From a radiation oncologic standpoint, follicular cell–derived cancer (FCDC) of the thyroid has some historic interest. In 1940, Hamilton and associates ¹⁵ at the University of California, San Francisco, first reported evidence of uptake of radioactive iodine (RAI) in FCDC. In 1942 at Columbia University, Keston and colleagues¹⁶ gave a patient with an iodophilic femoral metastasis a 10-mCi therapeutic dose of RAI. Since that time, RAI therapy has become the primary treatment for iodophilic distant metastases in the papillary or follicular histologic type of FCDC.¹⁷ The concept of radioiodine remnant ablation (RRA) to “complete” a thyroidectomy in FCDC, derived from pioneering work by Blahd and colleagues¹⁸ at the University of California, Los Angeles, in the late 1950s, has been increasingly employed in the routine management of FCDC.¹⁹ The use of external beam radiotherapy (EBRT) in managing thyroid cancer has its origin in the 1960s.²⁰ Subsequently, the role of EBRT generally has been restricted to the treatment of locoregional advanced FCDC, but it has also been employed as primary therapy in the management of two rare malignancies: undifferentiated (anaplastic) thyroid carcinoma and primary malignant lymphoma of the thyroid.²¹ In this chapter, we discuss the more common FCDCs, but where indicated, less common tumors derived from nonfollicular cell origin are considered.

Etiology and Epidemiology

The annual incidence of thyroid cancer is 0.5 to 10 per 100,000 people in the general population in most countries; a global estimate suggested 87,000 new cases worldwide each year.²² Clinical thyroid malignancy is relatively uncommon, accounting for only about 2% of human malignancies.^2,²³ Nonetheless, it was estimated that thyroid cancers would account for 95% of endocrine malignancies in the United States in 2010.² It was estimated by the American Cancer Society that 1690 patients with thyroid cancer would die during 2010, accounting for 66% of deaths from endocrine malignancies.²

Most thyroid cancers are derived from follicular epithelium. In most countries, incidence rates for papillary thyroid carcinoma (PTC) generally exceed the incidence rates for follicular thyroid carcinoma (FTC), and either PTC or FTC is far more common than the usually lethal anaplastic (undifferentiated) thyroid carcinoma.^24,²⁵ Very similar data on the frequency of the various histologic types are contained in three large series of more than 97,000 cases of thyroid cancer reported from the United States and Japan.²⁶^–²⁸ These data are summarized in Table 36-1.

TABLE 36-1 Frequency of Various Histologic Types of Thyroid Cancers Diagnosed during 1973-1995

Anaplastic carcinomas and FTCs tend to be relatively more common in areas endemic for goiter, and numerous case-control studies have strongly suggested that dietary iodine content is responsible for the increased incidence rates in these areas.²⁹ This hypothesis is supported by the fact that dietary iodine supplementation has been shown to increase the relative proportion of PTC and to decrease the frequency of FTC.³⁰

PTC and FTC are more than twice as common in women as men and tend to occur much more commonly in middle age and later, although patients with PTC are younger than patients with FTC.^24,²⁵ The preponderance of women with FCDC has led to speculation about the role of estrogens as a risk factor. Other putatively estrogen-dependent tumors, particularly breast cancer, and thyroid cancer occur more frequently in the same individual than expected by chance.³¹ Case-control studies have suggested a correlation between pregnancy, a high estrogen state, and the onset of thyroid cancer.³² Other studies have suggested that pregnancy per se, rather than the associated estrogen levels, may be associated with increased thyroid cancer risk.³³ The role of female sex hormones as a risk factor for thyroid cancer development must still be considered unresolved.

The most firmly established risk factor for development of thyroid cancer is prior exposure to ionizing radiation, particularly to the head and neck region during childhood, which used to be a common problem in some exposed populations in Japan and the Pacific during the 1960s and 1970s in the aftermath of atomic bomb use at the end of World War II and in areas exposed to atmospheric nuclear bomb tests during the 1950s and 1960s.^34,³⁵ In most other countries, radiation treatments for benign medical conditions, such as acne vulgaris, thymic enlargement, tinea capitis, or inflammatory connective tissue disorders, contributed to increasing numbers of patients with thyroid cancer.^36,³⁷ These practices have largely been abandoned, and radiation exposure as a risk factor for the most part has ceased to be of significant importance.³⁷ Exceptions are areas of high natural background radiation; radiotherapy used for malignant conditions³⁸; and areas where radioactive contamination of the environment from military or civilian sources is a notable problem, particularly in many countries in the southern part of the former Soviet Union that were heavily contaminated in the wake of the Chernobyl nuclear reactor accident.³⁹ The contamination in parts of Belarus and Ukraine was significant and prolonged, and there is mounting clinical evidence that this contamination has led to increased rates of thyroid malignancies, often of an unexpectedly aggressive nature.⁴⁰^–⁴² In the United States, it has been disclosed that a significant segment of the U.S. population was exposed to radioiodine during a series of nuclear bomb explosions at the Nevada test site in the 1950s.⁴³ Predictions of excess relative risk of thyroid cancer vary but may prove to be significant. However, it is impossible to determine whether individual thyroid cancers arose as a result of radiation exposure or as sporadic events.⁴³

Although most cases of thyroid cancer are sporadic occurrences, a small proportion of thyroid cancers may be familial, and in the case of C cell–derived malignancy, medullary thyroid carcinoma (MTC) may be associated with multiple endocrine neoplasia (MEN) type II (MEN-II) syndrome and its associated adrenal medullary tumors (pheochromocytomas). In patients with the MEN-II syndrome, a strong and typical family history can often be obtained. PTC may also be associated with other nonthyroid malignancies and with premalignant conditions such as Cowden’s syndrome and familial adenomatous polyposis coli (i.e., Gardner’s syndrome).⁴⁴ Several cases of familial PTC have been described.⁴⁵ No such distinct associations exist for FTC, but aggregation of FTC cases in families with dyshormonogenesis has been described.⁴⁶

Prevention and Early Detection

Most thyroid cancers are sporadic and not caused by an avoidable environmental agent. Prevention of thyroid malignancy usually is impossible.

Nodular thyroid disease generally can be detected early by careful neck palpation in the course of routine physical examinations. High-resolution ultrasound can readily detect impalpable thyroid nodules 2 to 3 mm in diameter. There is no evidence, however, that population screening for nodular thyroid disease using ultrasound is warranted because 70% to 90% of diagnosed thyroid malignancies are of the papillary histologic type. In such incidentally discovered papillary tumors, 85% to 90% are deemed low risk and are associated with mortality rates close to rates predicted by actuarial curves and comparable to rates for nonmelanoma skin cancer.

In contrast, early detection of the often more aggressive, familial MTC is possible when there is evidence within the family of an inherited mutation of the RET proto-oncogene in the pericentromeric region of the short arm of chromosome 10. Such mutations have been found in more than 90% of familial MTC cases, and since the successful cloning and sequencing of the RET gene, asymptomatic members of affected families can be tested for the presence of a mutation at this locus.⁴⁷ A positive test result obviates the need for any further testing. The present recommendation for persons shown to harbor the mutation is to undergo prophylactic total thyroidectomy, which completely prevents the development of the invariably multicentric MTC associated with these conditions.^11,⁴⁸ If such a mutation is found in infancy, the current practice is to perform a total thyroidectomy when the child is 3 to 5 years old.^11,⁴⁹

Pathology and Pathways of Spread

Most thyroid cancers are well-differentiated tumors derived from follicular cells. PTC accounts for 40% to 90% of cases, and 15% to 40% are classified as FTC, which includes the so-called oxyphilic or Hürthle cell variant. Anaplastic thyroid cancer (ATC), also derived from the follicular cell, is the least common FCDC, typically constituting 1% to 5% of most series. The frequency of C cell–derived MTC depends on the diligence with which reporting centers pursue the early diagnosis of patients with familial MTC and MEN-II, but it may account for 2% to 8% of thyroid cancers. Nonepithelial thyroid tumors include sarcomas, malignant hemangioendotheliomas, and malignant lymphomas. Lymphomas may involve the thyroid as the only manifestation of the disease or as part of a systemic disease; such tumors may rarely complicate Hashimoto’s (autoimmune) chronic lymphocytic thyroiditis, which in the United States is the most common cause of goiter and noniatrogenic hypothyroidism. Blood-borne metastases to the thyroid are common at autopsy in patients with widespread malignancy, but they rarely cause clinically detectable thyroid enlargement.

