Staging

Accurate staging is important in determining the prognosis and tailoring the treatment of patients with differentiated thyroid cancer. Unlike with many other tumor types, the presence of distant metastasis, for example in lungs and bones, does not obviate primary resection (thyroidectomy) because metastatic disease may respond to radioactive iodine therapy (RAI) after surgical removal of neck disease. Surgery comprises removal of all thyroid tissue along with the primary tumor, as well as that of regional nodal disease, and is one of the most important initial treatments. Complete resection of the thyroid gland and locoregional disease is particularly important for facilitating RAI for metastatic disease. Furthermore, patients having 5 or more clinically apparent metastases, a metastasis greater than 3 cm, or extranodal tumor extension were found to have a more adverse prognosis than those having none of these features. Therefore, it is important to assess the extent of local disease and regional lymph node involvement before surgery.

Imaging: Ultrasonography

Preoperative ultrasonography is the most important imaging modality in the evaluation of thyroid nodules and thyroid cancer. The ATA Surgery Working Group guidelines recommend ultrasonography of the lateral neck to assess for metastatic nodes when thyroid cancer is diagnosed. Ultrasonography identifies suspicious cervical adenopathy in the setting of thyroid malignancy. The sensitivity of detecting metastatic nodes that may alter overall management ranges from 20% to 31%. Sonographic features suggestive of metastatic lymph nodes are:

• •
  

  Cystic change

  
  
• •
  

  Calcifications

  
  
• •
  

  Loss of the fatty hilus

  
  
• •
  

  A rounded rather than oval shape

  
  
• •
  

  Hypoechogenicity

  
  
• •
  

  Increased vascularity

Of these, detection of loss of the fatty hilus is 100% sensitive, but has very low specificity (29%). The only criterion with high sensitivity as well as relatively high specificity is peripheral vascularity (86% sensitivity, 82% specificity). All other potential criteria have sensitivity of less than 60%, and are thus inadequate for use as a single criterion for the identification of malignancy.

In a series of 3874 patients, Ito and colleagues investigated the diagnostic accuracy of ultrasonography for lateral node metastasis in patients who underwent therapeutic or prophylactic modified neck dissection, reporting a specificity of 95% and a sensitivity of 43%. The presence of certain features, while low in sensitivity, can be highly specific for metastasis. For example, in a patient with known papillary thyroid cancer, the presence of cystic areas or punctate microcalcifications in a node are virtually diagnostic of metastasis (100% specificity). A lymph node short axis of less than 5 mm is also highly specific for metastasis (96%). Thus, an ultrasound scan can potentially alter the surgical approach in as many as 20% of patients.

Imaging: Computed Tomography, Magnetic Resonance Imaging, Positron Emission Tomography

The limitations of ultrasonography are that evaluation is uniquely operator dependent, and it cannot easily visualize retropharyngeal, deep paraesophageal, or mediastinal nodes. Therefore, alternative imaging procedures, such as computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET), may be preferable in some clinical settings. However, the sensitivities of these studies for the detection of cervical lymph node metastases are all relatively low (30%–40%). Of the 3 modalities, CT with contrast is probably the most sensitive. The disadvantage of obtaining a contrast CT is that the iodine load necessitates a delay in any planned subsequent RAI, often beyond 3 months postoperatively. Nevertheless, these alternative imaging modalities are necessary for assessing the presence of deep cervical nodes that ultrasonography cannot detect, as well as the extent of invasive tumors, such as with invasion of the trachea, involvement of esophagus, and encasement or invasion of major vessels.

Laryngoscopy, Esophagoscopy, and Tracheoscopy

In addition, laryngoscopy, esophagoscopy, and tracheoscopy may also be necessary in the assessment of large, rapidly growing, retrosternal, or invasive tumors to determine the possibility of involvement of extrathyroidal tissues.

