The Role of Sentinel Lymph Node Biopsy in Differentiated Thyroid Carcinoma

Objective  To determine whether sentinel lymph node (SLN) biopsy can accurately predict central compartment metastasis in patients with differentiated thyroid carcinoma.

Design  Prospective clinical study.

Setting  Academic tertiary care center.

Patients  Ninety-eight patients (82 women and 16 men; mean age, 48.3 years) underwent a total thyroidectomy and central compartment dissection.

Intervention  Peritumoral injection of methylene blue dye, 1%, followed by SLN biopsy.

Main Outcome Measures  The final pathology report established the presence of metastasis among SLNs and lymph nodes that did not stain blue (non-SLNs [NSLNs]).

Results  Differentiated thyroid carcinoma was found in 75 of 98 patients (77%). Seventy of 75 patients with differentiated thyroid carcinoma presented with SLNs and/or NSLNs within the central compartment. Fifteen of 70 patients had metastasis-positive SLNs, while 55 had metastasis-negative SLNs. Six of 15 patients with positive SLNs also had positive NSLNs. No patients with negative SLNs were found to have positive NSLNs. Sentinal lymph node status was a highly significant predictor of NSLN result (Fisher exact test, P < .001). The accuracy, sensitivity, specificity, and positive and negative predictive values of SLN biopsy were 87%, 100%, 86%, 40%, and 100%, respectively.

Conclusions  To our knowledge, this is the largest series of SLN biopsy in patients with differentiated thyroid carcinoma. Our experience suggests that this is an accurate and noninvasive means to identify subclinical lymph node metastasis. Because negative SLNs correlate strongly with a negative central compartment (100% in this study, P < .001), this technique can be used as an intraoperative guide when determining the extent of surgery necessary in cervical level VI.

The appropriate management of occult cervical lymph node metastasis in differentiated thyroid carcinoma (DTC) is controversial.¹ Subclinical lymph node metastasis can be found in 27% to 90% of patients with DTC at the time of surgery or histologic examination.^2-4 Although DTC has a high predilection to spread to regional lymph nodes, the prognostic significance of such regional metastasis is unclear. It has been shown that, in subsets of patients with DTC, particularly those younger than 45 years, lymph node metastasis is not associated with a decrease in survival.^5,6 Other reports, on the contrary, state that lymph node metastasis is a significant prognostic factor that increases the risk of locoregional recurrence^3,7,8 and survival.^9,10 The presence of cervical lymph node metastasis can account for up to 75% of locoregional recurrence cases.¹¹ In such instances, associated neck dissection improves the prognosis of patients with high-risk DTC.¹²

A routine adoption of prophylactic central compartment neck dissection in patients with DTC is not yet accepted as a standard management of occult metastasis given the inherent risks of hypocalcemia and recurrent laryngeal nerve injury.^4,13 The use of the sentinel lymph node (SLN) technique has therefore begun to gain interest as a means to detect occult lymph node metastasis in DTC.^14-24 The SLN is defined as the first lymph node in a regional lymphatic basin that receives lymph flow from a primary tumor. Lymphatic mapping with SLN permits staging of malignant tumors in an effort to avoid complete nodal dissection and its associated morbidity. The goals of this study were to prospectively determine the utility of SLN biopsy to detect occult metastasis and to evaluate the ability of the SLN to predict the nodal status of additional lymph nodes in the central compartment in patients with DTC.

This prospective study involved 98 patients who were treated at 3 teaching hospitals affiliated with McGill University, Montreal, Quebec, Canada. Patients who underwent a total thyroidectomy with ipsilateral central compartment neck dissection based on a preoperative diagnosis, or on findings highly suggestive, of DTC by fine-needle aspiration and clinical presentation were included in the study. Exclusion criteria included medullary and anaplastic thyroid cancer, benign pathologic findings (chronic thyroiditis or Graves disease), completion thyroidectomy, antecedent neck surgery or irradiation, distant metastasis, and concurrent parathyroidectomy.

