumor invades through the lymph node capsule.^13,47,91 In one study, 15% of patients with cervical node metastases died of thyroid carcinoma (P < .02), whereas all patients without cervical node metastases survived.⁹² Another study of patients with distant metastases from papillary carcinoma reported that 80% had mediastinal node metastases when cancer was diagnosed.⁹³ Still another study found that patients with papillary or follicular carcinoma who had cervical or mediastinal lymph node metastases had a significantly (P < .01) higher 30-year cancer-specific mortality (10%) than those without metastases (6%).¹³

Distant Metastases: Distant metastases are the principal cause of death from papillary and follicular carcinomas. Almost 10% of patients with papillary carcinoma and up to 25% of those with follicular carcinoma develop distant metastases. Approximately 50% of these metastases are present at diagnosis.⁴⁸ Distant metastases occur even more often in patients with Hürthle cell cancer (35%) and those diagnosed after age 40 years.^82,84 The sites of reported distant metastases among 1231 patients in 13 studies were lung (49%), bone (25%), both lung and bone (15%), and the central nervous system (CNS) or other soft tissues (10%). The main predictors of outcome for patients with distant metastases are patient age and the tumor's metastatic site, ability to concentrate ¹³¹I, and morphology on chest radiograph.^82,84,94,95

Although some patients, especially younger ones, with distant metastases survive for decades, approximately 50% die within 5 years regardless of tumor histology.⁴⁸ Even so, some pulmonary metastases are compatible with long-term survival. For example, one study found that when distant metastases were confined to the lung, more than 50% of the patients were alive and disease-free at 10 years, whereas no patients with skeletal metastases survived that long.⁹⁶ The survival rates are highest in young patients with diffuse lung metastases seen only on ¹³¹I imaging and not on radiograph,^95,96 which seems to be the most important feature governing an improved survival rate and prolonged disease-free interval with lung metastases.⁹⁷ Prognosis is worse with large pulmonary metastases that do not concentrate ¹³¹I and is intermediate with small nodular metastases that are seen on radiographs but that do concentrate ¹³¹I.^82,84,94

Tumor Staging and Prognostic Scoring Strategies

Several staging and clinical prognostic scoring strategies use patient age older than 40 years as a major feature to identify cancer mortality risk from differentiated thyroid carcinoma.^46,52,73,98 When applied to the papillary carcinoma data from the Mayo Clinic, 4 of the schemes that use age—EORTC, TNM 5th edition (tumor, node, metastasis), AMES (age, metastases, extent, and size), and AGES (age, tumor grade, extent, and size)—were effective in separating patients at low-risk (in whom the 20-year, cancer-specific mortality was 1%) from those at high-risk (in whom the 20-year, cancer-specific mortality was 30%–40%).⁸⁷ However, in patients with incrementally worsening MACIS (metastasis, age, completeness of resection, invasion, and size) scores of less than 6, 6 to 6.99, 7 to 7.99, and 8 or more, the 20-year survival rates decreased from 99% to 89%, 56%, and 24%, respectively.⁵² Notably, only “completeness of resection” is subject to intervention, and its contribution to prognosis is small.

Unfortunately, a study that classified 269 patients with papillary carcinoma according to 5 different prognostic paradigms found that some patients in the lowest risk group from each approach died of cancer.⁵⁵ This is particularly true of classification schemes that simply categorize patients dichotomously as low or high risk.^73,99 The AJCC TNM staging approach (see Table 1; available online, in these guidelines, at www.NCCN.org [ST-1]), which is perhaps the most widely used indicator of prognosis, classifies tumors in all patients younger than 45 years as stage I or II, even those with distant metastases. Although it predicts cancer mortality reasonably well,^100,101 TNM staging was not established to predict recurrence and therefore does not accurately forecast the recurrences that often occur in patients who develop thyroid cancer when they are young. Two studies have shown the poor predictive value of most staging approaches for thyroid carcinoma, including the TNM system.^46,102

Differentiated thyroid cancer staging systems are of value in epidemiology studies and as tools to stratify patients for prospective trials.¹⁰³ Staging systems, which are designed to segregate patients based on survival, offer gross indications of prognosis for groups of patients but are probably less useful in determining treatment for individual patients. When treating differentiated thyroid cancer, to which most patients do not succumb, many clinicians have emphasized potential morbidity more than mortality.

Systems designed to predict survival provide little guidance with respect to morbidity sustained by patients who are likely to be cured by their treatments. These NCCN Guidelines do not use TNM stages to guide therapy, but instead, many tumor and patient characteristics play important roles in the recommendations. Many specialists in thyroid cancer also follow this paradigm.

Surgical Management of Differentiated Thyroid Carcinoma

Ipsilateral Lobectomy Versus Total Thyroidectomy: The continuing debate surrounding the appropriate extent of thyroid resection reflects the limitations of prognostic scoring⁵⁴ and the morbidity often associated with total thyroidectomy performed outside of referral centers. Patients treated at the Mayo Clinic for low-risk papillary thyroid carcinomas (MACIS score ≤ 3.99) showed no improvement in survival rates after undergoing procedures more extensive than ipsilateral lobectomy; thus, Hay et al.¹⁰⁴ concluded that more aggressive surgery was indicated only for those with higher MACIS scores.However, cancer-specific mortality and recurrence rates after unilateral or bilateral lobectomy were assessed in patients with papillary carcinoma considered to be low risk according to AMES criteria.¹⁰⁵ No significant differences were found in cancer-specific mortality or distant metastasis rates between the 2 groups, but the 20-year frequencies of local recurrence and nodal metastasis after unilateral lobectomy were 14% and 19%, respectively, which were significantly higher (P = .0001) than the frequencies of 2% and 6% seen after bilateral thyroid lobe resection. Based on these observations, Hay et al.¹⁰⁵ concluded that bilateral thyroid resection is the preferable initial surgical approach for patients with AMES low-risk papillary carcinoma.