Biopsy

When a suspicious lymph node is identified, malignancy should be confirmed by ultrasound-guided fine-needle aspiration (FNA) if it will change the extent of the operative procedure. If the node is very small or cystic, it may be difficult to attain a sufficient sample for cytologic analysis, in which case the needle washout from the aspirate can also be sent for a thyroglobulin level. Elevated thyroglobulin from an FNA washout, even in a noncellular or hypocellular nondiagnostic aspirate, can confirm metastasis. This FNA measurement of thyroglobulin is valid even in patients with circulating thyroglobulin autoantibodies.

Staging

Accurate staging is important in determining the prognosis and tailoring the treatment of patients with differentiated thyroid cancer. Unlike with many other tumor types, the presence of distant metastasis, for example in lungs and bones, does not obviate primary resection (thyroidectomy) because metastatic disease may respond to radioactive iodine therapy (RAI) after surgical removal of neck disease. Surgery comprises removal of all thyroid tissue along with the primary tumor, as well as that of regional nodal disease, and is one of the most important initial treatments. Complete resection of the thyroid gland and locoregional disease is particularly important for facilitating RAI for metastatic disease. Furthermore, patients having 5 or more clinically apparent metastases, a metastasis greater than 3 cm, or extranodal tumor extension were found to have a more adverse prognosis than those having none of these features. Therefore, it is important to assess the extent of local disease and regional lymph node involvement before surgery.

Imaging: Ultrasonography

Preoperative ultrasonography is the most important imaging modality in the evaluation of thyroid nodules and thyroid cancer. The ATA Surgery Working Group guidelines recommend ultrasonography of the lateral neck to assess for metastatic nodes when thyroid cancer is diagnosed. Ultrasonography identifies suspicious cervical adenopathy in the setting of thyroid malignancy. The sensitivity of detecting metastatic nodes that may alter overall management ranges from 20% to 31%. Sonographic features suggestive of metastatic lymph nodes are:

• •
  

  Cystic change

  
  
• •
  

  Calcifications

  
  
• •
  

  Loss of the fatty hilus

  
  
• •
  

  A rounded rather than oval shape

  
  
• •
  

  Hypoechogenicity

  
  
• •
  

  Increased vascularity

Of these, detection of loss of the fatty hilus is 100% sensitive, but has very low specificity (29%). The only criterion with high sensitivity as well as relatively high specificity is peripheral vascularity (86% sensitivity, 82% specificity). All other potential criteria have sensitivity of less than 60%, and are thus inadequate for use as a single criterion for the identification of malignancy.

In a series of 3874 patients, Ito and colleagues investigated the diagnostic accuracy of ultrasonography for lateral node metastasis in patients who underwent therapeutic or prophylactic modified neck dissection, reporting a specificity of 95% and a sensitivity of 43%. The presence of certain features, while low in sensitivity, can be highly specific for metastasis. For example, in a patient with known papillary thyroid cancer, the presence of cystic areas or punctate microcalcifications in a node are virtually diagnostic of metastasis (100% specificity). A lymph node short axis of less than 5 mm is also highly specific for metastasis (96%). Thus, an ultrasound scan can potentially alter the surgical approach in as many as 20% of patients.

Imaging: Computed Tomography, Magnetic Resonance Imaging, Positron Emission Tomography

The limitations of ultrasonography are that evaluation is uniquely operator dependent, and it cannot easily visualize retropharyngeal, deep paraesophageal, or mediastinal nodes. Therefore, alternative imaging procedures, such as computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET), may be preferable in some clinical settings. However, the sensitivities of these studies for the detection of cervical lymph node metastases are all relatively low (30%–40%). Of the 3 modalities, CT with contrast is probably the most sensitive. The disadvantage of obtaining a contrast CT is that the iodine load necessitates a delay in any planned subsequent RAI, often beyond 3 months postoperatively. Nevertheless, these alternative imaging modalities are necessary for assessing the presence of deep cervical nodes that ultrasonography cannot detect, as well as the extent of invasive tumors, such as with invasion of the trachea, involvement of esophagus, and encasement or invasion of major vessels.