There were 82 women and 16 men involved in the study. The mean (SD) age of the patients was 48.3 (12.6) years, with a range of 19 to 78 years. The institutional review board of the university approved the study, and written informed consent was obtained from each patient.

At surgery, the thyroid nodule was exposed by lateralization of the strap muscles. Before mobilization of the gland, 0.2 to 0.3 mL of methylene blue dye, 1%, was injected peritumorally into the thyroid parenchyma at the 3, 6, 9, and 12 o’clock positions with a 27-gauge, 1¼-in needle and a tuberculin syringe. In cases of multiple tumor nodules, the predominant nodule was injected. Within seconds, the blue dye was seen spreading within the ispilateral thyroid lobe and via lymphatic channels to the central compartment (Figure 1). Injection required approximately 30 seconds, and 1 more minute was permitted to allow lymphatic drainage. The total time added to the operative procedure was less than 2 minutes.

A complete ipsilateral central compartment neck dissection was achieved in all patients. Lymph nodes that stained blue were defined as SLNs (Figure 2). Central compartment lymph nodes that did not stain blue were defined as non-SLNs (NSLNs). The SLN and NSLN samples were sent for permanent section and stained with hematoxylin-eosin. Data were recorded with respect to lymph node size, number, and anatomical site.

Patients were treated after surgery according to our institution's algorithm for calcium management of postthyroidectomy patients.²⁵ Normal values for serum calcium ranged from 8.48 to 10.48 mg/dL (to convert to millimoles per liter, multiply by 0.25). The serum calcium test was corrected for measured abnormal serum albumin levels.

Patients were considered hypocalcemic if 1 of the following laboratory or clinical conditions was present: a serum-corrected calcium level less than or equal to 7.6 mg/dL or signs and symptoms of hypocalcemia, such as perioral numbness, paresthesias of the upper-extremity digits, or a positive Trousseau sign. Transient hypocalcemia was defined as the presence of hypocalcemia conditions within 1 month of surgery. Permanent hypocalcemia was defined as the presence of persistent hypocalcemia conditions beyond 1 month. Nonmetabolic postoperative information on patients was obtained from follow-up clinical notes.

Descriptive statistics were used to analyze patient and tumor characteristics. Accuracy, sensitivity, specificity, and positive and negative predictive values of the SLN were calculated. The Fisher exact test was used to evaluate for statistically significant associations among SLNs and NSLNs. Intercooled Stata version 8.2 statistical software (StataCorp, College Station, Texas) was used to test for a difference of proportions of postoperative data.

The distribution of the final thyroid gland histologic type is listed in Table 1. Malignant tumors were found in 75 of the 98 patients (77%), with the tumor size ranging from 0.1 to 10.0 cm. Of the 75 malignant tumors, 48 were papillary thyroid carcinomas (PTCs), 2 were follicular carcinomas (FCs), 3 were Hürthle cell carcinomas (HCCs), and 22 were papillary microcarcinomas (PMCs). Benign histologic findings, including follicular adenomas, Hürthle cell adenomas, and adenomatous nodules, were detected in 23 patients (24%).

Overall, SLNs were detected with methylene blue dye in 60 of 75 patients (80%) with thyroid tumors. The rate of SLN detection among PTCs, PMCs, FCs, and HCCs was 79% (38 of 48), 77% (17 of 22), 100% (2 of 2), and 100% (3 of 3), respectively. Collectively, the number of methylene blue dye–stained SLNs ranged from 0 to 10 (mean [SD], 3.2 [2.4]), and the size ranged from 0.1 to 1.8 cm. Lymph node metastasis was found in 15 of 60 SLNs (25%).

Non-SLNs were detected in 59 of 75 patients (79%) with thyroid tumors. There was a presence of NSLNs in 75% (36 of 48), 86% (19 of 22), 100% (2 of 2), and 67% (2 of 3) of patients with PTCs, PMCs, FCs, and HCCs, respectively. The number of NSLNs collectively ranged from 0 to 13 (mean [SD], 2.5 [1.9]), and the lymph node size ranged from 0.1 to 1.7 cm. Metastatic tumor foci were found in 6 of 59 patients (10%) with NSLNs. The SLN and NSLN characteristics of patients with thyroid tumor foci are summarized in Table 2. Among the 23 patients with benign thyroid gland findings, no positive SLNs or NSLNs were identified.