Most NCCN panel members (and other authors) advise total thyroidectomy for all patients preoperatively diagnosed with thyroid carcinoma,^3,18,106 because these procedures are associated with improved disease-free survival, even in children and adults with low-risk tumors.^{43,56,105,107} Some centers report that patients treated with lobectomy alone have a 5% to 10% recurrence rate in the opposite thyroid lobe^48,104 with an overall long-term recurrence rate of more than 30% (compared with 1% after total thyroidectomy and ¹³¹I therapy)¹³ and the highest frequency (11%) of subsequent pulmonary metastases.¹⁰⁸ Higher recurrence rates are also observed with cervical lymph node metastases and multicentric tumors, providing some additional justification for more complete initial thyroid resection.¹³

However, some prominent thyroid cancer specialists (including some at NCCN Member Institutions) oppose this view and advocate unilateral lobectomy for most patients with papillary and follicular thyroid carcinoma based on both the low mortality among patients categorized as low risk according to the AMES and other prognostic classification schemes (i.e., most patients), and the high complication rates reported with more extensive thyroidectomy.^53,98,109 The large thyroid remnant after unilateral lobectomy, however, may complicate long-term follow-up with serum thyroglobulin (Tg) determinations and whole-body ¹³¹I imaging. In most clinical settings, decisions about the extent of thyroidectomy should be individualized and made together with the patient. Circumstances in which unilateral thyroidectomy is inadvisable are detailed in the NCCN Guidelines for Thyroid Carcinoma (to view the most recent version of these guidelines, visit the NCCN Web site at www.NCCN.org).

NCCN panelists believe that total lobectomy alone is adequate treatment for papillary microcarcinomas if the patient has not been exposed to radiation, has no other risk factors, and has a tumor smaller than 1 cm that is unifocal and confined to the thyroid without vascular invasion.^13,85 The same is true for minimally invasive follicular cancers (see page 1239).

Completion Thyroidectomy: Completion thyroidectomy is recommended when remnant ablation is anticipated or if long-term follow-up with serum Tg determinations with or without whole-body ¹³¹I imaging is planned, because large thyroid remnants are difficult to ablate with ¹³¹I.¹⁰⁸ This procedure has a complication rate similar to that of total thyroidectomy. Some experts recommend completion thyroidectomy for routine treatment of tumors 1 cm or larger, because approximately 50% of patients with cancers this size have additional cancer in the contralateral thyroid lobe.^88,110–114 In patients with local or distant tumor recurrence after lobectomy, cancer is found in more than 60% of the resected contralateral lobes.¹¹¹

Miccoli et al.⁵⁶ found that 61% of irradiated children from Chernobyl who developed thyroid carcinoma treated with lobectomy had unrecognized lung or lymph node metastases that could only be identified after completion thyroidectomy. In another study, patients who underwent completion thyroidectomy within 6 months of their primary operation developed significantly fewer lymph node and hematogenous recurrences, and survived significantly longer than those whose second operation was delayed for more than 6 months.¹¹²

Surgical Complications: The most common significant complications of thyroidectomy are hypoparathyroidism and recurrent laryngeal nerve injury, which occur with much higher frequency after total thyroidectomy. Transient clinical hypoparathyroidism after surgery is common in adults¹¹⁵ and children^56,116 undergoing total thyroidectomy. However, the rates of persistent hypocalcemia are reported to be much lower in the hands of experienced thyroid surgeons. In a review of 7 published surgical series, the average rates of long-term recurrent laryngeal nerve injury and hypoparathyroidism, respectively, were 3% and 2.6% after total thyroidectomy and 1.9% and 0.2% after subtotal thyroidectomy.¹¹⁷ One study reported hypocalcemia in 5.4% of patients immediately after total thyroidectomy, persisting in only 0.5% of patients 1 year later.¹¹⁸

When experienced surgeons perform thyroidectomies, complications occur at a lower rate. A study of 5860 patients treated in the state of Maryland found that surgeons who performed more than 100 thyroidectomies per year had the lowest overall complication rate (4.3%), whereas surgeons who performed fewer than 10 thyroidectomies per year had 4 times as many complications.¹¹⁹

Radioactive Iodine

Adjuvant Radioiodine Therapy: Postoperative ¹³¹I thyroid remnant ablation is performed when a tumor has the potential to recur.¹²⁰ Studies show decreased recurrence and disease-specific mortality when postoperative ¹³¹I therapy is administered as part of the initial treatment, but the supportive data are largely confined to higher-risk populations.^{13,47,55,121,122} In a study assessing outcomes in 1004 patients with differentiated thyroid carcinoma, tumor recurrence was approximately 3-fold higher in patients either treated with thyroid hormone alone or given no postoperative medical therapy when compared with those who underwent postoperative thyroid remnant ablation with ¹³¹I (P < .001). Moreover, fewer patients developed distant metastases (P < .002) after thyroid remnant ¹³¹I ablation than after other forms of postoperative treatment; however, this effect is observed only in patients with primary tumors 1.5 cm or more in diameter.¹²¹ Some experts may find that remnant ablation has a less therapeutic effect because more extensive locoregional surgery had been performed.⁸⁷

Debate continues about ablating the thyroid bed with ¹³¹I after total thyroidectomy.^87,121 Proposed mechanisms through which remnant ablation may decrease recurrences and disease-specific mortality include the ablation of normal tissue destined to become malignant, ablation of residual microscopic malignancy in the remnant, ablation of residual microscopic malignancy outside the remnant, ablation of residual malignancy outside the remnant obscured by uptake in a large thyroid remnant, and the demonstration of unsuspected residual malignancy on the posttherapy imaging, which alters disease stage and promotes further patient management. Several other reasons favor remnant ablation: 1) patient follow-up is simplified, because elimination of “thyroid bed” uptake eliminates misinterpretation of it as disease; 2) normal tissue is eliminated as a source of Tg production, which facilitates identification of patients who are disease-free and may simplify their care while promoting early identification of those with residual cancer; and 3) elimination of normal tissue may eliminate the nidus for continued confounding anti-Tg antibody production. However, long-term evaluation of recurrence risk after adjuvant radioiodine may be confounded by the accompanying improved specificity of diagnostic testing after elimination of the thyroid remnant, and the possibility that patients who undergo adjuvant therapy may be more likely to undergo more intensive follow-up testing.