Laryngoscopy, Esophagoscopy, and Tracheoscopy

In addition, laryngoscopy, esophagoscopy, and tracheoscopy may also be necessary in the assessment of large, rapidly growing, retrosternal, or invasive tumors to determine the possibility of involvement of extrathyroidal tissues.

Biopsy

When a suspicious lymph node is identified, malignancy should be confirmed by ultrasound-guided fine-needle aspiration (FNA) if it will change the extent of the operative procedure. If the node is very small or cystic, it may be difficult to attain a sufficient sample for cytologic analysis, in which case the needle washout from the aspirate can also be sent for a thyroglobulin level. Elevated thyroglobulin from an FNA washout, even in a noncellular or hypocellular nondiagnostic aspirate, can confirm metastasis. This FNA measurement of thyroglobulin is valid even in patients with circulating thyroglobulin autoantibodies.

Extent of Resection

Once the diagnosis of well-differentiated thyroid cancer has been established, the primary treatment modality is surgical resection. The current ATA guidelines recommend near-total or total thyroidectomy for all tumor sizes greater than 1 cm. Regarding less extensive surgery, studies have shown increased recurrence and decreased survival rates in the patient population who undergo subtotal thyroidectomy.

One argument for total thyroidectomy over lobectomy is the risk of thyroid cancer in the contralateral lobe. Several studies have demonstrated that of patients who originally underwent hemithyroidectomy and then subsequent completion thyroidectomy, the incidence of papillary thyroid cancer in the remaining lobe was found to be between 35% and 55%. Even more compelling is evidence that patients with larger tumors who undergo lobectomy rather than total thyroidectomy fare worse in terms of recurrence and survival. A retrospective study by Hay and colleagues that examined 2444 cases over 60 years showed that the recurrence and death rate was higher during the initial decade, in which the majority of surgical resections were lobectomies rather than total thyroidectomies, implying that lobectomy alone may be insufficient. Furthermore, an even larger study by Bilimoria and colleagues in 2007 of 52,173 patients in the National Cancer database demonstrated higher recurrence rates and lower survival rates in those patients with tumors larger than 1 cm who underwent lobectomy only. This finding was maintained even in the subset of patients whose tumors ranged in size from 1 to 2 cm.

Microcarcinoma Versus Macrocarcinoma

For tumors less than 1 cm in size at diagnosis, however, many studies have indicated that lobectomy alone may have equivalent outcomes to more extensive resection, and may therefore be the initial surgery of choice. In another study, Hay’s group reexamined nearly the same patient population over a 60-year period, focusing on those with tumors less than 1 cm in size, and found no significant difference in recurrence and survival rates between those undergoing lobectomy versus total thyroidectomy. Even the Bilimoria study did not find a difference in recurrence or survival rates in those patients undergoing either lobectomy or total thyroidectomy with tumors smaller than 1 cm.

A more recent study by Ogilvie and colleagues, however, raises the question of whether the “less than 1 cm” category of microcarcinoma should be further subdivided. These investigators retrospectively reviewed 130 records of node-negative patients with tumors smaller than 1 cm and found that when subdivided into a 6- to 10-mm group and a less than 6-mm group, a larger tumor size was still significantly associated with higher rates of adverse pathologic features and central node positivity. Nevertheless, there remains a lack of evidence to show whether lobectomy only in the 6- to 10-mm group results in any increase in recurrence or mortality rates in comparison with total thyroidectomy.

Tumor Involvement of the Recurrent Laryngeal Nerve

Invasive well-differentiated thyroid carcinoma is uncommon, but does occur in about 16% of patients. Thyroid surgeons strive to avoid damage to the recurrent laryngeal nerves during routine operations; however, in cases where local disease is found to involve the nerve, there is some question as to how aggressively one should resect. The recurrent laryngeal nerve has been found to be involved in 33% to 61% of invasive thyroid cancers and is the most commonly involved nerve in the central compartment. Recurrent laryngeal nerve involvement, however, may not predict the same morbidity or mortality as invasion of other structures. Chan and colleagues reported high long-term survival rates in patients with known preoperative vocal cord paralysis and complete nerve transection caused by gross involvement. McCaffrey and colleagues noted that invasion of the recurrent laryngeal nerve did not independently affect survival rates, in contrast to invasion of other structures, such as the trachea and esophagus, which independently decreased survival. McCaffrey’s group argues for incomplete excision and treatment of residual disease with adjuvant RAI, as data have not shown complete excision to provide a survival benefit. The development and use of adjunctive RAI has played a large role in allowing more surgical discretion regarding the extent of resection and preservation of vocal cord function despite nerve involvement.

