Differentiated Thyroid Carcinoma: Risk Group Assignment and Management Controversies

Introduction

Thyroid cancer has attracted interest because of its broad range of biologic phenotypes, ranging from the treatable papillary carcinomas to the often lethal anaplastic thyroid carcinoma. The most common thyroid cancers are the differentiated thyroid carcinomas, which, in general, carry high long-term survival rates.

 Partly because of the low mortality and prolonged recurrence patterns of differentiated thyroid carcinomas, prospective randomized trials have not been conducted. Hence, despite the well-documented natural history of this cancer, its treatment remains controversial.

 The emergence and widespread acceptance of risk group assignment in the management of differentiated thyroid cancer have diminished some of the clinical controversy surrounding its treatment. Recent progress in identifying the molecular events associated with the malignant transformation of the thyroid follicular cell has opened up the possibility that clinical treatment decisions can be based on the biologic mechanism of the disease process.

Epidemiology and Natural History

An estimated 15,600 thyroid malignancies will be treated in the United States this year. Of these, 80% to 90% will be well-differentiated thyroid cancers. Papillary carcinoma accounts for 70% of all differentiated thyroid cancer cases.[1] The incidence of papillary cancer is increasing, with the peak age of occurrence between the third and fourth decades of life.[1] Also, 90% of radiation-induced thyroid malignancies are diagnosed as papillary carcinomas. Despite its incidence, the overall mortality of papillary carcinoma remains low (6% to 12%).[2,3] Regional lymph node metastases occur in 30% to 60% of patients with papillary carcinoma. Nevertheless, when compared to other epithelial malignancies, the biologic behavior of papillary carcinoma is indolent, especially in adults less than 45 years of age. Long-term survival is common, even with persistent or recurrent disease (Table 1).

Follicular carcinomas are more commonly found in association with endemic goiter and geographic areas of chronic iodine(Drug information on iodine) deficiency.[4] The peak incidence of follicular carcinoma is in the fifth to sixth decades of life. Histologically, follicular carcinomas have been classified as microinvasive (low-grade) or macroinvasive (high-grade) carcinomas.[5] The distinctive histologic feature of high-grade lesions is angioinvasion. Microinvasive follicular carcinomas have a less aggressive clinical course with low rates of metastasis. Conversely, macroinvasive follicular carcinomas carry a poorer prognosis with higher rates of metastases.[6] The more aggressive follicular thyroid carcinomas tend to occur in men over 50 years of age.[7]

Although the pathogenesis of differentiated thyroid cancer is unknown, some theories suggest an association with such clinical syndromes as adenomatous polyposis.[8] The existence of a number of defective signal transduction factors (ras, ret, gsp, trk), as well as nuclear regulatory factors (eg, p53, retinoblastoma gene), has been implicated in the pathogenesis of differentiated thyroid carcinoma, implying potential familial inheritance patterns (Figure 1).[9,10]

One of the more clearly defined genetic events in differentiated thyroid cancer is the involvement of the RET proto-oncogene in papillary carcinoma (Figure 2).[11,12] Through a pericentromeric inversion in chromosome 10, the position of the RET proto-oncogene is rearranged. After the inversion, the gene is repositioned in front of a promoter, thereby activating it. The rearranged gene, called the papillary thyroid carcinoma oncogene (PTC), is seen in up to 40% of papillary cancers.[11]

Clinical Presentation and Diagnosis

Most patients with differentiated thyroid cancer present with a solitary, asymptomatic, hypofunctional nodule within the thyroid. Thirty percent of lesions are smaller than 1.5 cm.[12-14] On occasion, the only positive clinical finding is palpable cervical adenopathy harboring metastatic disease.[15]

From 6% to 10% of the adult US population have solitary thyroid nodules, and only 5% of these are malignant.[5] Hence, the first diagnostic challenge is to identify this relatively small population with malignancy. The next challenge is to detect thyroid carcinomas that are likely to display an aggressive biologic phenotype.

Fine-needle aspiration (FNA) remains the best single diagnostic tool (Figure 3). It carries a negative predictive value of 94% for benign disease and a positive predictive value of 96% for malignant disease. If one takes into account those patients with indeterminate or inadequate FNA, the positive predictive value is still over 50%. In comparison, scintigraphy and ultrasound have a positive predictive value of only 20%.[15]

A malignant FNA cytology mandates surgical resection with no further work-up. A cystic lesion that completely resolves after FNA warrants no other treatment. Surgical intervention is recommended only if the mass persists or reappears following aspiration.[16] Benign pathology on FNA can be followed by periodic clinical and ultrasound examination, as well as by thyroid-stimulating hormone (TSH) suppression.

Inadequate FNA cytology often results from technical error and should be repeated. Indeterminate FNA is usually secondary to the cytologist’s inability to differentiate follicular adenoma from follicular carcinoma (false-negative rate, 27%).[15] Under these circumstances, two clinical paradigms can be used.

