Controversies in the Surgical Management of Newly Diagnosed and Recurrent/Residual Thyroid Cancer

Introduction

T he incidence of thyroid cancer in the United States is rising, up nearly 2.4-fold since 1973 (1). Virtually the entire increase in cases is due to an increase in the diagnosis of small papillary thyroid cancers (PTCs). Despite the increasing incidence, the number of patients dying of thyroid cancer each year has remained stable (1). Most patients with a diagnosis of thyroid cancer do very well, and less than 10% of the patients will ultimately die of their disease (2).

The mainstay for the treatment of thyroid cancer is surgery, with adjuvant therapy consisting of radioactive iodine and thyroid hormone suppression. Each of these modalities provides benefit to the patient and is associated with their own set of risks. Our goal as physicians caring for patients with thyroid cancer is to find the right balance between these modalities, to provide the best and safest care for our patients. Careful consideration of treatment morbidity plays an especially important role in the management of patients with thyroid cancer, as few die of their disease, and patients may have to deal with the complications of their disease and/or treatment for many years.

Surgical Management of Newly Diagnosed Thyroid Cancer

The goal of the initial surgical treatment of thyroid cancer is to surgically remove all evident disease, minimize treatment and disease-related morbidity, and minimize the risk of recurrence and metastatic spread. A secondary goal of surgery is to permit accurate staging, facilitate postoperative treatment with radioactive iodine (if indicated), and permit accurate long-term surveillance (3).

One of the challenges in determining the optimal surgical approach for patients with thyroid cancer is that patients with thyroid cancer live long lives and rarely die of their disease. Recurrences, however, can be a problem for these patients, and because of the indolent nature of their disease, recurrences may occur up to 20 years after an initial diagnosis. Because of these issues, the ultimate effects of changes in our management may not be clinically evident for 10–20 years. An additional complicating feature for surgical clinical trials is that permanent complications after thyroid surgery are rare. Because of these challenges, we will likely never have a randomized prospective clinical trial to answer many of these questions. Therefore, we must rely on retrospective reviews and prospective studies, using surrogate markers, to determine the best treatment option for our patients.

Extent of Thyroidectomy

The debate over the extent of thyroidectomy (lobectomy vs. total thyroidectomy) for thyroid cancer has been going on for decades. For a variety of reasons, the tide has shifted, and most people are now in agreement that a total thyroidectomy is preferred over a thyroid lobectomy for the majority of patients with a diagnosis of PTC (4). The rationales for total thyroidectomy are several. First is the frequent multifocal nature of the disease. Between 30% and 80% of the patients with PTC develop or have multifocal disease. Therefore, treatment that only removes half of the thyroid leaves tumor in half of the patients (5,6). Second, to reduce the risk of recurrence, both contralateral lobes and also involved lymph nodes must be removed to reduce the need for additional operations. Third, total thyroidectomy instead of lobectomy will facilitate radioactive iodine treatment and the use of serum thyroglobulin (Tg) levels and whole-body radioiodine scans during follow-up. Some also advocate total thyroidectomy as a means to improve disease-specific survival and prevent progression to anaplastic thyroid cancer in unresected residual tumor.