In a previous Otolaryngologic Clinics review on the management of invasive thyroid cancer, Urken noted that several factors should play a role in the determination of surgical extent:

• •
  

  Preoperative preexisting vocal cord paralysis as documented by laryngoscopic examination

  
  
• •
  

  Involvement of the contralateral recurrent laryngeal nerve

  
  
• •
  

  Tumor histology

  
  
• •
  

  Likelihood of disease response to RAI

In most cases, all attempts at nerve preservation should be performed with resection of obvious gross tumor. If the tumor can be dissected off of the nerve, the nerve can be left intact. However, if the tumor encases the nerve and resection is impossible without nerve sacrifice, especially in cases of known ipsilateral preoperative paralysis, the nerve should be resected. In cases of known contralateral preoperative paralysis, however, the potential morbidity from ipsilateral sacrifice and resulting bilateral vocal cord paralysis with possible need for tracheostomy may reasonably cause one to decide against full nerve resection. Also of note are several studies that have shown the possibility of recovery of vocal cord function after tumor resection in cases where nerve involvement was in the form of compression rather than infiltration. Chiang and colleagues have shown that extensive dissection of the recurrent laryngeal nerve can be performed with relatively low rates of temporary nerve palsy, and in their particular study, no cases of permanent palsy. Complete resection, however, as shown by Nishida and colleagues, results in a high rate of permanent paralysis, even despite attempts at reanastomosis.

If there is need for recurrent laryngeal nerve sacrifice, rehabilitation procedures such as immediate reinnervation can be performed if sufficient distal stump remains for anastomosis. If direct reanastomosis is not possible because of a large gap, anastomosis of ansa cervicalis or ansa hypoglossi to the distal stump can be performed. Nerve grafting using a portion of the ansa cervicalis is another option. If accidental transection of the recurrent laryngeal nerve occurs during any operation, immediate repair is recommended to preserve muscle tone to laryngeal muscles, which can facilitate improvement in voice rehabilitation therapy.

Central Neck Dissection

The 2009 ATA guidelines provide recommendations regarding central neck dissection as an adjunct operation for thyroid cancer. These guidelines define central neck dissection as “at a minimum… consist[ing of] removal of the prelaryngeal, pretracheal, and paratracheal lymph nodes…[either] unilateral or bilateral.” The central neck compartment consists of level VI, and occasionally level VII, as defined by the Memorial Sloan-Kettering system. The superior boundary is the hyoid bone and the lateral boundaries, the carotid arteries. The definition for the inferior border is somewhat variable between the sternal notch or the innominate artery. Level VI is the main zone of lymphatic drainage for the majority of thyroid cancers. Those involving the upper pole, pyramidal lobe, and isthmus may also drain to levels II and III of the lateral neck. The lateral portion of the hemithyroid lobe may drain toward lateral neck levels III and IV.

The ATA guidelines recommend that therapeutic central-compartment dissection of level VI be performed for all patients with known clinical involvement of either the central or lateral neck compartments. However, in patients without evidence of nodal disease, the choice of whether to do an elective central neck dissection at the time of initial cancer resection is controversial. The ATA guidelines give only a C recommendation for elective central neck dissection in “patients with papillary thyroid carcinoma with clinically uninvolved central neck lymph nodes, especially for advanced primary tumors.” Likewise, they give a C rating for “thyroidectomy without prophylactic central neck dissection…[in patients with] small (T1–T2), noninvasive, clinically node-negative papillary thyroid carcinoma and most follicular cancer,” in recognition that the current evidence is not strong either for or against the procedure, with only expert opinion to guide which groups of patients might benefit the most.