The patients with either positive SLNs or positive NSLNs are described in Table 3. Fourteen patients had papillary thyroid disease, with 12 patients having clinically significant primary tumors measuring more than 1.0 cm (patients 1-12), and 2 patients having microcarcinomas (patients 13 and 14). One patient had primary HCC of the thyroid, with metastatic tumor found in 1 of 10 SLNs (patient 15). Collectively, of the 15 patients with positive SLNs, 6 also had positive NSLNs (patients 2, 4, 8, 9, 11, and 14). There were no cases in which a patient had a positive NSLN without having a positive SLN. In 3 patients, in addition to their cervical level VI metastasis, positive SLNs were found in the lateral aspect of the neck from cervical levels 2 to 5 (patients 4, 8, and 14). Each of these patients subsequently underwent an ipsilateral modified radical neck dissection (cervical levels II-V), and additional positive NSLNs were found. In no patients were there additional metastatic NSLNs in the lateral neck without positive SLNs being found within the same cervical level. There were no postoperative complications due to the use of methylene blue dye for SLN biopsy.

Seventy of 75 patients (93%) with thyroid tumors presented with SLNs and/or NSLNs within the central compartment. Fifteen of 70 patients (21%) had metastasis-positive SLNs, while 55 of the 70 (79%) had metastasis-negative SLNs. Six of the 15 patients (40%) with positive SLNs also presented with positive NSLNs. No patients with negative SLNs were ultimately found to have positive NSLNs (Table 4). Fisher exact test analysis indicated that SLN status was a highly significant predictor of NSLN result (P < .001). There was concordance among SLN findings and NSLN pathologic status in 61 of 70 patients: an accuracy rate of 87% (95% confidence interval [CI], 81.0%-90.4%). The sensitivity, specificity, and positive and negative predictive values of SLN biopsy to predict NSLN status were 100% (95% CI, 64.2%-100%), 86% (95% CI, 74.5%-92.9%), 40% (95% CI, 25.7%-67.1%), and 100% (95% CI, 91.9%-100%), respectively.

Transient hypocalcemia developed after surgery in 10 of the 98 patients (10%) who underwent SLN biopsy (95% CI, 0.05%-0.18%) (Table 5). No patients developed permanent hypocalcemia. One patient had postoperative unilateral vocal cord paralysis as a result of deliberate sectioning of the right recurrent laryngeal nerve, which was intraoperatively found to be encased by tumor. The results of frozen-section biopsy of the area indicated positive PTC with perineural invasion, a finding that was confirmed on the final pathology report. There were no other complications of thyroid surgery, such as a postoperative hematoma or hemorrhage.

The SLN is defined as the first lymph node in a regional lymphatic basin that drains a primary tumor. Conceptually, if lymphatic drainage is to occur in a stepwise fashion, then the SLN should reflect the pathologic status of the remaining lymph node basin. Sentinal lymph node biopsy, moreover, should beneficially detect early, subclinical metastasis. In the last 2 decades, SLN biopsy has gained significant consensus as the standard of management for identifying regional lymphatic spread in melanoma²⁶ and breast cancer.^27,28 The role of SLN biopsy in DTC, however, is not yet fully understood.