Thyroid cancer recurrence rates after a thyroid lobectomy are higher than after total thyroidectomy, but whether survival is improved is less clear. In 1977, Mazzeferri (7) published a series of 576 patients with PTC and noted the recurrence rate after thyroid lobectomy was twice as high as with total thyroidectomy (18.4% vs. 7.1%). This difference persisted during the 10 years of follow-up (19.2% vs. 10.9%) (8). The 6-year follow-up showed a significant survival advantage in the total thyroidectomy group (1.5% vs. 0.4%), but the differences were less and not significant at 10 years of follow-up (1.5% vs. 0.6%). The role of total thyroidectomy for low-risk disease was evaluated by Hay et al. (9) in 1998. In 1685 patients with low-risk papillary thyroid, they found even higher differences in local recurrences between those who had lobectomy (14%) and those who had total thyroidectomy (2%). Despite the large numbers of patients, however, they were still unable to demonstrate a survival advantage for a total thyroidectomy. Perhaps the largest series of patients relating to this issue was published in 2007 by Bilimoria and colleagues (10). They examined the National Cancer Database and found 52,173 patients with PTC. In this vast dataset they found small but significant advantages for total thyroidectomy compared with thyroid lobectomy for both recurrence (9.8% vs. 7.7%) and survival (98.4% vs. 97.1%). Notably, however, significant differences in either parameter were not found for tumors less than 1 cm. Similarly, in a Mayo Clinic series of 900 patients with the subset of papillary thyroid microcarcinoma, the 20-year recurrence rate was only 6%, and neither bilobar resection nor postoperative radioiodine remnant ablation decreased the recurrence rate (11).

So there are certainly data to support total thyroidectomy to reduce the recurrence of PTC and perhaps to improve survival, but the advantages appear to be small and must be carefully weighed against the risks of the added intervention that total thyroidectomy entails. The complications of thyroidectomy are well known and consistent of transient or permanent hypocalcemia, transient or permanent recurrent laryngeal injury, and neck hematoma. The rate of permanent complications is well established and, in experienced hands, should be less than 2–3%. Probably more important than the extent of surgery, however, is the experience of the surgeon. Although very low complication rates after total thyroidectomy have been published by many very experienced surgeons, those results should not be extrapolated to the majority of surgeons doing thyroid surgery who do less than 20 cases per year (12). Sosa et al. (13) published the first data regarding complications of thyroid surgery as it relates to the surgeon's case volume. She found that high-volume surgeons (defined as doing >100 cases/5 years) had 75% fewer complications than low-volume surgeons (4.3% vs. 16.1%). This was confirmed in another study by Stavrakis in 2007 (12) that showed the complication rate of thyroidectomy, when done by experienced surgeons (>50 cases/year), was 1.6%, but the complication rate rose to 4.3% if the surgeon did less than 10 cases per year and to 7.2% if the surgeon did less than 5 cases per year.

In summary, there are theoretical and established reasons why a total thyroidectomy should be the standard treatment for most forms of PTC. Although the advantages of total thyroidectomy over lobectomy are small in terms of recurrence and survival, the added risk of total thyroidectomy is also small when performed by experienced surgeons. Therefore, total thyroidectomy should be the standard for all patients with preoperatively diagnosed PTC. If, however, a PTC is found incidentally and is less than 1 cm, and the patient has no other risk factors such as a history of head and neck irradiation or a family history of thyroid cancer, thyroid lobectomy should be adequate.

Management of Lymph Node Disease

Lymph node involvement is common in PTC. The incidence of lymph node involvement depends on how it is defined. Palpable nodal disease is present in about 5–10% of the patients, but ultrasound detects lymph node disease in up to 30% of the patients (14 –17). When lymph nodes are removed routinely, standard histologic examination will reveal lymph node involvement in 20–50% of the patients, but when a more detailed inspection is performed, up to 90% of the patients will be found to have microscopic disease (18 –20). There is a large discrepancy between the number of lymph nodes that contain some cancer pathologically and the number that eventually become clinically apparent. This discrepancy makes it difficult to determine the best approach to lymph nodes in the neck.

Historically it was thought that lymph node involvement increases local recurrence, but does not affect survival (8,21). Therefore, surgeons in the United States have had a conservative approach to lymph node disease. More recent data have questioned the practices of the past and led many surgeons to be much more proactive in the management of lymph node disease, especially where it concerns prophylactic lymph node dissection (LND) (3,22 –24).