PTC most often occurs in patients 30 to 50 years old; the mean age at diagnosis is about 45 years. Most primary tumors are 1 to 4 cm in diameter; the average is about 2 to 3 cm in the greatest dimension. Of tumors, 95% are classified on the basis of degree of differentiation as histologic grade 1 (of 4); 80% of primary PTC tumors are assessed to be DNA diploid by flow cytometry.⁵⁰ Extrathyroidal invasion of adjacent soft tissues is present in about 15% (range 5% to 34%) at primary surgery, and about one-third of PTC patients have clinically evident lymphadenopathy at presentation.⁵¹ About 35% to 50% of excised neck nodes have histologic evidence of involvement, and in patients 17 years old or younger, nodal involvement may be present in up to 90%.⁵² The primary disease is confined to the neck in 93% to 99% of PTC patients at diagnosis.^50,⁵¹ Spread to superior mediastinal nodes is usually associated with extensive neck nodal involvement. Distant metastases are diagnosed in only 1% to 7% of patients with PTC before or within 30 days of primary treatment.^50,⁵³

FTC occurs in older patients, and the mean age in most studies is more than 50 years, about 10 years older than for typical PTC.^25,⁵⁴ Women affected by FTC outnumber men by more than 2 : 1. FTC patients rarely (4% to 6%) have clinically evident lymphadenopathy at presentation.⁵⁵ In most series, the average tumor size in FTC is larger than in PTC.⁵⁴ When tumor grading is performed, higher grade tumors are more common than with PTC.⁵⁴ DNA aneuploidy is present in about 60% of FTC tumors and in up to 90% of patients with oxyphilic or Hürthle cell variant tumors.^25,56 Direct extrathyroidal extension into adjacent soft tissues does not occur in the common “minimally invasive” FTC, but it is not unusual in the rare patient with “widely invasive” FTC. Of patients with FTC, 5% to 20% may have distant metastases at presentation, and the most common sites of distant spread are lung and bone.^25,57

MTC arises from the C cells of the thyroid rather than the follicular epithelium; secretes a characteristic hormone, calcitonin; is frequently associated with one or more paraendocrine manifestations; and provides an early biochemical signal (i.e., hypersecretion of calcitonin) that permits its early detection, treatment, and cure.11,58 The tumor occurs in sporadic and familial forms, with the latter constituting about 20% of the total. The familial variety usually appears at a younger age, is almost invariably bilateral, is less likely to have associated cervical metastases at presentation, and has a better prognosis.^11,58 Most importantly, the familial variety is preceded by a premalignant C cell hyperplasia that can be cured by total thyroidectomy.⁵⁹

ATC is the least common FCDC and typically constitutes less than 5% of most reported series.⁶⁰ It usually occurs after age 60 years, and it is only slightly more common in women than men (1.3 : 1 to 1.5 : 1). It is highly malignant, rapidly invading adjacent structures and metastasizing throughout the body. Pathologic examination of tumor biopsy specimens may reveal evidence of PTC or FTC, which may represent a precursor of ATC. Thorough biopsy sampling may be necessary to detect residual well-differentiated thyroid tissue. On histologic examination, the tumor is usually composed of atypical cells that exhibit numerous mitoses and form a variety of patterns. Spindle-shaped cells and multinucleated giant cells usually predominate, but in a third histologic pattern described as squamoid, the cells are undifferentiated but retain an epithelial appearance. It was formerly thought that there was a small cell ATC, but most of these tumors have been classified as malignant thyroid lymphomas.

Biologic Characteristics, Prognostic Factors, and Staging

It has long been recognized that prognosis in thyroid cancer largely depends on the age of the patient, the histology of the tumor, and the anatomic extent of disease at presentation. Although head and neck cancer is staged entirely on the basis of anatomic extent, thyroid cancer staging is unique in that the histologic diagnosis and the age of the patient are included because of their prognostic importance.

The seventh edition of the TNM (primary tumor, regional nodes, metastases) staging scheme for thyroid carcinoma, approved in 2009 by the International Union Against Cancer ⁶¹ and in 2010 by the American Joint Committee on Cancer (AJCC),⁶² is presented in Table 36-2. The primary tumor status is defined in this scheme on the basis of the size of the primary lesion (diameter in centimeters) and the presence of extrathyroidal extension. A T1 tumor is 2 cm or smaller and limited to the thyroid. T1a refers to tumors of 1 cm or less, and T1b refers to a tumor more than 1 cm but not more than 2 cm in greatest dimension. A T2 tumor is 2.1 to 4.0 cm in diameter and limited to the thyroid. A T3 tumor is larger than 4 cm and limited to the thyroid or any tumor with minimal extrathyroidal extension (i.e., to sternothyroid muscles or perithyroid soft tissues). A T4a tumor represents moderately advanced disease, which extends beyond the thyroid capsule and invades subcutaneous soft tissue, larynx, trachea, esophagus, or recurrent laryngeal nerve, and a T4b tumor represents very locally advanced disease, typically invading prevertebral fascia or encasing carotid artery or mediastinal vessels.

TABLE 36-2 American Joint Committee on Cancer Stage Groupings for Thyroid Carcinoma

When tumor spreads to lymph nodes, it is classified as N1. N1a refers to metastasis in level VI (i.e., pretracheal and paratracheal nodes, including prelaryngeal and delphian nodes), and N1b describes metastasis to other unilateral, bilateral, or contralateral cervical or superior mediastinal lymph nodes. M1 refers to the presence of distant metastasis involving nonregional lymph nodes, internal organs, or bones.

To make this TNM information clearer, several of these TNM descriptions can be grouped together into stages I through IV. For the seventh edition of the TNM/AJCC staging system, the grouping for MTC is the same as for PTC or FTC in patients older than 45 years with one exception. A T3 N0 M0 MTC patient is considered stage II, whereas such a patient with papillary and follicular cancers is considered stage III. All patients with ATC are considered stage IV, reflecting the poor prognosis of this type of cancer.

For patients younger than 45 years with a diagnosis of PTC or FTC, the TNM stage is I (any T, any N, M0) if there is no distant spread and stage II (any T, any N, M1) if there are distant metastases beyond the neck or upper mediastinal lymph nodes. For patients with MTC or who are 45 years or older with PTC or FTC, the seventh edition of the TNM/AJCC staging system is more complicated than before.

In patients with PTC or FTC who are 45 years of age or older, stage I (T1 N0 M0) cancers are 2 cm or smaller and have not spread to lymph nodes or distant sites. Stage II (T2 N0 M0) cancers are more than 2 cm in diameter but not bigger than 4 cm, and they are localized to the thyroid. Stage III (T3 N0 M0 or T1-3 N1a M0) encompasses tumors larger than 4 cm or with minimal extrathyroidal extension and tumors of any size that have spread to central neck nodes.

In former TNM/AJCC classifications, stage IV referred only to patients who had distant spread, but the later editions (since 2002) define these patients as IVC (any T, any N, M1), and the designations IVA and IVB define more aggressive tumors that have not spread to distant sites. Stage IVB (T4b, any N, M0) refers to the unusual situation of a tumor that has grown back to the spine (i.e., prevertebral fascia) or into nearby large blood vessels (i.e., carotid artery or mediastinal vessels). The new stage IVA (T1-4a, N0-1b, M0) was formerly considered to be stage III, and it encompasses locally invasive T4a tumors (i.e., with or without nodal involvement [any N]) and node-positive N1b tumors (with primary tumors of any size but localized or showing only minimal extrathyroidal extension [T1-3]).

Because of significant differences in the tumor biology displayed by the four principal types of thyroid carcinomas, the remainder of this discussion of tumors is divided according to histologic type.

Papillary Thyroid Carcinoma

Most patients with PTC present with localized, node-negative disease.28,50,65 Patients 45 years old or older with nodal metastases or extrathyroidal extension account for less than 20% of cases.⁵⁰ Only about 1% to 3% of older PTC patients present with distant metastases. Figure 36-1 shows cause-specific survival according to pathologic TNM (pTNM) stage in a cohort of 1851 patients who underwent surgical treatment at Mayo Clinic during the period 1940-1990.⁶³

Figure 36-1 Cause-specific survival according to pathologic TNM (primary tumor, regional nodes, metastases) stage for a cohort of 1851 patients with papillary thyroid carcinoma treated at the Mayo Clinic during the period 1940-1990. Numbers in parentheses are the percentages of patients in each pathologic TNM stage grouping.

From Larsen PR, Davies TF, Hay ID: The thyroid gland. In Wilson JD, Foster DW, Kronenberg HM, et al, editors: Williams Textbook of Endocrinology, 9th ed, Philadelphia, 1998, WB Saunders, pp 389-515.