Diagnostic Whole-Body Imaging and Thyroid Stunning: Whole-body ¹³¹I imaging may be performed (category 2B) after surgery when indicated to assess the completeness of thyroidectomy and whether residual disease is present (see page 1236). However, a phenomenon termed stunning may occur when imaging doses of ¹³¹I induce follicular cell damage.¹²³ Stunning decreases uptake in the thyroid remnant or metastases, thus impairing the therapeutic efficacy of subsequent ¹³¹I.¹²⁴

The use of ¹²³I or small (2 or 3 mCi) doses of ¹³¹I and/or a shortened interval of not more than 72 hours between the diagnostic ¹³¹I dose and the therapy dose has been recommended to avoid or reduce the stunning effect; however, ¹²³I is more expensive and smaller ¹³¹I doses have reduced sensitivity when compared with larger ¹³¹I doses.^123–125 Some experts recommend that diagnostic ¹³¹I imaging be avoided completely with decisions based on the combination of tumor stage and serum Tg.¹²³ Other experts advocate that the whole-body ¹³¹I diagnostic imaging may alter therapy when, for example, unsuspected metastases are identified or an unexpectedly large remnant is identified that requires additional surgery or a reduction in radioiodine dosage to avoid substantial radiation thyroiditis.^{3,123,126,127}

Administration of Radioiodine Therapy: Historically, the 3 methods of determining ¹³¹I therapy activities (doses) have included empiric fixed doses, quantitative dosimetry, and upper bound limits that are set by blood dosimetry.^3,123,128 Recently a fourth method that adjusts the activity to deliver a selected dose to the blood (as a surrogate of the activity available for the remnant or target tissue) became available using simplified single–time point whole-body dosimetry (Kloos, personal communication).

Previously, hospitalization was required for administering therapeutic doses of ¹³¹I larger than 30 mCi (1110 MBq). However, this is no longer necessary in most states, because a change in federal regulations permits the use of much larger ¹³¹I doses in ambulatory patients.¹²⁸

Fixed ¹³¹I Doses: Administration of a fixed dose of ¹³¹I is the most widely used and simplest method. Most clinics use this method regardless of the percentage uptake of ¹³¹I in the remnant or metastatic lesion. Patients with uptake in tumor are routinely treated with large, fixed amounts of ¹³¹I. Lymph node metastases may be treated with approximately 100 to 175 mCi (3700–6475 MBq) of ¹³¹I. Cancer growing through the thyroid capsule and incompletely resected is treated with 150 to 200 mCi (5550–7400 MBq). Patients with distant metastases are usually treated with 200 mCi (7400 MBq) of ¹³¹I, which typically will not induce radiation sickness or produce serious damage to other structures but may exceed generally accepted safety limits to the blood in elder patients and those with impaired kidney function.^129,130 Diffuse pulmonary metastases that concentrate 50% or more of the diagnostic dose of ¹³¹I (which is very uncommon) are treated with 150 mCi (5550 MBq) of ¹³¹I or less to avoid lung injury, which may occur when more than 80 mCi remain in the whole body 48 hours after treatment. The administered activity of radioiodine therapy should be adjusted for pediatric patients.

Quantitative Tumor ¹³¹I Dosimetry: A second method is to use quantitative dosimetry methods to estimate the amount of radiation delivered to the lesion per unit of ¹³¹I administered. This approach is attractive, because radiation exposure from arbitrarily fixed doses of ¹³¹I can vary substantially. When the calculated dose to the tumor is less than 3500 cGy, the cancer is unlikely to respond to ¹³¹I therapy.^128,131 Radioiodine activities that deliver more than 30,000 cGy to the residual normal tissue and more than 8000 cGy to metastatic foci are likely to be effective. It is necessary to serially measure the radiation activity in the target using a tracer dose and to estimate the tumor size to make these calculations, which is difficult to do and is impossible in the setting of diffuse or microscopic lung metastases.

Blood ¹³¹I Dosimetry: A third method is to administer a dose calculated to deliver a maximum of 200 cGy to the blood, while keeping the whole-body retention less than 120 mCi (4440 MBq) at 48 hours or less than 80 mCi (2960 MBq) when diffuse pulmonary uptake is present.¹³² Thyroid cancer dosimetry and radioiodine therapy with doses more than 200 mCi are best performed in medical centers with experience using these treatments.

Posttreatment ¹³¹I Imaging: When ¹³¹I therapy is given, whole-body radioiodine imaging should be performed several days later to document ¹³¹I uptake by the tumor, primarily because up to 25% of this imaging shows lesions that may be clinically important that were not detected on the diagnostic imaging.¹²⁸ A study of pre- and posttreatment imaging showed that these differed in 27% of the treatment cycles, but only 10% of the posttreatment imaging showed clinically significant new foci of metastatic disease.¹³³ Posttreatment imaging was most likely to reveal clinically important new information in patients younger than 45 years who had previously undergone ¹³¹I therapy. Conversely, in older patients and patients who had not previously undergone ¹³¹I therapy, posttreatment imaging rarely yielded new information that might have altered the patient's prognosis.¹³³ Thus, the panel only gives a category 2B recommendation for posttreatment radioiodine imaging.