Central Neck Dissection in the Setting of Known Nodal Involvement

The central compartment is the most likely region of nodal recurrence. Within the central compartment lymphatic drainage tends to flow to the ipsilateral side, and there is a higher rate of positive nodes found in the ipsilateral central compartment versus the contralateral side or the lateral neck.

It is rare to find patients with clinical evidence of lateral involvement without central involvement, but skip metastases can occur. Lateral node metastasis has been shown to have an independent correlation predicting central metastasis. Roh and colleagues performed a review of 22 patients who presented with lateral neck recurrences, none of whom had had previous central neck dissections with their initial thyroidectomies. At reoperation, elective central neck dissection showed 86% of patients to have central metastases, most frequently in the ipsilateral paratracheal compartment.

A study by Leboulleux and colleagues regarding prognostic factors for persistent or recurrent disease in patients with locally advanced thyroid cancer at the time of diagnosis found that the presence of metastasis, specifically to the central compartment, significantly increased the risk of persistent disease after RAI.

Lymph node positivity has been shown in many studies to affect the rate of recurrence in differentiated thyroid cancer.

• •
  

  Harwood and colleagues compared the records of well-differentiated thyroid cancer patients, half of whom were N0 at time of diagnosis, and found that there were significantly more recurrences in patients with nodal involvement than in those without.

  
  
• •
  

  Mazzaferri and Jhiang looked at 1355 patients over 40 years and discovered the presence of cervical node metastases to significantly correlate with higher recurrence rates at 30 years, regardless of cancer subtype.

  
  
• •
  

  The long-term study by Hay and colleagues on patients with papillary thyroid microcarcinoma also found that recurrence rates were higher in node-positive patients, with more than 80% of all recurrences localizing to regional neck nodes.

Furthermore, several studies have also indicated lymph node positivity to have an effect on survival. In a large study of 9904 patients in the Surveillance, Epidemiology, and End Results (SEER) database, 77% of whom were node negative on presentation, cervical lymph node metastasis had a risk ratio of 1.34 on overall survival rates and was statistically significant. This study refuted a previous study in 2003 that used the same database and did not find positive cervical nodes to affect mortality.

Yet there are other studies showing that cervical node metastases have an independently significant effect.

• •
  

  Harwood’s group also found that when matching for age, nodal metastasis also resulted in a worse survival prognosis, especially in older populations.

  
  
• •
  

  Another study by Scheumann and colleagues confirmed a decrease in survival for node-positive patients, even when controlling for age, tumor invasion, and distant metastasis.

  
  
• •
  

  In addition, a more recent study by Lundgren and colleagues performed a large population-based study of 5123 patients that showed cervical metastases to have an odds ratio of 2.5 on mortality, even after adjusting for TNM stage.

  
  
• •
  

  Finally, a 2010 study by Grant and colleagues examined 420 patients who underwent thyroidectomy with or without neck dissection based on the 4 recommendations set out by the 2009 ATA guidelines for management of thyroid cancer, and found only a 5% nodal recurrence rate in this group.

The evidence also supports the recommendation for compartment-oriented central compartment dissection of the involved compartments rather than “berry picking” when metastatic disease is identified in the central compartment either radiographically or intraoperatively. A study by Musacchio and colleagues demonstrated increased local recurrence rates in patients whose neck disease was managed only through berry picking of visualized metastases rather than thorough neck dissections.

Role of Prophylactic or Elective Central Neck Dissection

Lateral Neck Dissection

Metastasis to the lateral neck compartments warrants dissection of the involved compartments, given that positive lymph nodes, as noted previously, increase the risk of recurrence and decrease survival rates. The 2009 ATA guidelines recommend that “therapeutic lateral neck compartmental lymph node dissection should be performed for patients with biopsy-proven metastatic lateral cervical lymphadenopathy.” As already noted, lateral neck involvement without central neck involvement is rare, although skip metastases can occur. Chung and colleagues found a 7.7% rate of skip metastases in 7.7% of patients with papillary microcarcinoma who had undergone both central and lateral neck dissections for preoperative evidence of lateral metastasis.