Three types of tumor relapse may occur with PTC: postoperative regional nodal metastasis, local recurrences, and postoperative distant metastases. Local recurrence has been defined as “histologically confirmed tumor occurring in the resected thyroid bed, thyroid remnant, or other adjacent tissues of the neck (excluding lymph nodes)” after complete surgical removal of the primary tumor.⁶⁴ Nodal or distant spread is considered postoperative if the metastases are discovered within 180 days or 30 days, respectively. In a cohort of 2370 patients with PTC who did not have initial distant metastases and had complete surgical resection of their primary tumors during treatment at Mayo Clinic from 1945-2000, relapse rates at nodal, local, and distant sites were 9.8%, 5.5%, and 4.6% after 25 years of follow-up.⁶⁵ In a larger cohort of 2512 PTC patients treated during the period 1940-1990 at Mayo Clinic, the cause-specific survival rates at 5, 10, and 20 postoperative years were 98%, 96%, and 95%.⁶⁵ Of patients with lethal PTC, 20% of deaths occurred within the first year after diagnosis, and 80% of the deaths occurred within 10 postoperative years.^50,51,65

Only a fraction (≈15%) of PTC patients have relapse of disease, and even fewer (≈5%) have a lethal outcome.50–52,65 The exceptional patient who experiences an aggressive course tends to relapse early, and the rare fatalities usually occur within 5 to 10 years of initial diagnoses.^50,51,65

Multivariate analyses have been used to identify variables predictive of cause-specific mortality.66,67,68,69 Increasing patient age and presence of extrathyroidal invasion are independent prognostic factors in all such studies.^66,67,68,69 The presence of initial distant metastases and large size of the primary tumor are also significant variables in most studies,^66,68 and some groups^{26,28,50,66,67} have reported that histopathologic grade (i.e., degree of differentiation) is an independent variable. The completeness of initial tumor resection (i.e., postoperative status) is also a predictor of mortality.^50,⁶⁹ The presence of initial neck nodal metastasis, although relevant to future nodal relapse, does not apparently influence cause-specific mortality.^50,^51,⁶⁹

From a multivariate analysis of more than 14,200 patient-years of experience, a prognostic scoring system was devised and named the AGES system after five independent variables: patient’s age, tumor grade, tumor extent (e.g., local invasion, distant metastasis), and tumor size.^50,⁶⁶ With the use of such a scoring system, 86% of PTC patients were in the minimal-risk group (AGES score <4), and they had a cause-specific mortality rate of only 1%.⁵⁰ In contrast, patients with AGES scores of 4 or greater (i.e., high-risk group is 14% of the total) had a 20-year cause-specific mortality rate of 40%. Based on the description of the Mayo Clinic–derived AGES system,⁶⁶ Cady and Rossi,⁷⁰ working at the Lahey Clinic, devised a simplified version of the AGES system, which they called the AMES multifactorial system. The AMES system disregarded tumor grade because this information was not readily available to the authors, but they took advantage of the other four variables: age, metastasis, extent, and size. The details of the AMES system are presented in Table 36-3. Of the 1961-1980 cohort of patients studied by Cady and Rossi, 89% were deemed low risk and had a death rate of 1.8%—outcome results virtually identical to results defined by the AGES prognostic scores.^50,⁶⁶

TABLE 36-3 Classification of Risk Group Categories According to the AMES System

Low Risk

1 Younger patients (men <41 yr; women <51 yr), no distant metastasis

2 Older patients with (a) minor tumor capsule involvement and (b) primary cancers <5 cm in diameter and (c) no distant metastasis

High Risk

1 All patients with distant metastasis

2 All older patients with (a) major tumor capsule involvement or (b) primary cancers ≥5 cm in diameter

AMES, age, metastasis, extent, and size.

Data from Cady B, Rossi R: An expanded view of risk-group definition in differentiated thyroid carcinoma. Surgery 104:947-953, 1988.

Although the AGES scheme had the potential for universal application, some academic centers could not include the differentiation variable (tumor grade [G]) because their surgical pathologists did not recognize higher grade PTC tumors.⁷¹ Another prognostic scoring system for predicting PTC mortality rates was devised with the use of candidate variables that included completeness of primary tumor resection but excluded histologic grade.⁶⁹ Cox model analysis and stepwise variable selection led to a final prognostic model that included five variables: metastasis, age, completeness of resection, invasion, and size (MACIS). The final score was defined as MACIS = 3.1 (if age is ≤39 years) or 0.08 × age (if age is ≥40 years), +0.3 × tumor size (in centimeters), +1 (if tumor not completely resected), +1 (if locally invasive), +3 (if distant metastases are present). As illustrated by Figure 36-2, the MACIS scoring system permitted identification of patient groups with a broad range of risk of death from PTC. Twenty-year cause-specific survival rates for patients with MACIS scores of less than 6, 6 to 6.99, 7 to 7.99, and 8 or more were 99%, 84%, 56%, and 24%. When cumulative mortality from all causes of death was considered, approximately 85% of PTC patients (who had AGES scores <4 or MACIS scores <6) experienced no excess mortality over rates predicted for control subjects.^50,65,66,69

Figure 36-2 Cause-specific survival according to MACIS (metastasis, age, completeness of resection, invasion, and size) scores of <6, 6-6.99, 7-7.99, and ≥8 for a cohort of 1851 consecutive patients with papillary thyroid carcinoma (PTC) undergoing initial treatment at the Mayo Clinic during the period 1940-1990. Numbers in parentheses are the numbers and percentages of patients with PTC in each of the four risk groups.

Follicular Thyroid Carcinoma

When more than 75% of cells in an FTC exhibit Hürthle cell or oncocytic features, the tumor is classified as a Hürthle cell carcinoma (HCC), oncocytic carcinoma, or oxyphilic variant FTC.²⁵ Most patients with FTC or HCC present with tumors 2 cm or larger and confined to the neck. Patients 45 years old or older with nodal metastases or extrathyroidal extension account for only about 4% to 7% of FTCs and 8% to 10% of HCCs.^28,⁶³ In contrast to PTC patients, among whom only 1% to 3% present with distant metastases, about 4% to 6% of HCC patients and 7% to 15% of non–oxyphilic variant FTC patients have distant metastases at the time of initial diagnosis.^25,^28,⁶³ Figure 36-3 shows cause-specific survival according to pTNM stage in a cohort of 153 patients with nonoxyphilic FTC surgically treated at Mayo Clinic during the period 1940-1990.⁶³

Figure 36-3 Cause-specific survival according to pathologic TNM (primary tumor, regional nodes, metastases) stages for a cohort of 153 patients with nonoxyphilic follicular thyroid carcinoma treated at the Mayo Clinic during the period 1940-1990. Numbers in parentheses are the percentages of patients in each pathologic TNM stage grouping.

Nodal metastases are rare in typical FTC, and nodal relapse rates at 10 and 20 postoperative years are 1% and 2%. About 6% of patients with HCC have nodal involvement at presentation,⁵⁵ but at 20 and 30 years after primary surgery, 18% and 24% of these patients have nodal relapse.²⁵ When relapse at neck or distant sites is considered, patients with HCC have the highest rates of tumor relapse after 10 or 20 years. Local recurrences account for most of these postoperative events because the numbers of distant metastases in patients with HCC are comparable with the numbers of metastases found in nonoxyphilic FTC (i.e., about 20% after 20 postoperative years).²⁵

Cause-specific mortality rates vary with the presenting TNM stage for patients with FTC or HCC. The death rates tend to parallel the curves for development of distant metastases. In 5 decades of Mayo Clinic experience, the mortality rate for FTC initially exceeded that of HCC, but by 20 and 30 postoperative years, there were no significant differences in cause-specific survival (CSS) rates between FTC and HCC,²⁵ with both survival rates about 80% at 20 postoperative years and 70% at 30 postoperative years.

The risk factors that predict outcome of patients with FTC are largely the same as for patients with PTC: distant metastases at presentation, increased patient age, large tumor size, and presence of local (extrathyroidal) invasion.²⁵ To a lesser degree, increased mortality is associated with male sex and higher grade tumors. Vascular invasiveness, lymphatic involvement at presentation, DNA aneuploidy, and oxyphilic histology are potential prognostic variables unique to FTC.²⁵ The importance of vascular invasion is underscored by a study showing that FTC patients with minimal capsular invasion and no evidence of vascular invasion had 0% cause-specific mortality after 10 years of postoperative follow-up.⁷²

Prognostic scoring systems for FTC allow stratification of patients into high-risk and low-risk categories.25,73 A multivariate analysis at the Mayo Clinic found that distant metastases, patient age older than 50 years, and marked vascular invasion predicted a poor outcome.⁵⁴ As illustrated by Figure 36-4, if two or more of these factors are present, the 5-year survival rate is 47%, and the 20-year survival rate is 8%. If only one of these factors is present, the 5-year survival rate is 94%, and the 20-year survival rate is 86%.⁵⁴

Figure 36-4 Cause-specific survival among 100 patients with nonoxyphilic follicular thyroid carcinoma treated at the Mayo Clinic during the period 1946-1970 and plotted by high-risk and low-risk categories. High risk is characterized by two or more of the following factors: age older than 50 years, marked vascular invasion, and metastatic disease at the time of initial diagnosis.

From Brennan MD, Bergstralh EJ, van Heerden JA, et al: Follicular thyroid cancer treated at Mayo Clinic, 1946 through 1970: initial manifestations, pathologic findings, therapy, and outcome. Mayo Clin Proc 66:11-19, 1991.