In this series, the rate of SLN detection of thyroid tumors with methylene blue dye was 80%. This result is similar to other studies that have reported a rate of SLN identification of 76% to 100% in DTC using a vital dye technique.^14-22 Additional methods of identifying SLNs in DTC have been tried, including intraoperative gamma probe and lymphoscintigraphy.^23,24 Using a radiocolloid for SLN mapping, however, is both more costly and time-consuming and is not readily available in many centers. For these reasons, radionuclide injection was not evaluated in the present study. Hypersensitivity reactions, particularly anaphylaxis, have been reported in 1% to 3% of cases in which isosulfan blue dye was used,^29,30 whereas, to our knowledge, there have been no studies documenting allergic reactions with the use of methylene blue. In breast cancer, no significant differences have been reported in the identification rates of positive SLNs with isosulfan blue and methylene blue,^31,32 and the latter is available at a lower price.^32,33 We elected to use methylene blue dye for SLN biopsy in DTC because of its equivalent efficacy, low cost, ease of use, and limited adverse effects. We found that methylene blue stains lymph nodes efficiently and that there were no complications with its use.

Subclinical nodal disease is found histopathologically in the majority of patients with DTC; however, the management and impact on the prognosis of such lymph node metastasis is unclear.^2-4 Nodal involvement in DTC has been shown in a number of series to be associated with an increased risk of locoregional recurrence but not with overall survival.^1,34 Other reports indicate that nodal dissection in DTC can advantageously decrease locoregional recurrence and improve survival.^7,9,35 Recommendations of the management of adenopathy associated with DTC are quite varied and include blind nodal sampling, “berry picking” of palpable lymph nodes, central compartment neck dissection, and elective modified radical neck dissection. Routine cervical node dissection in thyroid disease, however, necessitates longer surgical time and is associated with increased patient morbidity.^4,13 Accordingly, being able to identify those patients who would benefit from nodal dissection before a more extensive procedure is undertaken would improve DTC management.

To be of clinical utility, the SLN biopsy technique must have a high sensitivity and specificity for detecting the presence of subclinical lymph node metastasis in the central compartment. In this study, 15 of 75 patients (20%) with DTC were found to have positive central compartment lymph node metastasis, and, in each instance, the SLN was correctly identified. There were no cases in which a patient had a positive NSLN without having a positive SLN. Our results confirm that the SLN can predict lymph node metastasis among the NSLNs with a sensitivity and a specificity of 100% and 86%, respectively. We found that the SLN identified in DTC was highly accurate (87%) and also a highly significant predictor (P < .001) of the status of NSLNs in the central compartment. Our findings agree with those of recent studies, albeit with a smaller sample size, that state that SLN biopsy can predict the nodal status of the remaining central compartment lymph nodes with a sensitivity of 71% to 88%, a specificity of 86% to 100%, and an accuracy of 75% to 92%.^19-22 In one report, the sensitivity of SLN biopsy was increased to 100% with the addition of immunohistochemical analysis with cytokeratin 7.¹⁹

Papillary thyroid carcinomas are well known to metastasize to regional lymph nodes and make up the majority of our malignant cases. Shaha et al³⁶ describe the incidence of lymph node metastasis among histologic varieties of DTC as 61%, 30%, and 21% for PTCs, FCs, and HCCs, respectively. The role of additional lymph node resection in FCs and HCCs has been under question.^15,37 We identified SLNs in all 5 cases of FC and HCC, and 1 case involved positive metastasis in the SLN. Excluding cases of FC and HCC from SLN biopsy in this series would have missed regional metastatic disease.

Sentinal lymph node biopsy with methylene blue dye has also been shown to be an accurate method for estimating lymph node status in the lateral aspect of the neck in DTC and to support the need for lateral neck dissection.^2,20,21 We found additional SLNs in the lateral aspect of the neck in 3 cases, and, in each instance, SLN biopsy correctly identified lymph node metastasis. Each of the 3 patients underwent lateral neck dissection, and, as in the central compartment, we found no patients with positive NSLNs in the lateral aspect of the neck without the presence of positive SLNs in the same cervical level. In the central and lateral regions of the neck, SLN biopsy may resultantly prove to guide compartment-oriented neck dissection.