When discussing surgery for lymph node disease, it is important to define terminology. An elective or prophylactic LND involves removing lymph nodes that are normal on physical exam, imaging, and intraoperative assessment. The rationale to remove these nodes is that they may contain microscopic disease and their removal may prevent recurrence or additional treatment with radioactive iodine. A therapeutic LND involves removing lymph nodes that are clinically or pathologically abnormal based on exam, imaging, or intraoperative assessment. The goal of removing these lymph nodes is to aid in local control and prevent recurrences as well as to potentially improve survival. The standard for LND for thyroid cancer is a compartment-mediated dissection. This involves removing all of the fibrofatty tissue and nodes en bloc from a given compartment, while preserving the critical structures when possible. For thyroid cancer, two types of nodal dissection are frequently performed, a central neck dissection and a lateral neck dissection.

When considering surgically removing lymph node disease, it is important to ask two questions: What are the benefits for the patient—does it decrease recurrences and does it improve survival? And secondly what are the risks of removing the nodes? The presence of lymph node disease does appear to be associated with a higher risk of recurrence. In a study by Wada et al. (25), patients with pathologic-positive lymph nodes had a recurrence rate of 16.3% compared with 0% in patients without lymph nodes pathologically. Of note, most of the disease recurrences occurred in patients with palpable nodes at presentation (24% vs. 7.6% with no palpable nodes). In another study by Ito and colleagues (26), they examined the impact of lateral nodal disease on recurrences. They found that lymph node disease was associated with recurrence, but it was most significant in patients with lymph node disease recognized preoperatively. If lymph nodes were not seen preoperatively, then the risk of nodal recurrence was only 1.5%. Interestingly, in this study of 590 patients, all whom had papillary microcarcinomas, 40% were found to have lateral neck lymph node disease with routine prophylactic resection. This is a 40% incidence of lymph node disease, in a patient population that most would agree is at very low risk for developing recurrent disease. This study and others suggest that lymph node involvement does affect recurrences, but that all lymph node disease does not behave the same and there appears to be a difference between pathologic and clinically significant nodal involvement (15,27).

Whether or not lymph node involvement affects survival has been looked at in several studies, but the answer is still not clear (23,25,28 –32). Eustatia-Rutten et al. (30) studied 52 patients who died because of thyroid cancer. Lymph node involvement had a relative risk of 2.01 on univariate analysis, but on multivariate analysis was no longer a significant predictor of mortality. Another article by Podnos et al. (31) used the Surveillance Epidemiology and End Results database to address this question. They noted that the 10-year survival for patients with thyroid cancer was 79% in lymph node–positive patients and 82% in lymph node–negative patients, a significant but small difference. Probably the most cited reference supporting the concept that lymph node disease negatively impacts survival was published by Lundgren et al. (23) in 2006. This was a nested case–control study of 595 patients who died of thyroid cancer. Predictors of death included distant metastases (odds ratio [OR] 6.6), incomplete resection (OR 4.2), and lymph node metastases (OR 1.9). Lymph node disease was predictive of decreased survival in their multivariate analysis. However, they did not control for incomplete tumor resection that occurred in 51% of the patients who died as compared with 23% of the patients who did not die. Without controlling for this, it is difficult to know if the study truly evaluates the impact of lymph node disease on survival.

In summary, lymph node involvement does appear to increase recurrence rates, but microscopic involvement does not predict as great an increase in risk of recurrence as macroscopic involvement (25,33). The impact of lymph node disease on survival remains unclear, but its overall impact appears to be small.

Role of Prophylactic Central Neck Dissection

The role of prophylactic central neck dissection (PCND) at the time of thyroidectomy for well-differentiated thyroid cancer is a controversial and hotly debated topic, not only in the literature (34) but also at national meetings. In 2006, the American Thyroid Association published their initial guidelines for the management of patients with thyroid nodules and differentiated thyroid cancer (3). In these guidelines they indicated that “routine central neck dissection should be considered for patients with PTC.” They based this recommendation on the notion that this might improve survival and reduce the risk for nodal recurrence and could be achieved with low morbidity. This aspect of the initial guidelines (3) engendered much debate. Recently, the American Thyroid Association reconvened their panel to update and revise their original guidelines and they have been modified to state that a “prophylactic central neck dissection may be performed, especially in patients with advanced primary tumors” and that a “total thyroidectomy without prophylactic central neck dissection may be appropriate for small (T1 or T2), non-invasive, clinically node negative patients” (4). Unfortunately, the literature is still fairly inconclusive on this subject.