Systems developed to predict outcomes for PTC or FTC have been applied to patients with FTC. Specifically, the AMES risk-group categorization by Cady and Rossi has proved useful in FTC.28,70 From a multivariate analysis of 228 patients with FTC treated at Memorial Sloan-Kettering Cancer Center, the independent adverse prognostic factors were identified as age older than 45 years, Hürthle cell histologic type, extrathyroidal extension, tumor size exceeding 4 cm, and presence of distant metastasis.⁷⁴ The prognostic importance in FTC of histologic grade has also been confirmed^74,⁷⁵ by the Memorial Sloan-Kettering group, who have included this factor in their assignment of risk groups to low-risk, intermediate-risk, and high-risk categories, with 20-year survival rates of 97%, 87%, and 49%.⁷⁴ The AGES and MACIS prognostic scoring systems, originally developed for PTC, have also been successfully applied to FTC.^76,77 It seems that scoring systems used in PTC can be cautiously applied in FTC as long as some of the unique features of this tumor, such as vascular invasiveness and the remarkable significance of DNA aneuploidy in HCC, are considered.²⁵

Medullary Thyroid Carcinoma

In reported studies of treated MTC, the proportion of patients with intrathyroidal node-negative tumors of 2 cm in diameter or smaller (TNM stage I) varies, depending on the number of familial cases detected by biochemical testing or DNA screening. The number of patients who present with TNM stage I MTC ranges from 5% to 25%, with the lower numbers representing the older series. Of patients, 25% to 50% present with positive neck nodes; the proportion of patients presenting with distant metastases usually exceeds the proportion with PTC but is typically less than the proportion with FTC. Stage IV cases constitute 3% to 10% of most MTC series. Figure 36-5 illustrates CSS according to pTNM stage in a cohort of 181 patients with MTC surgically treated at Mayo Clinic during the period 1940-1990.⁶³

Figure 36-5 Cause-specific survival according to pathologic TNM (primary tumor, regional nodes, metastases) stage for a cohort of 181 patients with medullary thyroid carcinoma treated at the Mayo Clinic during the period 1940-1990. Numbers in parentheses are the percentages of patients in each pathologic TNM stage grouping.

From Larsen PR, Davies TF, Hay ID: The thyroid gland. In Wilson JD, Foster DW, Kronenberg HM, Larsen PR, editors: Williams Textbook of Endocrinology, 9th ed, Philadelphia, 1998, WB Saunders, pp 389-515.

Other prognostic factors relevant to outcome in MTC include age at diagnosis, male gender, vascular invasion, calcitonin immunoreactivity, amyloid staining, presence or absence of postoperative gross residual disease, and abnormal postoperative plasma calcitonin levels.11,78–80 In a multivariate analysis from Toronto, only the presence of extrathyroidal invasion and postoperative gross residual disease were significant in the prediction of CSS.⁸⁰ In another multivariate study from Mayo Clinic, however, the only factors remaining in the final Cox model were pTNM stage III or IV disease, negative Congo red staining for amyloid, and postoperative gross residual disease.⁷⁹ Based on these three independent prognostic variables, a scoring system was devised to define four risk groups with 10-year mortality rates ranging from 5% to 100%.⁷⁹

Anaplastic Carcinoma

Patients with ATC have a very poor prognosis, and survival beyond 1 year is rare. There is no real use for an elaborate staging system for these patients, although distant and lymphatic spread at presentation, advanced age, and extremely poor differentiation have been identified as factors that make an already grim picture worse.60,81 Figure 36-6 shows survival curves for 82 patients with ATC, stratified by extent of disease at presentation.⁶⁰

Figure 36-6 Survival to death from anaplastic thyroid cancer is plotted by extent of disease at presentation. Notice the logarithmic scale of the ordinate.

From Nel CJ, van Heerden JA, Goellner JR, et al: Anaplastic carcinoma of the thyroid: a clinicopathological study of 82 cases. Mayo Clin Proc 60:51-58, 1985.

Clinical Manifestations and Patient Evaluation

At initial assessment, most patients with thyroid cancer have a palpable neck mass, which may represent the primary intrathyroidal tumor or metastatic regional lymphadenopathy. In some patients, the tumor may be clinically occult, and the impalpable lesion may first be recognized on high-resolution neck imaging ³ or during the course of a neck exploration for presumed benign thyroid disease. In patients with a family history of MTC or MEN-II syndrome, the finding of a RET oncogene mutation (identical to the proband) or abnormal calcitonin (or stimulated calcitonin) levels, or both, may necessitate elective prophylactic thyroidectomy in a patient who may prove to have early MTC visible only under the surgical pathologist’s microscope.^48,⁴⁹

Features of the history and physical examination rarely provide convincing evidence for a diagnosis of thyroid malignancy.^8,⁹ The diagnosis of thyroid cancer necessitates pathologic confirmation from cytologic or histologic material. It is generally recognized that fine-needle aspiration (FNA) biopsy is the most effective method available of preoperatively distinguishing between benign and malignant thyroid nodules.^8,⁹ All cancer diagnoses should be verified, however, by careful examination of histologic material after surgical excision of affected tissues. This approach is particularly relevant to the problem of the cellular follicular lesions described by cytologists as “suspicious” for follicular or Hürthle cell neoplasm. The diagnosis of FTC or HCC depends on the demonstration of invasion of the thyroid capsule or the adjacent blood vessels (i.e., angioinvasion), a process that typically necessitates careful evaluation of serial sections from the excised specimen for the presence or absence of such microinvasion. Even if FTC or HCC is apparently excluded at intraoperative frozen section, the resected specimen must be carefully reviewed in multiple sections from paraffin-embedded material.²⁵

FNA biopsy can usually allow a confident diagnosis of PTC, which typically represents more than 75% of clinically recognized thyroid cancers in most contemporary series.^12,^26,^28,⁵⁰ Some authorities claim that the characteristic nuclear abnormalities diagnostic for PTC may be best seen in cytologic preparations from an FNA biopsy specimen, rather than in frozen sections or in paraffin-embedded histologic material. MTC may be readily diagnosed by FNA biopsy, but in equivocal cases, amyloid stained by Congo red or immunoperoxidase labeling of intracytoplasmic calcitonin may allow a definitive preoperative diagnosis.^11,58 ATC may often be diagnosed by FNA biopsy, but it sometimes may be difficult to distinguish from carcinoma metastatic to the thyroid.⁶⁰ When the biopsy specimen is found to be suspicious for ATC, immunostaining for thyroglobulin can help confirm the diagnosis of ATC. FNA diagnosis of thyroid lymphoma is difficult to make, and verification may necessitate examination of open biopsy material and specific immunostaining for clonal B-cell and T-cell populations.⁸²

The evaluation of a patient with thyroid cancer requires obtaining a thorough history and performing physical examination with particular attention to signs and symptoms associated with local (extrathyroidal) invasion, involvement of regional (neck) nodes, and distant spread. Historical features suggesting a possible thyroid cancer include growth of a thyroid nodule over weeks or months; changes in speaking, breathing, or swallowing; and systemic symptoms of malignancy, such as weight loss, fatigue, and night sweats. On thyroid palpation, typical signs may include firm consistency of the dominant nodule, irregular shape, and fixation to underlying or overlying tissues. Evidence of suspicious regional lymphadenopathy may be present in up to one-third of patients with PTC and MTC, but it is absent in most patients with FTC.

Patients with thyroid cancer are typically euthyroid, and when measured, the serum thyroid-stimulating hormone (TSH) level is usually normal. FNA biopsy can provide a confident diagnosis of PTC, MTC, and usually ATC. Diagnosing follicular cancer preoperatively remains a problem because capsule or vascular invasion cannot be shown in cytologic material. When FCDC is suspected and the decision is made for surgical exploration, it is often advisable to draw blood for a baseline serum thyroglobulin. This thyroid-specific protein can be used as a tumor marker to assess the efficacy of future therapies. Similarly, for patients with MTC, baseline measurements of calcitonin and carcinoembryonic antigen may permit a more accurate subsequent assessment of tumor control.

A preoperative vocal cord examination may permit the demonstration of unilateral vocal cord paralysis, which may raise suspicion about pressure or invasion of the ipsilateral recurrent laryngeal nerve. Particularly with PTC, MTC, and HCC, in which initial nodal involvement is common, high-resolution, real-time ultrasound examination of the neck can provide a surgeon with valuable information regarding the extent of disease within the thyroid gland and in adjacent regional lymph nodes. With extensive nodal disease or suspected locally invasive thyroid cancer, examination of the lower central neck with computed tomography (CT) may delineate the presence of an intraluminal tumor or show more subtle evidence of tracheal or esophageal involvement with invasive disease. A high-resolution spiral CT scan of the chest or a whole skeleton isotopic bone scan is not routinely performed before initial neck exploration, but it can be useful postoperatively to exclude metastases in the mediastinum, lung, or bone, especially in patients without gross residual disease but with persistently elevated tumor markers.