Assessment and Management After Initial Treatment

Serum Tg determinations, neck ultrasound, and whole-body ¹³¹I imaging detect recurrent or residual disease in most patients who have undergone total thyroid ablation.¹³⁴ In contrast, neither serum Tg or whole-body radioiodine imaging is specific for thyroid cancer in patients who have not undergone thyroidectomy and remnant ablation. When initial ablative therapy has been completed, serum Tg should be measured periodically. Serum Tg can be measured while the patient is taking thyroxine, but the test is more sensitive when thyroxine has been stopped or when recombinant human TSH (rhTSH) is given to increase the serum TSH.^135,136 Using current Tg assays, patients with measurable serum Tg levels during TSH suppression and those with stimulated Tg levels more than 2 ng/mL are likely to have residual/recurrent disease that may be localized promptly in almost 50% of patients and over the next 3 to 5 years in an additional 30%.¹³⁷ Approximately 6% of patients with detectable serum Tg levels, which are less than 2 ng/mL after stimulation, experience recurrences over the next 3 to 5 years, whereas this is true for approximately 2% of patients with completely undetectable serum Tg after stimulation. Conversely, the long-term clinical significance is uncertain for disease only detected through minimally elevated Tg levels after stimulation.

rhTSH: During follow-up, periodic withdrawal of thyroid hormone therapy traditionally has been required to increase the serum TSH concentrations sufficiently to stimulate thyroid tissue so that serum Tg measurements (with or without ¹³¹I imaging) could be performed to detect residual thyroid tissue or carcinoma. An alternative to thyroid hormone withdrawal is the administration of rhTSH intramuscularly, which stimulates thyroidal ¹³¹I uptake and Tg release while the patient continues thyroid hormone suppressive therapy and avoids symptomatic hypothyroidism.¹³⁸

A second multicenter international study was performed to assess the effects of 2 rhTSH dosing schedules on whole-body ¹³¹I imaging and serum Tg levels when compared with imaging and Tg levels obtained after thyroid hormone withdrawal. The imaging method in this study was more carefully standardized and accounted for the fact that ¹³¹I retention was higher in patients rendered hypothyroid than in patients given rhTSH.¹³⁶ Imaging was concordant in 89% of the patients and superior in 4% after rhTSH, and superior in 8% of patients after thyroid hormone withdrawal, but these differences were not statistically significant. The main finding in this study was that the combination of rhTSH-stimulated whole-body imaging and serum Tg measurements detected 100% of metastatic carcinoma.¹³⁶ In this study, 0.9 mg of rhTSH was given intramuscularly every day for 2 days, followed by a minimum of 4 mCi of ¹³¹I on the third day. Whole-body imaging and Tg measurements were performed on the fifth day. Whole-body ¹³¹I images were acquired after 30 minutes of imaging or after obtaining 140,000 counts, whichever came first. A serum Tg of 2.0 ng/mL or higher obtained 72 hours after the last rhTSH injection indicates that thyroid tissue or thyroid carcinoma is present, regardless of the whole-body imaging findings.^136,139

rhTSH is well tolerated. Nausea (10.5%) and transient mild headache (7.3%) are its main adverse effects.¹³⁶ It is associated with significantly fewer symptoms and dysphoric mood states than hypothyroidism induced by thyroid hormone withdrawal.¹³⁸

Measuring Serum Tg: Serum Tg measurement is the best means of detecting thyroid tissue. Tg should be measured when TSH has been stimulated either by thyroid hormone withdrawal or by rhTSH, when serum Tg has a lower false-negative rate than whole-body ¹³¹I imaging.^{135–137,140} Serum Tg levels vary in response to the increase in serum TSH after thyroid hormone withdrawal or rhTSH stimulation. Serum Tg generally does not rise as high after rhTSH administration as after withdrawal of thyroid hormone. The conditions for rhTSH-stimulated whole-body ¹³¹I imaging stipulate using 4-mCi ¹³¹I doses (based on the doses used in the pivotal phase III trial)¹³⁶ and an imaging time of 30 minutes or until 140,000 counts are obtained.

The sensitivity and specificity of various Tg assays, however, vary widely in different laboratories, even with the use of an international standard (CRM 457).^141,142 Therefore, experts recommend that patients undergo Tg monitoring using the same Tg assay performed in the same laboratory. Ideally, serum is frozen and saved for future analyses if needed, especially if a change in Tg assay is necessary.

Anti-Tg antibodies should be measured in the serum sample taken for Tg assay because these antibodies (which are found in up to 25% of patients with thyroid carcinoma) invalidate serum Tg measurements in most assays.¹⁴³ These antibodies typically falsely lower the Tg value in immunochemiluminometric (ICMA) and immunoradiometric (IRMA) assays, while raising the value in older radioimmunoassay. Although the clinical importance of these antibodies is unclear, their persistence for more than 1 year or so after thyroidectomy and radioiodine ablation probably indicates the presence of residual thyroid tissue and possibly an increased risk for recurrence.¹⁴³ In one study, 49% of patients with undetectable serum Tg concentrations and serum anti-Tg antibody concentrations of 100 U/mL or more experienced a recurrence, compared with only 3% of patients with undetectable serum Tg concentrations and serum anti-Tg antibody concentrations of less than 100 U/mL.¹⁴⁴ In patients with coexistent autoimmune thyroid disease at surgery, anti-Tg antibodies may persist much longer. In a study of 116 patients with anti-Tg antibodies before thyroidectomy, antibodies remained detectable for up to 20 years in some patients without detectable thyroid tissue, and the median time to disappearance of antibodies was 3 years.¹⁴⁵

Heterophile antibodies may falsely increase or decrease serum Tg measurements in the absence of anti-Tg antibodies. Clues to the presence of a false Tg elevation are the lack of Tg rise with TSH stimulation and the lack of linear results with serum sample dilution. Heterophile blocking tubes may be used to correct this problem.