The central neck lymph nodes, or level 6 lymph nodes, include the paratracheal, perithryoidal, and precricoid lymph nodes. These are the nodes that run along and often behind the recurrent laryngeal nerve and frequently surround the lower parathyroid gland. These lymph nodes are believed to be the first site of spread from thyroid cancer (20). These nodes are macroscopically involved in about 10% of the cases. However, when these nodes are removed routinely, microscopic central neck lymph node involvement is present in 32–69% of the patients (20,27,35). If these nodes are macroscopically involved, most would agree that a formal central neck dissection should be performed to prevent local recurrence. The debate centers upon whether or not these nodes should be removed prophylactically when there is no evidence of pathologic involvement on physical exam, imaging, or intraoperative assessment.

The argument for PCND focuses on the inability to fully evaluate these nodes preoperatively with ultrasound, the high frequency of pathologic involvement when they are removed, the risks of having to reoperate in the central neck if the nodes become involved in the future, and the current use of radioactive iodine to ablate microscopic disease that may have its own risks. Those who favor PCND consider that the safest and easiest time to remove potentially involved nodes is at the initial operation, and the goal is to prevent local recurrence and reoperation in the central neck. Those who do not favor PCND consider that if nodes are not macroscopically involved they are usually clinically insignificant, being unlikely to grow or become problematic. Moreover, they believe that if patients do develop a central neck recurrence, this can be removed safely by experienced surgeons. Lastly they acknowledge that the risks of PCND are low in highly experienced hands, but that they would probably be much higher if PCND becomes widespread and therefore likely to be performed by those with less experience.

The main argument in favor of PCND is that it should prevent central neck recurrences. To consider this there is a need to establish the chances of central neck recurrence in a patient without evidence of nodal disease on preoperative ultrasound or intraoperative assessment. We are of the opinion that most central neck recurrences occur in patients who have significant nodal disease at their first operation, and even with a formal neck dissection, developed a recurrence in the central neck. In patients who require a therapeutic LND for central neck lymph node disease, the prevalence of central neck recurrence ranges from 7% to 21% (25,36). In a study of patients with papillary thyroid microcarcinoma, Wada et al. (27) found the recurrence rate in patients who underwent therapeutic LND to be 21%, a very high value. In contrast, patients who had PCND, and therefore by definition no overt nodal disease, did have microscopic nodal involvement in 61% but a recurrence rate of only 0.43%. Importantly, however, those patients without overt nodal disease who did not have PCND also had a very low recurrence rate of 0.65% despite the likelihood that they had a high prevalence, perhaps as much as 60%, of microscopic lymph node cancer. Similar findings were reported by Gemsenjager et al. (36) in 2003. This study showed a central neck recurrence rate of 7.1% after a therapeutic LND and a recurrence rate of 0% after PCND with microscopically positive nodes. Again, these results for PCND appear very positive, but in this study those who were candidates for PCND, but did not have this procedure, also had a central neck recurrence rate of 0%. Other studies support the concept that microscopically positive lymph nodes, whether they are removed or not, do not appear to progress to recurrence (36 –39), at least within the time frames of the studies. In summary, recurrences rates are high in PTC patients with macroscopic but not with microscopic lymph node involvement. This is emphasized again in the study of Bardet et al. (37) in which macroscopic lymph node involvement was associated with a 10–30% incidence of recurrence, whereas those without had a much lower recurrence, regardless of whether they had PCND.