Scanning with RAI has little role to play in the preoperative evaluation of thyroid malignancy. However, a postoperative whole body scan has become the gold standard for identification of iodophilic distant metastasis in patients with PTC or FTC.^8,¹⁰

Primary Therapy

Surgery

Surgery is the standard primary treatment for differentiated thyroid cancers, but the extent of initial surgery is an area of considerable controversy.5,66,83,84 Various investigators have advocated ipsilateral thyroid lobectomy with isthmectomy, bilateral subtotal thyroidectomy, and near-total or total thyroidectomy. Most institutions favor the initial resection of both lobes of the thyroid. Clinical guidelines promote near-total or total thyroidectomy as the recommended operative procedure for FCDC and total thyroidectomy for MTC.^8,^10,¹¹ Even thyroid carcinomas that are locally extensive, involving extrathyroidal structures, may be most effectively treated by a more aggressive surgical approach.^85,⁸⁶ Most studies that have evaluated the impact of surgical treatment in thyroid cancers have compared total or near-total thyroidectomy with less aggressive surgical procedures.^5,64,66,87 These reports suggest that survival rates in higher risk cases are significantly improved after the more extensive surgery.^5,⁶⁶

An often-cited reason for removing all or nearly all of the thyroid gland is reducing the risk of future recurrence of FCDC in the contralateral thyroid lobe. This approach may be most appropriate in cases of PTC, in which multicentricity is a frequent phenomenon. Some investigators claim, however, that a subsequent local recurrence in the contralateral lobe also can be managed with a secondary operative procedure.⁸⁸ Such “completion” thyroidectomies carry a higher morbidity except in the hands of highly skilled surgeons.^64,89 One advantage of initial resection of the entire gland is the subsequent ability to follow serum thyroglobulin levels and to perform meaningful radioiodine scanning more easily.^7,8,10 Whole body scans may be difficult to interpret in the setting of residual thyroid remnant, as would be found after unilateral or bilateral subtotal resection.

In addition to the initial resection of the thyroid, the question of the extent of neck node removal remains controversial. When clinically evident nodal metastases are identified by palpation or ultrasound examination, surgical removal is performed. The role of prophylactic or prognostic lymph node dissections is unclear.* For patients undergoing total or near-total thyroidectomy, removal of the central compartment lymph nodes may not add significantly to the morbidity of the procedure and should be included routinely with the primary resection.^8,89,90–92 Modified radical neck dissection is indicated for gross nodal metastases found before or during surgery, and it may improve the outcome of older PTC patients with larger tumors or evidence of extrathyroidal invasion.^93,⁹⁴

The advantage of total thyroidectomy over near-total thyroidectomy may be nonexistent or limited to patients with more advanced malignancies.64,66,95 The use of more limited surgery, such as unilateral total lobectomy, is probably appropriate only for a selected population of low-risk patients, such as patients with papillary microcarcinoma.^8,^10,⁹⁶

Adjuvant Therapy

Thyroid Hormone

Postoperative thyroid suppression therapy for FCDC is based on administration of supraphysiologic oral doses of levothyroxine. This therapy has been widely used for more than 40 years, and it is assumed that suppression of endogenous TSH deprives TSH-dependent differentiated FCDC cells of an important growth-promoting influence. Traditionally, the goal of thyroxine therapy has been complete suppression of pituitary TSH secretion, as indicated by undetectable levels of serum TSH when measured in sensitive immunometric assays. Uncertainty surrounds the level to which the serum TSH level must be suppressed to maximize benefit while avoiding potential long-term complications.4,6,8,10

With the increasing availability of sensitive TSH assays, meticulous titration of the level of TSH suppression has become possible. If patients with FCDC are deemed to be at high risk for relapse or mortality on the basis of scoring or staging systems,62,69,70 or if they have persistent or recurrent carcinoma that cannot be eradicated by surgery, RAI, or other therapeutic measures, the goal would be to maintain the TSH level at less than 0.1 mU/L indefinitely in the absence of specific contraindications. In most patients with PTC who would be classified as low risk by prognostic scoring systems,^66,69,70,75 it is generally accepted that the degree of TSH suppression would be less stringent, and the goal for basal serum TSH should be in the readily detectable, barely subnormal range of 0.1 to 0.5 mU/L.^6,8,10,63,95

Radioactive Iodine Therapy

Radioiodine remnant ablation (RRA) has been defined as “the destruction of residual macroscopically normal thyroid tissue after surgical thyroidectomy.”^95,⁹⁷ Typically, RRA is used to complete the initial therapy in a patient whose FCDC has been completely resected (i.e., no gross residual disease is reported at the conclusion of the primary neck exploration). RRA is a procedure that is offered to patients with FCDC who have undergone “potentially curative” surgical treatment, and it should not be confused with RAI therapy, in which larger doses of iodine-131 (¹³¹I) are used in an attempt to destroy persistent neck disease or distant metastatic lesions.^{82,95,97,98,99}

Proponents of RRA describe at least three advantages for this adjunctive therapy. First, by being actively trapped by normal thyroid cells, RRA is thought to destroy occult microscopic carcinoma cells within the thyroid remnant. Second, the later detection of persistent or recurrent disease, particularly in the neck by RAI scanning, is facilitated by the destruction of remaining normal tissue. Third, the performance of RRA is thought to increase the value of serum thyroglobulin measurement during follow-up. This last benefit has convinced many physicians to consider RRA in patients with FCDC who are, by prognostic scoring or staging, deemed to be at low risk for tumor relapse or cause-specific mortality.^19,⁹⁷ Other physicians have not advocated RRA in these low-risk patients, however, because of a lack of evidence of improved outcome.^94,95,97,98

Investigators have reported benefit and a lack of benefit from the use of RRA.* It is unclear from these reports whether outcome is improved with the routine use of RRA in the initial therapy of all patients with differentiated thyroid carcinoma, largely because of difficulties obtaining a comparable control group and the potential for selection biases, which may account for the outcomes reported. Sawka and colleagues⁹⁷ systematically reviewed 1543 English references to determine whether RRA decreases the risk of thyroid cancer–related death or relapse after bilateral thyroidectomy for PTC or FTC. In 18 cohort studies not adjusted for prognostic factors or interventions, the benefit of RRA in decreasing thyroid cancer–related mortality and any relapse by 10 years was inconsistent among centers. However, pooled analyses suggested a statistically significant treatment effect of RRA in terms of 10-year outcomes for locoregional recurrence (relative risk 0.31) and distant metastases (absolute decrease in risk 3%). The conclusion of this meta-analysis was that RRA might be beneficial in decreasing relapse of well-differentiated thyroid cancer; however, Sawka and colleagues⁹⁷ said that “the incremental benefit of remnant ablation in low-risk patients treated with bilateral thyroidectomy and thyroid hormone suppressive therapy is unclear.”

The National Thyroid Cancer Treatment Cooperative Study Group (NTCTCSG) was founded in 1986 and has maintained over more than two decades a registry contributed to by 11 North American institutions prospectively following a large nonrandomized cohort of patients with FCDC, with the objective of assessing the effects of initial and longitudinal management on outcomes. By June 2001, 2936 patients were registered, and the more recently published results of outcome showed that “no treatment modality, including lack of radioactive iodine, was associated with altered survival in stage I patients.” Jonklaas and colleagues ⁴ “were unable to show any impact, positive or negative, of specific therapies in stage I patients.” These authors concluded that “postoperative RAI (RRA) therapy does not provide significant benefit in stage I patients, and could even be harmful,” and suggested that “further evaluation of the potential risks and benefits of treatment in stage I patients is, therefore, indicated.”⁴

Use of prognostic scoring schemes may help to define patients at significant risk for relapse and who are more appropriate for consideration of remnant ablation. At the Mayo Clinic, patients with PTC who have MACIS scores greater than 6.0 or are found to have persistent elevation of serum thyroglobulin levels after 6 to 8 postoperative weeks are considered for an ablative procedure. Use of such a scheme is applicable only if the patient undergoes a near-total or total thyroidectomy. If the goal is to minimize the use of RAI, more extensive surgery such as near-total or total thyroidectomy is more appropriate in low-risk patients. However, if RRA is planned, more extensive surgery may not give any advantage. In high-risk patients, the use of RAI in addition to more aggressive surgery seems warranted.94,98,100

After determining the need for remnant ablation, the question of dose after the primary surgical procedure is also an important consideration. In the United States, low-dose regimens of less than 30 mCi have been used when the amount of residual thyroid tissue is small. Higher doses of 75 to 150 mCi may be appropriate when microscopic residual is present or when the initial scan suggests an unsuspected metastatic site. Use of low-dose RAI in such a patient may decrease the ability of the tumor subsequently to concentrate sufficient radioiodine and should be avoided.

Further treatment with RAI ideally should reflect the local or distant uptake of iodine in residual differentiated thyroid cancer tissue. Adequate ability of thyroid cancer cells to concentrate radioiodine is variable, and this approach may not be appropriate in all patients. Patients with FTC may concentrate RAI better than patients with PTC.⁹⁹ In a multicenter Canadian study, Simpson and associates⁹⁹ reviewed the outcomes of more than 1500 patients with PTC or FTC and studied 321 patients who were treated with RAI. The use of RAI resulted in significantly improved local control in patients with microscopic residual disease. Cause-specific survival was also significantly better for PTC patients with microscopic residual disease who received RAI. Most patients received only thyroid remnant ablation doses (50 to 125 mCi), but total (cumulative) therapeutic doses given in one to seven treatments ranged from 75 to 1000 mCi.