RNA-based detection strategies, including the sodium–iodine symporter, TSH receptor, and Tg mRNAs, or DNA-based strategies to detect thyroid oncogenes in peripheral blood, represent current areas of active research that may improve the detection of residual cancer and the monitoring of these patients, especially during thyroxine treatment or when circulating anti-Tg antibodies are present.^146–149

Treating Tg-Positive/Image-Negative Patients: Posttreatment ¹³¹I imaging may indicate the location of metastases when the serum Tg level is increased, but a tumor (or metastases) cannot be found by physical examination or other localizing techniques (i.e., diagnostic ¹³¹I imaging, neck ultrasonography, CT, MRI, or PET). Pulmonary metastases may be found only after administering therapeutic doses of ¹³¹I and obtaining whole-body imaging within a few days of treatment.¹⁵⁰ In a study of 283 patients treated with 100 mCi (3700 MBq) of ¹³¹I, 6.4% had lung and bone metastases detected after treatment that had been suspected based on high serum Tg concentrations alone but had not been detected after 2-mCi (74 MBq) diagnostic imaging.¹⁵¹ In another study in 17 patients with increased serum Tg concentrations and negative 5-mCi (185 MBq) diagnostic imaging, 16 patients showed ¹³¹I uptake after 75 to 140 mCi (2775–5180 MBq) of ¹³¹I; more than 50% of these patients had lung metastases.¹⁵²

Unfortunately, most diagnostic imaging–negative/Tg-positive patients are not rendered disease-free from ¹³¹I therapy; however, the tumor burden may be diminished.¹⁵³ Thus, most patients with residual or recurrent disease confined to the neck undergo reoperation rather than radioiodine therapy with the hope of a higher chance for cure. Radioiodine therapy is more commonly considered for those with distant metastases or inoperable local disease. Patients not benefiting from this therapy can be considered for clinical trials, especially those patients with progressive metastatic disease. When a large tumor is not visible on diagnostic whole-body imaging, its ability to concentrate ¹³¹I is very low; thus, the tumor will not respond to ¹³¹I therapy.

Currently, the Tg level recommended for empiric treatment is approximately 10 ng/mL.^150,154 However, no study has shown a decrease in morbidity or mortality in patients treated with ¹³¹I based on increased Tg measurements alone. In a long-term follow-up study, no survival advantage was associated with empiric high-dose radioiodine in patients who had a negative image.¹⁵⁵ Furthermore, potential long-term side effects (i.e., xerostomia, nasolacrimal duct stenosis, bone marrow and gonadal compromise, risk for hematologic and other malignancies) may negate any benefit.^120,156

Thyroid Hormone Suppression of TSH: Because TSH is a trophic hormone that can stimulate the growth of cells derived from thyroid follicular epithelium, the use of levothyroxine to maintain low TSH levels is considered optimal for treating patients with papillary, follicular, or Hürthle cell carcinoma.^3,157 However, data are lacking to permit precise specification of the appropriate serum levels of TSH. In general, patients with known residual carcinoma or those at high risk for recurrence should have their TSH levels maintained near the lower limit of the reference range (either slightly below or slightly above). Patients who remain disease-free for several years can probably have their TSH levels maintained within the reference range. The risk and benefit of TSH-suppressive therapy must be balanced for each individual patient because of the potential toxicities associated with TSH-suppressive doses of levothyroxine, including cardiac tachyarrhythmias (especially in the elderly), bone demineralization (particularly in postmenopausal women), and frank symptoms of thyrotoxicosis.³ Patients whose TSH levels are chronically suppressed should be counseled to ensure an adequate daily intake of calcium (1200 mg/d) and vitamin D (1000 units/d).

Decreased recurrence and cancer-specific mortality rates for differentiated thyroid carcinoma are reported by some authors for patients treated with thyroid hormone suppressive therapy.^{13,121,157–161} The average dosage needed to attain serum TSH levels in the euthyroid range is higher in patients with thyroid carcinoma (2.11 mcg/kg per day) than in those with spontaneously occurring primary hypothyroidism (1.62 mcg/kg per day),¹⁶⁰ and still higher doses are required to suppress serum TSH in patients with thyroid carcinoma. Still, the optimal TSH level to be achieved in patients with thyroid carcinoma is uncertain. Superior outcomes were associated with aggressive thyroid hormone suppression therapy in high-risk patients but were achieved with modest suppression in patients with stage II disease.¹⁶¹ Excessive TSH suppression (into the undetectable, thyrotoxic range) is not required to prevent disease progression in all patients with differentiated thyroid cancer.

Adjuvant External-Beam RT: No prospective controlled trials have been completed using adjuvant external-beam RT.¹⁶² One retrospective study showed a benefit of adjuvant external-beam RT after radioactive iodine in patients older than 40 years with invasive papillary thyroid cancer (T4) and lymph node involvement (N1).¹⁶³ Local recurrence and locoregional and distant failure were significantly improved. A second study showed improved cause-specific survival and local relapse-free rate in selected patients treated with adjuvant external-beam RT (in addition to total thyroidectomy and TSH-suppressive therapy with thyroid hormone) for papillary thyroid carcinoma with microscopic residuum. Not all patients underwent radioactive iodine therapy.⁴⁷ Benefit was not shown in patients with follicular thyroid carcinoma or other subgroups of papillary thyroid carcinoma. Similarly, patients with microscopic residual papillary carcinoma after surgery are more commonly rendered disease-free after undergoing external RT (90%) than when they do not (26%).¹⁶⁴ In another study, patients with microscopically invasive follicular carcinoma were also more often disease-free after surgery when postoperative external RT was given (53%) than when it was not (38%)¹⁶⁴; however, these patients had not received radioactive iodine. Similar benefit was shown among patients treated with radioactive iodine alone and those treated with it after surgery.¹⁶⁴