There are other features besides macroscopic lymph node involvement that are associated with recurrence of PTC and/or a worse overall prognosis. These include extrathyroidal extension, older age, and more aggressive subtypes (40). Although it is has yet to be proven that PCND decreases the odds of recurrence in these patients, it seems logical that they may be more likely to benefit from PCND, particularly if they segregated into a group that was less likely to respond to radioactive iodine. BRAF mutations have also been shown in several studies to be associated with a higher risk of nodal involvement and nodal recurrence (41 –43). New technologies are now available at some institutions to test an FNA (surgery for thyroid cancer fine needle aspiration) sample for BRAF status, which then allows a surgeon to know preoperatively which patients may be at higher risk for recurrence and therefore benefit from a more aggressive surgical approach.

The risks of central neck dissection are the same as a thyroidectomy: recurrent laryngeal nerve injury, hypoparathyroidism, and hematoma. Although many argue that the risks of a central neck dissection are no higher than a total thyroidectomy, there is debate about this. To perform a central neck dissection, the recurrent laryngeal nerve is dissected along its length, potentially exposing the patient to a higher risk of either transient or permanent hoarseness. Despite the theoretical higher risk, most reports looking at recurrent laryngeal nerve dysfunction or injury after central neck dissection have shown that the incidence of permanent nerve injury remains acceptably low at 1–3% (44). Transient and permanent hypoparathyroidism does appear to be higher when a central neck dissection is performed (45). Most surgeons agree that to do a complete central neck dissection, it is very difficult to leave the lower parathyroid glands in situ and some advocate routine autotransplantation (46). Studies of transient hypoparathyroidism report this complication in 14–60% of the patients who have central neck dissection with rates for permanent hypoparathyroidism ranging from 0% to 16% (45,47,48) (Table 1). Although some of these reports showed very low rates of complications, this may reflect ascertainment bias.

In summary, central neck nodal involvement is common in patients with PTC. Clinically evident nodal disease does impact recurrence and potentially survival and should be resected. PCND may reduce an already low recurrence rate or reduce the need for radioactive iodine, but is associated with risks, namely of hypoparathyroidism. There may be a subset of patients at higher risk for recurrence who may benefit from PCND. These include those with locally advanced tumors, tumors that do not respond to radioactive iodine, and those with a BRAF mutation. More work needs to be done to define the putative population that would benefit from PCND. Until more convincing evidence is available, we cannot recommend that PCND be performed on a routine basis for PTC.

Extent of Resection for Lateral Neck Lymph Node Disease

A preoperative ultrasound looking for metastatic disease is essential in all patients with a preoperative diagnosis of thyroid cancer (3). If abnormal lateral neck lymph nodes are identified on ultrasound or physical exam, then they should be evaluated with FNA, and a lateral compartment LND should be performed at the initial operation (4). Most authors agree that at the present time there is no role for a prophylactic lateral neck dissection for the management of PTC. A compartment-oriented dissection is the standard of care and there is no role for berry-picking of lymph nodes.

For patients who are candidates for LND, there is considerable debate as to how extensive this procedure should be (49 –53). Most patients with PTC have disease limited to levels 3 and 4. Is it necessary, therefore, to remove levels 1, 2, and 5 routinely or can they be removed selectively? Caron et al. (50) studied recurrence rates of patients treated with a selective neck dissection that consisted of routine removal of levels 3 and 4 and selective removal of 1, 2, and 5 based on preoperative imaging. In their study they made several important observations. First, they found that only 3% of the patients had a recurrence in either level 1 or 5. Further, level 1 nodes were never positive unless there was significant level 2 disease. Second, they noted that when they performed a selective removal of only level 3 and 4 lymph nodes, 21% of the patients ultimately required a level 2 neck dissection, and all of those patients had positive level 3 lymph nodes. Lastly, even when they performed a level 2 dissection, 19% of the patients developed recurrence in level 2. Both Lee et al. (52) and Farrag et al. (49) have found that most positive level 2 nodes are level 2A nodes and without significant involvement, level 2B dissection is unnecessary in most patients.