If a technique is to be used as a screening tool for metastasis, it must not miss potential metastatic disease. In our view, the most critical qualitative descriptor for SLN biopsy is the false-negative rate. Early SLN biopsy studies identified the status of the SLN, but not all patients in these reports underwent a central compartment dissection; therefore, the false-negative rate and the predictive utility were unknown.^14,15,18 False-negative rates among recent reports in which a vital dye technique was used range from 7% to 25%, with a negative predictive value of 37% to 89%.^16,17,19-22 In this series, among 75 patients with DTC, we found no false negatives, which resulted in a negative predictive value of 100%. For SLN biopsy to gain acceptance as a suitable technique, it is imperative to have a small number of false-negative SLNs, as we observed.

All patients in this series underwent a complete ipsilateral central compartment neck dissection. We aimed to establish the morbidity if SLNs were identified, and we therefore suggest that a corresponding neck dissection be completed within that cervical compartment. We compared our current results with those of a recent study from our institution that involved 2 (M.P.H. and M.J.B.) of the current 3 (M.P.H., M.J.B., and R.J.P.) thyroid surgeons.²⁵ We found no significant difference in the rate of transient postoperative hypocalcemia with the addition of a central compartment neck dissection to a total thyroidectomy (Table 5). Furthermore, there were no additional postoperative complications as a result of SLN biopsy in this series. Completing a central neck dissection at the time of initial surgery can therefore potentially avoid the higher complication rates that have been reported with reoperation in the central compartment.^38,39 Another advantage of SLN biopsy is that it may help identify patients who are likely to develop a challenging central compartment recurrence at the time of initial operation. The SLN technique may also permit early detection of patients who may benefit from adjuvant iodine 131 ablation.

Our findings have established the feasibility of the SLN biopsy technique in DTC. Frozen-section analysis of the SLN by step-sectioning with immunohistochemical studies is the next step at our institution and is currently under way. To our knowledge, this study is the largest series of SLN biopsy in DTC in the literature. Our findings support the following conclusions:

• Sentinal lymph node biopsy with methylene blue dye is an effective and safe technique to identify subclinical lymph node metastasis in DTC.

• The SLN in DTC is a highly significant predictor of the status of the NSLN in the central compartment (P < .001), displaying high accuracy, sensitivity, specificity, and negative predictive utility.

• Our findings enable us to confidently state that, if the SLN in DTC is negative, then additional NSLNs in the central compartment will also likely be negative.

• Patients without the presence of identifiable SLNs in DTC are unlikely to have subclinical central compartment metastasis.

• Ultimately, this technique may allow a distinct selection of patients who would benefit from central compartment dissection.

Correspondence: Richard J. Payne, MD, MSc, FRCSC, Department of Otolaryngology–Head and Neck Surgery, Jewish General Hospital, Pavilion E, Room E-904, 3755 Chemin de la Cote-Ste-Catherine Rd, Montreal, QC H3T 1E2, Canada (rpayne@jgh.mcgill.ca).

Submitted for Publication: May 3, 2009; final revision received July 28, 2009; accepted August 29, 2009.

Author Contributions: Dr Anand had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Anand, Gologan, Rochon, Tamilia, How, Hier, Black, Trifiro, Tabah, and Payne. Acquisition of data: Anand, Gologan, Rochon, Hier, Black, Richardson, Hakami, Marzouki, and Payne. Analysis and interpretation of data: Anand, Tamilia, Hier, and Payne. Drafting of the manuscript: Anand and Payne. Critical revision of the manuscript for important intellectual content: Anand, Gologan, Rochon, Tamilia, How, Hier, Black, Richardson, Hakami, Marzouki, Trifiro, Tabah, and Payne. Statistical analysis: Anand. Obtained funding: Anand and Payne. Administrative, technical, and material support: Anand, Gologan, Rochon, How, Hier, Black, and Payne. Study supervision: Gologan, Rochon, Tamilia, How, Hier, Black, Trifiro, Tabah, and Payne.

Financial Disclosure: None reported.

Previous Presentation: This study was presented in part at the 2009 Annual Meeting of the American Head and Neck Society; May 30, 2009; Phoenix, Arizona.

Additional Contributions: Christina Holcroft, PhD, Centre for Clinical Epidemiology and Community Studies, Jewish General Hospital, assisted in analyzing the results.