Chemotherapy, External-Beam Radiation, and Surgical Excision of Metastases: Isolated skeletal metastases should be considered for surgical excision or external irradiation. Brain metastases pose a special problem, because ¹³¹I therapy may induce cerebral edema. Neurosurgical resection can be considered for brain metastases. For solitary lesions, either neurosurgical resection or stereotactic radiosurgery is preferred. Once brain metastases are diagnosed, disease-specific mortality is very high (67%), with one retrospective study showing a reported median survival of 12.4 months. Survival was significantly improved after surgical resection of one or more tumor foci.¹⁶⁵

Life-threatening tumors refractory to all other forms of therapy may be palliatively treated with doxorubicin, although the response rate is poor.⁴⁸ The experience with chemotherapy in patients with differentiated thyroid carcinoma is limited, because most recurrent tumors respond well to surgery, ¹³¹I therapy, or external-beam RT. Chemotherapy is mainly used for tumors that are not surgically resectable, are not responsive to ¹³¹I, and have either been treated with or are not amenable to therapy with external-beam RT. Among 49 patients with metastatic differentiated thyroid carcinoma who were treated with 5 chemotherapy protocols, only 2 (3%) patients experienced objective responses.¹⁶⁶ In a review of published series, 38% of patients experienced a response (defined as a decrease in tumor mass) to doxorubicin.¹⁶⁷ Combination chemotherapy is not clearly superior to doxorubicin therapy alone.⁴⁸ Overall, traditional cytotoxic systemic chemotherapy (e.g., doxorubicin) has minimal efficacy in patients with metastatic differentiated thyroid disease.

Several phase II trials are evaluating novel treatments for patients with metastatic differentiated thyroid carcinoma. Although 12-month progression-free survival was only 3% in one study assessing celecoxib (400 mg twice daily) in patients with progressive, radioiodine-unresponsive disease,¹⁶⁸ 38% of the patients had stable disease, representing a possible alteration in their disease course. Other agents are now in clinical trials, including 1) multitargeted kinase inhibitors, such as motesanib diphosphate (AMG-706),^169,170 sorafenib,^171–173 sunitinib,^174,175 axitinib,¹⁷⁶ and vandetanib; 2) the histone deacetylase inhibitors, vorinostat and depsipeptide; 3) the DNA methylation inhibitor, decitabine; 4) the heat-shock protein 90 (HSP-90) inhibitor, 17allylamino-17-demethoxygeldanamycin (17-AAG); 5) the proteasome inhibitor, bortezomib; and 6) a derivative of thalidomide, lenalidomide.^177,178 Recent reviews of the completed phase II clinical trials suggest that tyrosine kinase inhibitors have a clinical benefit (partial response rates plus stable disease) in 50% to 60% of subjects, usually for 12 to 24 months.^179,180

Surgical Therapy

A CT/MRI should be performed if the lesion is fixed or substernal (iodinated contrast should be avoided unless essential). A thyroid ultrasound (including lateral neck) is recommended if not previously performed. In one report, cervical ultrasound performed before primary surgery for newly diagnosed disease identified metastatic sites not appreciated on physical examination in 20% of patients, and surgical strategy was altered in many patients.¹⁸¹ Vocal cord mobility should also be evaluated. A chest radiograph can be considered.

The panel members agreed on the characteristics of patients who require total thyroidectomy and neck dissection (if lymph nodes are palpable or biopsy positive) as the primary treatment (see page 1233). If the nodes are negative, prophylactic central neck dissection can be considered (category 2B) but is not required in all cases.^182–184

Panel members did not agree about the preferred primary surgery for patients who are assumed to be at lower risk for cancer-specific mortality. Most panel members opted for total thyroidectomy (category 2B) in any patient in whom papillary thyroid carcinoma was identified preoperatively or during surgery. However, a few panel members felt strongly that, initially, lobectomy plus isthmusectomy (category 2B) is adequate surgery for patients at lower risk. A study in more than 5000 patients found that survival of patients after partial thyroidectomy was similar to survival after total thyroidectomy for both low- and high-risk patients.¹⁸⁵ However, another study in 2784 patients with differentiated thyroid cancer (86% with papillary thyroid cancer) found that total thyroidectomy was associated with increased survival in high-risk patients.¹⁶¹ A more recent study in 52,173 patients found that, compared with lobectomy, total thyroidectomy improves survival in patients with papillary thyroid carcinoma greater than 1 cm.¹⁸⁶

For patients who undergo lobectomy plus isthmusectomy (lower-risk patients), completion of thyroidectomy is warranted for aggressive variant disease, macroscopic multifocal disease, positive isthmus margins, cervical lymph node metastases, or gross extrathyroidal extension. Aggressive variant disease is defined as tall cell variant, columnar cell, or poorly differentiated features.

The panel agreed that completion of thyroidectomy is appropriate for any large tumor (> 4 cm), positive margins, gross extrathyroidal extension, macroscopic multifocal disease, or confirmed nodal metastases. Incidentally discovered papillary carcinomas measuring 1 to 4 cm may warrant a completion thyroidectomy (category 2B) for an aggressive variant (see page 1234); observation is another option for these patients (i.e., with Tg measurement plus anti-Tg antibodies). The TSH levels of these patients should be suppressed with levothyroxine therapy (see page 1232). Lobectomy is sufficient for tumors resected with negative margins, no contralateral lesion, no suspicious lymph node, or small (< 1 cm) papillary carcinomas found incidentally on the final pathology sections during thyroid surgery for benign disease; these patients are observed (i.e., with Tg measurement plus anti-Tg antibodies). Levothyroxine therapy to reduce serum TSH to low or low-normal concentrations is recommended for these patients.