The definitive role of RRA continues to evolve.100,101 Multiple studies have suggested that RRA be used routinely in high-risk patients. Although the definition of “high risk” varies, most investigators favor using RAI in patients with large tumors or extensive extrathyroidal invasion and in patients with incompletely resected disease or distant metastases. Robbins and Schlumberger¹⁰⁰ recommended RRA for “any individual with a carcinoma larger than 1.5 cm or with a thyroid carcinoma of any size with obvious lymph node involvement, extrathyroidal extension, or multicentricity.” The clinical guidelines proposed by the European Thyroid Association and the American Thyroid Association should allow some greater uniformity of treatment approach.^8,^10,¹⁰¹

External Beam Irradiation

The appropriate application of EBRT in PTC and FTC is controversial, and this modality is not uniformly used.101,102 A better understanding of prognostic factors that are specific to local recurrence may help identify an appropriate population of patients who would be candidates for EBRT. Most prognostic schemes consider only the overall risk of relapse and CSS.

Historically, indications for irradiation have included residual macroscopic or microscopic disease, extrathyroidal extension, multiple lymph node involvement, Hürthle cell histology, older age of patients, large tumor size, and residual disease that does not take up RAI. Several investigators have retrospectively reviewed their experience with EBRT and suggested improved local control or survival benefit or both.* The patient populations in these studies are not uniform with respect to other treatments that also effect local control, making interpretation difficult.

Most series support the use of adjuvant irradiation for patients with microscopic residual disease or narrow surgical margins of resection. An early Canadian study of follicular and papillary carcinomas evaluated the role of EBRT according to completeness of surgical resection.¹⁰³ Table 36-4 summarizes these findings. Although the overall numbers are small, they suggest that patients with surgically narrow or microscopically positive margins may benefit from the addition of EBRT in moderate doses. These results were achieved with minimal morbidity.

TABLE 36-4 Local Recurrence of PTC and FTC at Princess Margaret Hospital by Method of Treatment

A larger synopsis of 13 Canadian institutions in which EBRT was used predominantly in patients with high-risk features, such as extrathyroidal invasion, high-grade malignancies, and older age, revealed a significant improvement in local control for patients who had microscopically positive or narrow surgical margins.¹⁰¹ In this study, the overall survival (OS) of PTC patients with presumed or microscopic residual was significantly improved with the addition of adjuvant EBRT. These improvements were seen with or without the use of RAI. Results from Germany for patients with pathologic T4 disease (based on the older staging system) with or without lymph node involvement also suggest a benefit from EBRT with respect to local control and distant failure rates, especially in patients older than 40 years.¹⁰⁵ In a review from Princess Margaret Hospital in Toronto of 207 patients with microscopic residual, the addition of EBRT did not significantly improve the CSS or the local control (p = .38 and p = .14).¹⁰⁶ However, a more recent review with more patients and longer follow-up revealed an improved 10-year CSS in high-risk patients (age >60 years, extrathyroid extension, microscopic residual disease, no gross residual disease) with the addition of EBRT.^118,¹¹⁹ A review from Mount Sinai Hospital in Toronto found that the use of EBRT was the only predictor of significantly improved disease-specific survival and OS in patients with well-differentiated thyroid cancer with extrathyroidal extension.¹²²

Benker and associates ¹⁰⁸ described 932 patients with differentiated thyroid malignancies and found no OS benefit with the addition of EBRT. On subset analysis, however, patients older than 40 years with T3 or T4 disease experienced improved survival that approached statistical significance (p = .09). All patients in this series received thyroxine-suppressive therapy and RAI.

The role of EBRT in the setting of gross residual disease has been evaluated by multiple investigators, with a wide range of outcomes. A study from the Royal Marsden Hospital showed complete response in 37% of patients and partial response in an additional 25%.¹²¹ The Princess Margaret Hospital study had 33 patients with macroscopic disease treated with EBRT with or without RAI; the 5-year local recurrence-free rate was 62%.¹⁰⁶ Tubiana and associates¹⁰⁴ from Institut Gustav-Roussy evaluated 97 patients with macroscopic residual based on the surgeon’s clinical assessment. Local recurrence after EBRT was only 15% compared with 32% in a similar surgical category without postoperative EBRT. None of these patients received RAI. With the ability to deliver higher doses of EBRT using techniques such as intensity-modulated radiotherapy (IMRT), it may be feasible to give higher doses than were used in these studies and perhaps increase the long-term control of disease in this difficult patient population.

Although the definitive role of EBRT for differentiated thyroid carcinoma remains controversial, retrospective data support its use in a selected population of patients who have microscopic or presumed microscopic residual cancer.¹²³ Evidence exists not only for PTC and FTC but also for Hürthle cell carcinoma and MTC.¹²⁴^–¹²⁶ The only randomized trial evaluating the role of EBRT was initiated in 2000 in patients with pathologic T4, N0-1, M0 disease after surgical resection and ablative ¹³¹I therapy.¹²⁷ The study closed because of inadequate accrual in March 2003 despite acceptable toxicity.¹²⁸ The study authors indicated that at early follow-up, there are lower than expected relapse rates in the nontreatment arm. The improvements in surgical management of these patients and the appropriate application of RAI may have accounted for the decreased relapse and mortality rates compared with historical studies.

The use of EBRT for patients with differentiated thyroid cancer needs to be carefully tailored to the specific patient and applied after use of RAI. Although efforts to improve on surgical outcome have focused mainly on RAI treatment in patients with a high risk of relapse, there seems to be an evolving role for EBRT in their management. The rationale for inclusion of EBRT up-front in the management of this subset of patients is to (1) treat tissues that are known to take up less RAI (areas of extrathyroidal or extranodal extension, which are known poor prognostic indicators); (2) treat carcinomas that fail to concentrate and retain RAI (20% of all differentiated thyroid cancers, patients with Hürthle cell tumors, and patients >40 years old who commonly have less sodium iodide symporter); (3) spare some of the systemic toxicity of RAI treatment (when adequate treatment of residual disease entails excessive doses of radioiodine or when repeated treatments are likely to be necessary, increasing cumulative systemic toxicity); and (4) spare some of the morbidity associated with uncontrolled locoregional cancer, such as obstruction of the esophagus or trachea or both, neurovascular compromise, pain, hemorrhage, need for laryngectomy, and need for repeated surgical procedures.

The current American Thyroid Association guidelines suggest that EBRT should be considered in patients older than age 45 with grossly visible extrathyroid extension at the time of surgery and a high likelihood of microscopic residual disease and patients with gross residual tumor in whom further surgery or RAI would likely be ineffective.¹¹⁷ Changes in the staging of locally advanced thyroid cancers in 2002 by categorizing minimal extrathyroidal extension as T3 disease and creating two T4 categories should help delineate further the highest risk population. In the clinical setting of T4b disease, it is unlikely that complete surgical resection can be achieved, and use of EBRT in addition to RAI seems to help control gross disease. For patients with locally recurrent disease, early use of EBRT as a component of treatment should be considered.

Treatment of the thyroid bed or gland and its draining lymph node areas has historically been a challenging undertaking. Advances in treatment modalities include three-dimensional conformal multiple-beam treatments and IMRT (and intensity-modulated arc or volumetric therapy) treatments. Typical total doses are 50 to 60 Gy administered as 2 Gy per treatment, 5 days per week, over 5 to 6 weeks for microscopic disease (adjuvant treatment). Total doses of 60 to 70 Gy over 6 to 7 weeks are employed for gross residual disease. Treatment volumes typically encompass all lymph node regions “at risk” (typically from the hyoid bone to the carina, levels II, III, IV, V, and VI). Although there is no definitive evidence of a dose-response relationship between radiation dose and local control probability, there is some suggestion that higher doses are associated with lower local recurrence rates.109,116 With the increasing availability of IMRT and proton radiotherapy, dose escalation may be more readily achievable in the adjuvant setting without increasing toxicity.¹²⁹^–¹³¹ Stereotactic body radiotherapy (SBRT) has been proposed as a highly targeted technique that is efficient (one to three treatments) and effective in controlling cervical lymph node recurrence of nonanaplastic thyroid cancer (100% tumor control) with no serious adverse events.¹³²

Chemotherapy

The role of chemotherapy in the treatment of differentiated thyroid carcinoma is limited. Most agents have low response rates. The most active single agent reported in the treatment of thyroid cancers is doxorubicin (Adriamycin). Other reportedly active agents include cisplatin, carboplatin, and etoposide (VP-16).

The Eastern Cooperative Oncology Group (ECOG) reported the only randomized study of the use of chemotherapy in advanced thyroid cancer.¹³³ Patients with locally advanced or metastatic FTC, PTC, MTC, and ATC were eligible. In this study, patients were randomly assigned to chemotherapy with doxorubicin alone or doxorubicin with cisplatin. Of the 84 eligible patients, the investigators reported a complete response rate of 6% and a partial response rate of 15%. A partial or complete response to chemotherapy occurred in 8 (23%) of 35 patients with advanced differentiated thyroid carcinomas. Overall, there was no advantage to the combination-drug regimen over doxorubicin alone. All five complete responders were in the combination arm, however.

Thyroid Lymphoma and Anaplastic Thyroid Carcinoma

The most common poorly differentiated tumors of the thyroid, lymphoma and ATC, are treated differently from well-differentiated FCDC and MTC. For that reason, these two tumors and their treatment plans are considered separately.