In patients in whom the TSH level is normal or elevated, the clinician can reasonably assume that the nodule is hypofunctioning. This implies an increased probability of malignancy, and thus, surgical resection can be recommended without further diagnostic work-up. An equally accepted clinical strategy is to perform scintigraphy in these patients. Patients with a hypofunctioning nodule would be triaged to surgery. Those with a hyperfunctioning nodule receive a trial of suppression therapy. Interestingly, malignant foci occur in 80% of lesions larger than 4 cm.[5] Therefore, in patients with large lesions, immediate surgical intervention can be considered without additional diagnostic testing.

Response to TSH suppression is not a reliable predictor of malignant potential in a solitary thyroid nodule. Fifteen percent of patients placed on TSH suppression and followed by serial ultrasounds showed continued lesion growth, 20% demonstrated regression, and 65% showed no change.[17] In the case of total resolution with TSH suppression, further treatment, aside from periodic surveillance, is unnecessary.

Thyroid ultrasound is helpful in detecting small nodules or ruling out contralateral disease. Occult nodules less than 1 cm in size detected incidentally on imaging studies (ultrasound, CT) can simply be followed by serial examinations in most cases. To assuage clinical concern about the benign nature of these lesions, an image-guided FNA should be appropriate.

Prognostic Factors and Staging

Over the past decade, prognostic factors for differentiated thyroid cancer were identified using multivariate analysis from large retrospective series. The major prognostic variables can be divided into patient- and tumor-related risk factors. Age is the only patient-related risk factor that significantly affects survival (Table 2). The tumor-related risk factors are size, presence or absence of extrathyroidal extension, and distant metastasis.

A number of patient-risk assignment systems were developed using these prognostic factors (Table 3). The four systems we will discuss are the AMES (Lahey Clinic),[18] AGES-MACIS (Mayo Clinic),[19,20] University of Chicago,[8] and Memorial Sloan-Kettering systems.[21,22] These systems use patient- and tumor-related risk factors (age, distant metastasis, extrathyroidal extension, and tumor size) as prognostic variables. The only exception is the University of Chicago system, which uses all of these variables except tumor size (Table 3). The differences between the various systems are based on the weighted influence these prognostic factors have on the risk of a patient developing local, regional, or distant metastases.

The AMES system assigns patients to the low-risk group if they are men less than 40 years old or women younger than 50 and are free of distant metastasis. Given these criteria of age or gender, patients are further assigned to the low-risk group if their tumor is less than 5 cm, has no extrathyroidal extension, and has a favorable histology (microinvasive follicular car- cinoma). All other patients who do not meet the above criteria are considered to be at high risk.[18]

The AGES (Mayo) system determines risk group assignment by using a weighted scoring system based on the aforementioned prognostic factors.[19] The MACIS system is a modification of the AGES system in which tumor grade has been replaced by distant metastases and margin status.[20]

The Memorial Sloan-Kettering staging system assigns patients to low-, intermediate-, and high-risk groups. Low-risk patients are those who are less than 45 years old and have favorable tumor-related factors (Table 3). Patients in the high-risk group are age 45 or older with adverse tumor-related risk factors. Those assigned to the intermediate-risk group are under age 45 with adverse tumor-related factors or older than 45 with favorable tumor-related factors.[23]

Histologic Subtypes

The impact of the more aggressive histologic subtypes, such as tall cell variant of papillary carcinoma, is not well understood. Most studies do not define histologic subtypes sufficiently to assign a significant predictive value to histologic grade. However, poor differentiation[19] and tall cell variants[24] of papillary carcinoma should be considered unfavorable subtypes.

Multifocality and Cervical Node Metastasis

Despite the high incidence of multifocality in papillary thyroid cancer, the recurrence rate in unresected thyroid tissue is less then 10%.[25,26] Several studies corroborate the finding that multifocality does not carry prognostic significance.[21,27-29]

Similarly, up to 40% of patients with differentiated thyroid carcinoma have microscopic, occult cervical lymph node metastases.[30] The impact on survival of occult or clinically evident cervical metastases is unclear. However, most studies agree that N1 disease does not affect overall survival.[27-29,31] The Lahey group and others[21,32,33] reported that, after univariate analysis, cervical lymph node metastases are a favorable indicator of survival. This particular outcome analysis failed to adjust for patient age, however.

In both the Sloan-Kettering and Lahey series, the favorable impact of nodal metastases on survival disappeared with multivariate analysis.[21,24] In a matched pair study at Sloan-Kettering, all N1 patients had more recurrences and yet similar survival when compared to N0 patients. However, for patients over 45 years of age with nodal metastases, survival was negatively affected.[35] This finding was supported by two other retrospective studies.[36,37]