Radioactive Iodine

Postoperative Whole-Body ¹³¹I Diagnostic Imaging: Performing diagnostic whole-body ¹³¹I imaging with adequate TSH stimulation (thyroid withdrawal or rhTSH stimulation) before ¹³¹I therapy is a category 2B recommendation (see page 1236). The panel advises that this decision be weighed against the problem of stunning that occurs with diagnostic ¹³¹I imaging.¹⁸⁷ Diagnostic whole-body ¹²³I imaging does not carry a risk for stunning. The alternatives to performing diagnostic ¹³¹I imaging are to obtain an ¹²³I image before ¹³¹I therapy, obtain a thyroid uptake measurement with microcurie quantities of radioiodine to confirm neck uptake, or forgo diagnostic imaging. If radioiodine is administered after a diagnostic ¹³¹I study, the interval between radioiodine doses should be minimized. Whenever therapeutic radioiodine is administered, whole-body imaging should be obtained 5 to 8 days after treatment with ¹³¹I, which is termed posttreatment ¹³¹I imaging in the guidelines.

Thyroid Remnant Ablation With Radioactive Iodine: The decision to ablate uptake in the thyroid bed is closely linked to the extent of thyroid surgery and is not recommended for patients who have undergone lobectomy or lobectomy plus isthmusectomy as initial surgery. Adjuvant radioiodine ablation (30–100 mCi) to destroy residual thyroid function is recommended for suspected (based on pathology, postoperative Tg, and intraoperative findings) or proven thyroid bed uptake in patients who have undergone total thyroidectomy and have no gross residual disease in the neck (see page 1236). The administered activity of radioiodine therapy should be adjusted for pediatric patients. The panel does not routinely recommend empiric administration of radioiodine without diagnostic imaging.

Radioactive Iodine Treatment: Therapy with ¹³¹I is advised for patients with tumors found on examination, by imaging studies, or through increased serum Tg levels if the tumors are not amenable to surgical removal and if they concentrate ¹³¹I. All patients should be examined, and palpable neck disease should be surgically resected before any radioiodine treatment. A negative pregnancy test is required before the administration of radioiodine in women of childbearing potential. The panel agrees that radioiodine treatment is not needed for patients with Tg levels less than 1 ng/mL, negative radioiodine imaging, and negative anti-Tg antibodies. For patients with suspected or proven radioiodine responsive residual tumor, radioiodine treatment can be given at 100 to 200 mCi along with posttreatment imaging; dosimetry can be considered for distant metastases (see page 1236). Again, the administered activity of radioiodine therapy should be adjusted for pediatric patients.

For unresectable locoregional recurrence, radioiodine treatment with RT can be given if the radioiodine imaging is positive; RT alone is another option in the absence of radioiodine uptake. When recurrent disease is suspected based on high stimulated Tg values (> 10 ng/mL) and negative imaging studies (including PET scans), radioiodine therapy can be considered (category 3) using an empiric fixed dose of 100 to 150 mCi of ¹³¹I (see page 1237); however, there was major disagreement about this recommendation. For patients with metastatic disease that is not locoregional, the panel recommends individualized treatment based on the tumor locations (e.g., CNS, bone, or sites other than CNS; see page 1238).

Adjuvant External-Beam RT

The guidelines recommend that external RT be considered for patients older than 45 years with T4 (surgically evident gross extrathyroidal extension) and without gross residual disease in their neck (see page 1236).

Thyroxine Suppression of TSH

Thyroxine therapy is required after total thyroid resection, and is advisable even after lobectomy and isthmusectomy. The level of TSH suppression is not stipulated, because data currently conflict. Practically, the most appropriate dose of thyroid hormone for most low-risk patients with differentiated thyroid cancer is one that decreases the serum TSH concentration to just below the lower limit of the normal range (see page 1232).³ At a minimum, patients should not be permitted to have increased TSH levels, because this would represent inadequate treatment of both postsurgical hypothyroidism and differentiated thyroid carcinoma. A greater degree of TSH suppression is generally recommended for higher-risk patients, including those with metastatic disease.³ The risk and benefit of TSH-suppressive therapy must be balanced for each patient. Patients whose TSH levels are chronically suppressed should be counseled to ensure an adequate daily intake of calcium (1200 mg/d) and vitamin D (1000 units/d).

Surveillance and Maintenance

The guidelines recommend the following for surveillance and maintenance (see page 1237): 1) physical examination, TSH, Tg, and anti-Tg antibody measurements every 6 to 12 months, then annually if patients remain disease-free; 2) periodic neck ultrasound; 3) TSH-stimulated Tg (without radioiodine imaging) in patients previously treated with radioiodine with negative TSH-suppressed Tg and negative anti-Tg antibodies; 4) consider TSH-stimulated radioiodine imaging in patients with T3–4 or M1 at initial staging, or with abnormal Tg levels (either TSH-suppressed or TSH-stimulated), abnormal anti-Tg antibodies, or abnormal ultrasound during surveillance; 5) whole-body ¹³¹I imaging every 12 months until no response is seen to radioiodine treatment in iodine responsive tumors (either withdrawal of thyroid hormone or rhTSH) for patients with detectable Tg, distant metastases, or soft tissue invasion on initial staging; and 6) consider additional nonradioiodine imaging (e.g., FDG-PET with or without CT) if Tg levels are ≥ 10 ng/mL for patients whose ¹³¹I imaging is negative and stimulated Tg is more than 2 to 5 ng/mL.

The panel acknowledges that the suggested Tg cutoff levels will continue to evolve as new Tg assays are introduced. In selected patients who may be at higher risk for residual or recurrent disease (e.g., those with N1 disease), a stimulated Tg should be obtained and concomitant diagnostic radioiodine (RAI) imaging considered. In patients with a positive stimulated Tg, concomitant RAI imaging may help determine whether treatment with RAI is indicated (i.e., RAI is often beneficial in iodine-avid disease but not in non-iodine avid disease; see page 1237). A subgroup of very–low-risk patients (micropapillary carcinomas entirely confined to the thyroid gland) may only require periodic ultrasound follow-up, without stimulated Tg or follow-up whole-body imaging, as long as the basal Tg remains low (see page 1237).