Primary Thyroid Lymphoma

Since the development of immunocytochemical staining, this distinct entity has become more accurately recognized.¹³⁴ Most thyroid lymphomas are diffuse, large B-cell lymphomas (DLBCLs), but significant numbers are in the category of lymphomas known as mucosa-associated lymphoid tissue (MALT) lymphomas or mixed tumors containing elements of DLBCL and MALT. Rarely, other types of lymphoma, including follicular lymphomas and Hodgkin’s lymphomas, have been reported.^134,¹³⁵ Thyroid lymphomas usually manifest with a rapidly enlarging mass that is sometimes associated with respiratory and esophageal symptoms. These lymphomas typically occur in middle-aged and older women and are often associated with chronic thyroiditis.^134,¹³⁵ MALT lymphomas have a more indolent natural history, whereas DLBCLs tend to be more aggressive.

An early review by Doria and colleagues ¹³⁶ before the widespread use of immunohistochemical staining suggests that the distant failure rate with thyroid lymphomas was seen primarily in patients with non-MALT lymphomas. Their review included 211 patients in 11 series and showed an 11% distant relapse rate for patients with MALT lymphomas compared with 45% for patients with localized non-MALT lymphomas. This review suggests that local therapy alone in patients with DLBCL is inadequate and that systemic therapy is warranted.

Surgery has a diagnostic role only in the management of thyroid lymphomas, and this role is diminishing with the increasing accuracy of FNA.¹³⁷ Some investigators contend that surgery can provide prognostic information, influence treatment options, and improve outcome.^135,¹³⁸^–¹⁴⁰ They assert that the use of surgery may obviate the need for chemotherapy in some patients or identify a population requiring local therapy only. One investigator reported five patients with MALT lymphoma or small lymphocytic lymphoma treated with surgery alone who had no relapses.¹³⁵ Surgery may have a palliative role in selected patients with obstructive symptoms.¹⁴¹ In the absence of prospective or randomized data, the improved outcome of patients who are able to undergo resection may indicate only selection of a favorable subgroup and not provide clear evidence of survival advantage associated with the surgery itself, and it does not eliminate the consideration of irradiation or chemotherapy. The infiltrative nature of this disease can lend itself to a higher rate of surgical complications involving the parathyroid glands and recurrent laryngeal nerves. Some investigators have failed to show any advantage to more aggressive surgery, provided that the patient receives adequate irradiation with or without chemotherapy.^142,¹⁴³

The use of irradiation alone in the treatment of thyroid lymphomas has been reported by multiple institutions.^138,¹⁴²^–¹⁴⁶ The reports indicate that the OS for patients with Ann Arbor stage I and II disease ranges from 40% to 93%. Patients with stage I disease have a better outcome. Some of these reports indicate improved outcome with neck and mediastinal irradiation over neck treatment alone.^142,¹⁴⁷^–¹⁴⁹ Other authors have not found this association.^138,¹⁴⁰ In one report of 27 patients treated with irradiation alone or in combination with chemotherapy, all four local failures occurred in patients treated solely with irradiation.¹⁴⁶

The role of EBRT alone in the treatment of thyroid lymphomas is limited, and given the randomized clinical data for DLBCL with combined chemoradiation, EBRT as the sole treatment modality should be given only when chemotherapy is not feasible or tolerated. With the increased risk of distant relapse reported by Doria and colleagues,¹³⁶ survival was improved with the use of systemic chemotherapy in addition to local EBRT. Other authors have concluded that chemotherapy improves the outcome in these patients.^137,¹⁴⁴

Two large prospective randomized studies have evaluated the use of combined-modality therapy in thyroid lymphomas. Studies by ECOG and Southwestern Oncology Group (SWOG) evaluated the use of CHOP (cyclophosphamide, hydroxydaunomycin, Oncovin [vincristine], prednisone) chemotherapy with or without involved-field irradiation for stage I or II non-Hodgkin’s lymphomas, including thyroid lymphomas.150,151 The number of thyroid lymphoma patients enrolled in these studies is not published, but 18% of the patients in the ECOG study had stage IE disease (i.e., disease confined to one extranodal site). Both studies indicated a survival benefit for combined-modality therapy. Doses of 30 Gy were used in the ECOG study after eight cycles of CHOP chemotherapy except when there was only a partial response to chemotherapy. In that circumstance, a dose of 40 Gy was delivered using standard fractionation. In the SWOG study, the combined-modality arm received only three cycles of CHOP along with 40 to 55 Gy. The OS at 6 years in the ECOG study was 84% for combined-modality therapy compared with 70% with chemotherapy alone (p <.05).¹⁵⁰ Further data from this study await publication. The SWOG report estimated an 82% overall survival rate compared with 72% for CHOP alone (p = .02).¹⁵¹ Based on these randomized studies, it seems that the use of combined-modality therapy offers the most appropriate management of this disease process.

The radiotherapy treatment volume should include the thyroid region and regional lymph nodes. This should include the superior mediastinum and low cervical, pretracheal, internal jugular, and supraclavicular nodes.

Further improvement in outcomes for DLBCLs has been shown with the addition of monoclonal antibodies, especially rituximab, to combination chemotherapy.^152,¹⁵³ Other immunologic and radiolabeled agents are being evaluated.

There is minimal information about the use of chemotherapy alone in this particular disease.¹⁵⁴ Data from the SWOG and ECOG studies show a high response rate with CHOP chemotherapy alone. Given the locally aggressive nature of this extranodal site and the bulky disease that can extend into the mediastinum, however, the use of combined-modality therapy is preferable over chemotherapy alone.

Anaplastic Thyroid Carcinoma

The treatment of ATC is frustrating because of the rapidly progressive nature of this disease and its high fatality rate. Most patients are women presenting in the sixth to seventh decade of life, typically with a mass rapidly enlarging over a short time interval. Symptoms of hoarseness, dyspnea, dysphagia, and pain are common in this form of thyroid malignancy. Persistent or recurrent local disease is often the cause of death. Distant metastases are present in almost half of patients at presentation.

The optimal role of surgery has not been well defined. Surgery may be performed initially in some patients, but this probably reflects the selection of a favorable population of patients. The M.D. Anderson Cancer Center reported that more extensive surgery by itself did not provide a survival advantage.⁸¹ Overall, most patients present with unresectable or metastatic disease, and the initial surgery in these situations is primarily for diagnostic purposes. In most reports, however, the only long-term survivors were patients who had undergone surgical resection.^155,¹⁵⁷^–¹⁶⁰ For this reason, surgery should be strongly considered as a component of treatment. Surgery has been performed before or after irradiation, chemotherapy, or both.^81,¹⁶¹ The sequence of surgery relative to radiation and chemotherapy may be primarily a reflection of the extent of disease at time of patient presentation to the treating physicians.¹⁶²

Radiation has been used in the treatment of these advanced tumors, but because of the high metastatic rate and high rate of local failure with irradiation alone, the use of chemotherapy in addition to irradiation is warranted.¹⁶³ Multiple studies have investigated the use of multimodality therapy.¹⁵⁶^–¹⁵⁸ ¹⁶¹ The use of hyperfractionated irradiation has been examined because of the rapid proliferation rate of these malignancies.^156,^158,^161,¹⁶⁴ Various approaches have been reported, but most studies use doxorubicin-based chemotherapy given concurrently with local irradiation.^133,^156,^157,¹⁶⁵

In approaching ATC, we view two issues to be paramount: (1) the eradication of localized disease in the neck, with protection of the airway, esophagus, and neurovascular structures, and (2) the control of metastatic disease, which almost always contributes to the death of ATC patients. It follows that a strategy combining locoregional and systemic therapy may offer the greatest promise in ATC.

Although long-term survival (>1 year) after combined-modality adjuvant therapy in ATC has been so rare as to provoke reports of single cases,¹⁶⁶ subsequent series support the benefit of combined-modality approaches. Swaak-Kragten and colleagues¹⁶⁷ conducted a study of 30 ATC patients treated with an adjuvant approach combining surgery, radiotherapy, and single-agent chemotherapy with doxorubicin administered concurrently with radiotherapy and continued after radiotherapy completion and reported an encouraging improvement in 1-year survival from 9% (historical controls) to 23% (new patient cohort). Haigh and associates¹⁶⁸ reported a 75% 2-year survival in eight patients undergoing complete surgical resection and adjuvant radiotherapy and chemotherapy. Tan and coworkers¹⁶⁹ reported an estimated 5-year survival of 60% for five patients who underwent complete resection. Four patients had received postoperative radiotherapy, three with postradiotherapy doxorubicin-based chemotherapy. Pudney and colleagues¹⁷⁰ reported a median OS of 13 months for five patients with stage IVB ATC treated with either radiotherapy concurrent with doxorubicin or induction docetaxel, doxorubicin, and cyclophosphamide. Three patients developed progression of local disease, and one developed distant metastases. Higashiyama and associates¹⁷¹ reported a 12-month survival rate of 44% in nine patients with stage IVb ATC treated with induction paclitaxel. This survival rate compared favorably with a historical control group of 50 patients treated without induction paclitaxel; their 12-month survival was just 5.9%.