Risk factors for lateral compartment involvement have been found to include :

• •
  

  Younger patient age

  
  
• •
  

  Gender

  
  
• •
  

  Tumor multifocality

  
  
• •
  

  Tumor calcifications

  
  
• •
  

  Upper pole tumor location

  
  
• •
  

  Larger tumor size

  
  
• •
  

  Extrathyroidal extension

  
  
• •
  

  Central node positivity

  
  
• •
  

  Ipsilateral involvement of other lateral neck levels

  
  
• •
  

  Contralateral lateral involvement

Some investigators recommend more aggressive treatment of the lateral compartment if any positive lymph nodes are found in the central compartment.

• •
  

  Goropoulous and colleagues looked at 39 patients who underwent central and bilateral lateral neck dissections in addition to total thyroidectomy, and found that of the patients who were found to have positivity in the central neck, 80% also had concurrent ipsilateral lateral neck disease and 52% were also positive in the contralateral lateral neck. However, this study included patients with a variety of tumor sizes, with a trend toward lateral positivity with larger tumors.

  
  
• •
  

  Machens and colleagues conducted a retrospective review of patients who underwent both primary and reoperative neck dissections for both papillary and medullary thyroid cancer, and found that the ipsilateral lateral compartment was involved almost as frequently as the central compartment for all groups.

  
  
• •
  

  By contrast, however, a study by Wang and colleagues found that only 3% of N0 necks were found to have lateral disease, and that risk factors for lateral recurrence and decreased survival included older patients and tumor size greater than 4 cm.

Extent of Lateral Neck Dissection

There is also some controversy over whether performing lateral neck dissection requires full clearance of levels II through V, or whether selective dissection may be performed based on clinical data. The most common neck levels involved with lateral metastases are II, III, and IV, with level III having the highest probability of positive nodes in most studies. Ahmadi and colleagues, however, found level IV to be the most commonly involved level in their recent study of 49 patients, with level IV also being the most common site of recurrence.

Some argue that the surgeon’s judgment, based on both preoperative clinical data and intraoperative findings, may be sufficient to determine the extent of lateral dissection. Caron and colleagues also found a low rate of recurrence to levels I and V (3% for both levels combined) in a population that had had previous resection rates of 3.9% for level I and 18.6% for level V. Nevertheless, several studies have shown that multiple-level involvement is common in the lateral neck. There is also concern that preoperative assessment may not be very sensitive at finding involved lymph nodes.

• •
  

  Wu and colleagues showed that of 100 patients who had at least 1 positive lateral neck node, 77% had involvement of multiple lateral neck levels and that the sensitivity of preoperative ultrasonography for the lateral neck was only in the 40% to 60% range per level.

  
  
• •
  

  Kupferman and colleagues also found preoperative ultrasonography to be only 20% sensitive for level V involvement.

Many investigators, however, have advocated thorough lateral neck dissection for any known lateral involvement.

• •
  

  Kupferman and colleagues found a relatively high rate of involvement of level V metastases, of 21%. This study, however, involved a variety of patient and tumor characteristics.

  
  
• •
  

  Farrag and colleagues also found a high rate of level V metastases, with 40% of patients who underwent lateral neck dissection for clinically proven involvement of at least one lateral level. This study further differentiated within level V, finding that all of the level V metastases were in V-B, with 0% involvement of V-A. Differentiation within level II showed a 60% overall involvement of level II, with only an 8.5% involvement of II-B; however, all positive level II-B were also positive in II-A.

  
  
• •
  

  Ahmadi and colleagues noted equal rates of level V involvement for both primary resections and lateral neck dissection for recurrence.

  
  
• •
  

  In another retrospective review of recurrence in the lateral neck after initial lateral neck dissection, one group found a relatively high rate of level II recurrence, of 19% to 21%, whether or not the patient had undergone previous dissection of that level.