Recurrent and Metastatic Disease

The panel agrees that the preferred therapy for recurrent disease is surgery if the tumor can be localized and is resectable (see page 1237). Preoperative vocal cord assessment should be considered for those with central neck recurrence. For unresectable locoregional recurrences, ¹³¹I therapy is recommended for tumors that concentrate ¹³¹I (i.e., radioiodine imaging–positive), and external-beam RT alone is recommended for those that do not (i.e., radioiodine imaging–negative). Unresectable iodine-responsive locoregional disease that is unlikely to respond to radioiodine therapy alone may also be treated with external-beam RT.

For metastatic disease, several therapeutic approaches are recommended (see page 1238), depending on the site and number of tumor foci. Patients should continue to receive levothyroxine to suppress TSH levels. For skeletal metastases, surgical palliation is recommended for symptomatic or asymptomatic tumors in weight-bearing extremities; other therapeutic options are ¹³¹I treatment (if the radioiodine imaging is positive) with consideration of dosimetry to maximize dosing and/or external-beam RT. Intravenous bisphosphonate (pamidronate or zoledronic acid) therapy may be considered for symptomatic bone metastases, and embolization of metastases can also be considered.¹⁸⁸ For metastases to the CNS, neurosurgical resection should be considered for appropriate cases, and/or radioiodine treatment (with rhTSH and steroid prophylaxis) if the radioiodine imaging is positive (with consideration of dosimetry to maximize dosing), and/or image-guided RT (see page 1238). For solitary CNS lesions, either neurosurgical resection or stereotactic radiosurgery is preferred.

For sites other than the CNS, surgical resection and/or RT of selected, enlarging, or symptomatic metastases can be considered, and/or ¹³¹I is recommended if the tumor concentrates the radioisotope (with consideration of dosimetry to maximize the dosing). For clinically progressive or symptomatic disease, patients should consider 1) clinical trials; 2) small molecule kinase inhibitors (i.e., sorafenib or sunitinib) or systemic therapy if a clinical trial is not available; or 3) best supportive care.¹⁸⁹ Because chemotherapy has been generally disappointing, the guidelines recommend clinical trials for non–radioiodine avid tumors; sorafenib, sunitinib, or traditional cytotoxic systemic therapy can be considered if a clinical trial is not available.^3,171,173 Several agents are currently being investigated in clinical trials (http://www.thyroidtrials.org, http://www.nci.nih.gov/clinicaltrials).

Treatment and Prognosis

No effective therapy exists for anaplastic carcinoma, and the disease is almost uniformly fatal. The median survival from diagnosis ranges from 3 to 7 months.¹⁹⁸ The 1- and 5-year survival rates are approximately 17% and 8%, respectively.^196,198 Death is attributable to upper airway obstruction and suffocation (often despite tracheostomy) in 50% of these patients, and to a combination of complications of local and distant disease and/or therapy in the remaining patients. Patients with disease confined to the neck at diagnosis have a mean survival of 8 months compared with 3 months if the disease extends beyond the neck.¹⁹⁹ Other variables that may predict a worse prognosis include older age at diagnosis, male sex, and dyspnea as a presenting symptom.

Except for patients whose tumors are small and confined entirely to the thyroid or readily excised structures, total thyroidectomy with attempted complete tumor resection has not been shown to prolong survival.^198–201 External-beam RT can increase short-term survival in some patients; RT can also improve local control and can also be used for palliation (e.g., to prevent asphyxiation).^193,202 Treatment with single-drug chemotherapy also does not improve survival or disease control in the neck, although perhaps 20% of patients experience some response in distant metastases. The introduction of hyperfractionated RT, combined with radiosensitizing doses of doxorubicin, may increase the local response rate to approximately 80%, with subsequent median survival of 1 year. Distant metastases then become the leading cause of death.²⁰³ Similar improvement in local disease control has been reported with a combination of hyperfractionated RT and doxorubicin, followed by debulking surgery in patients with responsive disease.²⁰⁴ However, the addition of larger doses of other chemotherapeutic drugs has not been associated with improved control of distant disease or survival. Paclitaxel has been tested in newly diagnosed patients and may provide some palliative benefit.^205,206

Once the diagnosis of anaplastic carcinoma is confirmed pathologically, the panel recognizes the importance of rapidly determining the potential for local resection.¹⁹³ If the disease is deemed likely to be resectable, an attempt at total thyroidectomy should be made, with selective resection of all involved local or regional structures and nodes. The patency of the airway should be considered throughout the patient's course. Given the poor outcome with current standard therapy, all patients, regardless of surgical resection, should be considered for clinical trials. Currently, ongoing clinical trials include combretastatin A4 phosphate (CA4P; a vascular disrupting agent), CS-7107 (an oral PPAR gamma [peroxisome proliferator–activated receptors] agonist), and novel multitargeted therapies, including bevacizumab with doxorubicin, sorafenib, sunitinib, and imatinib (http://clinicaltrials.gov/ct2/results?term=thyroid+cancer).^178,207,208 A patient with anaplastic thyroid cancer had a durable complete response in a phase I trial with CA4P, and has been disease-free for more than 3 years.^209,210 Recent data using fosbretabulin, which is another vascular disrupting agent, in 26 patients with advanced anaplastic thyroid cancer showed that 33% of patients survived more than 6 months.²¹¹

Multimodality therapy should also be considered. Although optimal results have been reported with hyperfractionated RT combined with chemotherapy, the panel acknowledged that considerable toxicity is associated with this treatment and that prolonged remission is uncommonly reported.²¹² A recent study found that surgery and RT were associated with improved survival, but not chemotherapy.²¹³ The guidelines do not recommend particular chemotherapeutic agents, either for radiosensitization or full-dose therapy, because of a lack of clear evidence of efficacy for any particular regimen.