The optimal EBRT fractionation schedule has not yet been determined. Of patients treated with moderate-dose (57.6 Gy) slightly accelerated radiotherapy, given twice daily 3 days per week, combined with concurrent weekly doxorubicin, 70% required a 1-week interruption in treatment because of pharyngoesophagitis and tracheitis.¹⁵⁶ Moderate-dose (60.8 Gy) accelerated fractionation even without concurrent chemotherapy may be too toxic with a median OS of just 10 weeks.¹⁷² Low-dose (40 Gy) hyperfractionated accelerated radiotherapy with sequential but not concurrent chemotherapy (doxorubicin) seems to be well tolerated with 1-year and 3-year OS of 50% and 35%.¹⁶⁵ Moderate-dose (60 Gy) hyperfractionated accelerated radiotherapy without concurrent chemotherapy seems to be well tolerated with a median OS of 13.6 months and 1-year and 5-year OS of 66.7% and 0%.¹⁷³ However, others have reported less than 10% 1-year survival with hyperfractionated accelerated radiotherapy to a median dose of 57 Gy in patients with nonmetastatic disease.¹⁷⁴

In a report of the Surveillance, Epidemiology and End Results database from the National Cancer Institute, the only factors predicting for lower cause-specific mortality were age younger than 60 years, intrathyroidal tumor, and combined use of surgery and irradiation.¹⁷⁵ However, the overall cause-specific mortality rate was 80.7% at 1 year. The use of multimodality therapy, including surgery, irradiation, and chemotherapy, is appropriate and offers the best chance for long-term survival, but more active systemic agents need to be found to improve survival significantly.

Irradiation Techniques and tolerance

Treatment Volumes and Doses

Treatment of the thyroid bed or gland and the regional lymph nodes is challenging because of the contour of the body; the potential extension of disease to the upper mediastinum; involvement of lymph nodes in the neck and mediastinum; and the close proximity of the lungs, spinal cord, esophagus, larynx, pharyngeal constrictors, and brachial plexus. Historically, several approaches have been reported for the initial extended fields, including a single anterior electron field,¹⁰⁶ an anterior field with a posterior supplemental field for the mediastinum, and lateral fields.¹⁰⁴

For optimal treatment in an adjuvant setting or for locoregionally advanced or recurrent disease with curative intent, the clinical target volume should include the thyroid bed, lymph nodes of the neck (typically from the hyoid bone to the carina, levels II through VI), and upper mediastinum. When treating for palliation, it is reasonable to consider treating only the gross tumor volume plus margin. Treatment planning should be image based (CT simulation) and incorporate CT, ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography with CT (fused PET/CT), or magnetic resonance imaging (MRI).

At Mayo Clinic, we have chosen to use IMRT or intensity-modulated arc therapy in the treatment of these patients, which allows us to administer a more conformal high dose (60 to 70 Gy) with improved homogeneity to the thyroid bed or gross disease and high-risk areas (lymph nodes and tracheoesophageal groove), while lowering the dose to normal organs at risk, including the salivary glands, epiglottis, false vocal cords, true vocal cords, arytenoids, central and posterior aspects of the pharyngeal constrictors and esophagus, brachial plexus, and spinal cord. The use of IMRT has allowed for more elegant treatment of this disease site compared with conventional and three-dimensional treatment planning (Fig. 36-7). IMRT improves the minimum and mean dose to the planning tumor volume and significantly reduces the dose to the spinal cord.^129,^131,^176,¹⁷⁷ This is true for treatment of the thyroid tumor bed or the thyroid region and locoregional nodal sites.

Figure 36-7 A, Three-dimensional treatment planning isodose curves for a patient with locally recurrent and progressive papillary thyroid carcinoma, deemed unresectable. Treatment uses anteroposterior fields followed by oblique fields to the left neck for a total dose of 70 Gy. B, Intensity-modulated radiotherapy (IMRT) treatment plan for a patient with resected recurrent Hürthle cell carcinoma. Treatment uses 9 IMRT fields delivering 60 Gy to central lower neck and 54 Gy to lateral upper neck and mediastinum.

The radiation dose used in the postoperative setting for microscopic residual disease should be 60 Gy in 6 weeks. However, some centers have suggested 40 Gy given over 3 to 3.5 weeks is adequate.¹⁰³ For known gross residual disease, a dose escalation to 66 to 70 Gy is reasonable. One retrospective study from the United Kingdom suggested a possible dose-response effect at greater than 50 Gy for patients treated with curative intent.¹⁰² The use of higher doses without added morbidity may be feasible with IMRT. Although there is no definitive evidence of a dose-response relationship between radiation dose and local control probability, there is some suggestion that higher doses are associated with lower local recurrence rates.^109,116

Treatment Tolerance

With the increasing availability of IMRT, intensity-modulated arc therapy, and proton radiotherapy, dose escalation may be more readily achievable in the primary setting without increasing toxicity and lowering toxicity in the adjuvant setting.¹²⁹^–¹³¹ Ideal planning target volumes and dose remain inadequately characterized, however. Early experiences with IMRT seem to suggest that escalation of dose to subclinical sites of possible involvement (nodal regions outside the thyroid bed to 54 Gy), the margin-negative thyroid bed (60-63 Gy), margin-positive regions (66 Gy), and gross disease (70 Gy) is technically feasible without increasing toxicity.^129,¹³¹ One treatment is given each day, 5 days a week, 1.8 to 2.0 Gy per treatment. Acute radiation reactions include dermatitis, laryngitis, pharyngitis, tracheitis, and esophagitis. Late sequelae are uncommon but may include laryngeal edema, cartilage necrosis, esophageal stenosis, myelitis, brachial plexopathy, and pulmonary fibrosis.

Treatment Algorithm, Conclusions, and Future Possibilities

Treatment Algorithm

The initial approach to a potentially malignant thyroid mass is to perform FNA biopsy. This technique permits the identification of patients with definite cancer or for whom there is a high degree of suspicion for malignancy. At initial operation (i.e., open biopsy, thyroid lobectomy, and near-total or total thyroidectomy), the surgeon can assess tumor size, determine the presence or absence of gross local invasion or regional lymph nodal metastasis, and define the postoperative status (i.e., presence or absence of gross residual disease). The surgical pathologist can identify the degree of tumor differentiation (i.e., histologic grade); the tumor histologic type; and, if requested, the DNA ploidy.

Typically, FCDC is treated by near-total or total thyroidectomy unless the primary tumor is 1 cm or less in diameter. MTC is treated by total thyroidectomy, whereas ATC and lymphoma probably require only an open biopsy for verification of diagnosis. Postoperatively, all patients undergoing more than a lobectomy require thyroid hormone therapy, but only patients with FCDC should receive TSH-suppressive doses of thyroxine.

Knowledge of patient age, cell type, pTNM stage, and postoperative status permits classification of patients into risk groups, which may influence the selection of subsequent postoperative adjuvant therapy. The treatment of patients with different tumor histologies is outlined in the following paragraphs and summarized as four steps in the accompanying treatment algorithm (Table 36-5).

TABLE 36-5 Treatment Algorithm

Step I	Initial neck exploration/thyroid resection
	FCDC: usually near-total or total thyroidectomy
	MTC: total thyroidectomy
	ATC: open biopsy or subtotal thyroidectomy
	Lymphoma: FNA or open biopsy
	Lymph nodes: removal of central compartment nodes in FCDC and MTC; modified radical neck dissection for involved lateral nodes
Step II	Thyroid hormone therapy
	Replacement doses for MTC, ATC, lymphoma
	TSH-suppressive doses for FCDC except microcarcinoma
Step III	Outcome prediction by risk-group classification
	Gauged according to age, stage, histologic type, and cancer type-specific scoring systems (e.g., AMES, MACIS)
Step IV	Patient selection for RAI therapy, EBRT, chemotherapy, or combinations
	FCDC: RAI therapy indicated for distant spread, unresectable or residual neck tumor, possibly invasive disease in PTC and most cases of FTC or HCC; EBRT for local or metastatic tumor nonresponsive to RAI therapy; almost no role for chemotherapy in differentiated FCDC
	MTC: residual or recurrent neck disease considered for EBRT; octreotide or chemotherapy considered for palliation only
	ATC: postbiopsy EBRT and concurrent chemotherapy
	Lymphoma: CHOP chemotherapy and radiotherapy

AMES, age, metastasis, extent, and size; ATC, anaplastic thyroid cancer; CHOP, cyclophosphamide, hydroxydaunomycin, Oncovin (vincristine), prednisone; EBRT, external beam radiotherapy; FNA, fine-needle aspiration, FCDC, follicular cell–derived cancer; FTC, follicular thyroid carcinoma; HCC, Hürthle cell carcinoma; MACIS, metastasis, age, completeness of resection, invasion, and size; MTC, medullary thyroid carcinoma; PTC, papillary thyroid carcinoma; RAI, radioactive iodine; TSH, thyroid-stimulating hormone.