Ultrasound is essential to identify the location of abnormal lymph nodes and should be done preoperatively to help determine the extent of resection required. As long as a compartment-mediated dissection of the involved levels is performed, complete dissection of levels 2–5 is not always essential for patients with lateral lymph node involvement.

Challenges of Reoperative Neck Dissection

Up to a third of patients with PTC will develop recurrent disease after thyroidectomy (54 –56). Postoperative follow-up for patients with thyroid cancer has become much more rigorous so that, increasingly, recurrent thyroid cancer is being identified when it is macroscopic, but quite small. The question for surgeons has become, when do we intervene? What are the benefits of resecting metastatic disease, and what are the risks?

Surgical removal of recurrent lymph node disease can successfully prevent further recurrence in 62% of the patients (57) and can convert 41% of the patients to an undetectable stimulated Tg postoperatively (58). Overall, reoperation in the central neck is associated with more complications, but when performed by an experienced surgeon, the risks are not significantly different from similar surgery performed at the initial operation (58 –60). Therefore, when there is a biopsy-proven recurrence that is >1 cm in size, we and most surgeons advocate surgical excision of the lesion. When the recurrence occurs in a field that has not had a formal compartment-oriented dissection, then a formal dissection should be performed. If the recurrence occurs after a formal neck dissection, then the focused removal of the involved lymph nodes is appropriate.

The first step when considering reoperation for the possibility of nodal disease is to confirm the presence of a metastatic lesion. This is generally done by either FNA cytology or Tg positivity within the lymph node (61). Although there are features on ultrasound that are suggestive of lymph node involvement (microcalcifications, cystic appearance), no feature is 100% diagnostic (62). When lymph nodes are detected on positron emission tomography (PET) scans, it is essential to prove the presence of malignancy with a biopsy as PET has a low specificity for thyroid cancer, and false-positive PET scans are common (63 –65).

Distortion of the tissue planes due to scar tissue can make surgery very technically demanding and can increase the morbidity for the patient and decrease the success of the operation. The key to minimizing the morbidity of a reoperation is careful preoperative and intraoperative localization.

Intraoperatively several adjuncts have been described to facilitate the localization of disease. Preincision ultrasound is probably the most commonly employed adjunct and can guide incision placement. Although preincision ultrasound can give an approximate location for the disease, it frequently is not helpful once an incision is made and scar tissue is encountered. Intraoperative ultrasound has been described and can be helpful especially when operating in a radiated field or a paratracheal location (66,67). Needle localization has been attempted by some, but because of the superficial nature of the lesions, it is difficult to prevent the needle from dislodging during the dissection (68,69). More recently, several techniques have described high success using ultrasound preincision to inject either blue due (70) or a charcoal suspension (71) to mark the pathologic nodes, so that pathologic nodes can be found among the neck scar tissue.

In summary, neck recurrences are common in PTC. Surgical removal decreases the need for additional therapy and can minimize complications from local recurrence. Therefore, any patient with a local recurrence should be offered surgical excision. Since reoperations in the central neck can be challenging, it is essential that a surgeon have accurate preoperative imaging and extensive experience in thyroid surgery.

Conclusions

The risks of thyroidectomy and nodal dissection must be carefully weighed against the potential benefits of therapy. Ultrasound and Tg testing are very sensitive and detect very low levels of disease, and the clinical significance of this disease is still not known. Clinically evident lymph node disease does appear to be associated with an increased risk of recurrence and may also adversely affect survival. Therefore, in addition to a total thyroidectomy, a careful preoperative evaluation for lymph node disease should be performed and if found, it should be resected. The benefits of prophylactic neck dissection, in the absence of macroscopic involvement, have not been proven, and we do not recommend performing a routine central neck dissection. In all cases, to minimize morbidity, patients should only be operated on by surgeons with significant experience in thyroid surgery.