The Role of SPECT/CT Lymphoscintigraphy and Radioguided Sentinel Lymph Node Biopsy in Managing Papillary Thyroid Cancer

Importance  Single-photon emission computed tomography/computed tomography (SPECT/CT) and radioguided sentinel lymph node biopsy (rSLNB) are techniques that could potentially benefit surgeons and pathologists in the identification of sentinel lymph node (SLN) metastases in patients with papillary thyroid carcinoma (PTC). Evidence suggests that these novel techniques lead to substantial changes in PTC management by reducing understaging and of occult lymph node (LN) metastases and optimizing neck surgery by increasing the necessity of lateral lymphadenectomy and decreasing central lymphadenectomy.

Objectives  To correlate the presence of LN metastases in PTC with clinical and pathological features using SPECT/CT and rSLNB.

Design, Setting, and Participants  For this prospective cohort study from June 2010 to November 2013, 42 patients with thyroid nodules suspicious for papillary carcinoma or classified as malignant on cytology examination without suspicion of lymph node metastases by clinical and ultrasound examinations were recruited from a single public medical institution.

Interventions  All 42 patients underwent preoperative lymphoscintigraphy after an ultrasound-guided peritumoral injection of Technetium Tc 99m nanocolloid. Cervical images were acquired with a SPECT/CT scanner 15 minutes after radiotracer injection. Approximately 2 hours after lymphoscintigraphy, the patients were submitted to intraoperative rSLNB using a handheld gamma probe. All SLNs identified were removed alongside with non-SLNs from the same compartment. Papillary thyroid carcinoma, SLNs and non-SLNs were submitted for histopathology and immunohistochemical analyses.

Results  Of the 42 patients initially enrolled, 37 were included in analysis, including 6 men and 31 women with a mean (range) age of 47 (22-83) years. Overall, T stage was as follows: T1, 23 patients (62.2%); T2, 8 patients (21.6%); and T3, 6 patients (16.2%). Sentinel lymph nodes were identified in 92% of the patients, and among these metastases were present in 17 patients (46%). The SLNs were false-negative in 3 patients. Metastases in the lateral compartment ocurred in 7 patients (18%). There was a significant association between LN metastases and tumor size (odds ratio, 1.06; 95% CI, 1.00-1.13; P = .02), with a Cohen d effect of 0.683 (medium to large effect). Overall, 17 patients (46%) with LN metastases had management changed because they were submitted to higher radioiodine ablation doses and closer clinical surveillance.

Conclusions And Relevance  Radioguided SLNB is able to detect occult cervical lymph node metastases in patients with papillary thyroid carcinoma, and in 7 patients (18%) rSLNB detected lymph node metastases in the lateral compartments. The rSLNB technique lead to management change in 14 patients (37.8%).

Papillary thyroid carcinoma (PTC) is the most common type of differentiated thyroid cancer with a high probability of spreading to regional lymph nodes (LNs). Although some studies have shown that the presence of LN metastases does not alter survival,^1-3 other studies have shown that PTC recurrence increases morbidity and mortality and that the presence of cervical LN metastasis at diagnosis is the single most important factor that ultimately increases the risk of local recurrence.^4,5

The American Thyroid Association (ATA) recommends therapeutic central compartment (level VI) neck dissection only for those patients with clinically involved central LNs. Additionally, the ATA recommends prophylactic central compartment neck dissection (ipsilateral or bilateral) in patients with clinically uninvolved central neck LNs (cN0) who have advanced primary tumors (T3 or T4) or clinically involved lateral neck nodes (cN1b), or if the information will be used to plan further steps in therapy. The ATA guidelines do not recommend prophylactic central compartment neck dissection in patients with T1 and T2 tumors that are cN0.

Unfortunately, patients classified cN0 with PTC are categorized as such based on clinical examination and neck ultrasonography. However, clinical examination may present false-negative rates of up to 30%^6-8 because of the incapacity to locate occult cervical LN metastases. Ultrasonography has a higher accuracy in detecting cervical LN metastases from PTC,^9,10 and is thus routinely used for preoperative evaluation.¹¹ Although ultrasonography has a high specificity for detecting cervical LN metastases (95%-97%), the range of sensitivity is wide (52%-84%) mainly because of occult disease.^12,13

Central neck dissection for T1 and T2 tumors could detect microscopic LN metastases and reduce the possibility of locoregional recurrence. Nevertheless, this procedure is not routinely performed because it is associated with significant morbidity and increased time of surgery, anesthesia, and hospitalization.^14-17

Sentinel lymph node biopsy (SLNB) is an alternative to elective LN dissection in cN0 patients, and this concept has been addressed in a few studies in patients with PTC using patent blue^18,19 and radiotracers with dynamic or single-photon emission computed tomography and computed tomography scanner (SPECT/CT) lymphoscintigraphy.^20-23 Garcia-Burillo et al²⁴ performed SPECT/CT lymphoscinthigraphy to evaluate the effect of intraoperative sentinel lymph node (SLN) identification. The authors found that this method detected occult LN metastases and improved surgical management in PTC in patients with cN0 tumors by upstaging a significant number of patients (17 patients [46%] vs 5 patients [12.5%]).

We believe it is important to further evaluate if SPECT/CT lymphoscintigraphy for radioguided SLNB (rSLNB) can change not only the surgical management of PTC, as has been addressed in the literature, but the clinical management of patients with PTC.

The aims of this study were to investigate the added value of SPECT/CT lymphoscintigraphy and rSLNB to detect LN metastases in PTC. We intend to correlate the presence of LN metastases with the clinical and pathological features of the primary tumor and to evaluate the potential change in patient management, whether clinical or surgical.

Box Section Ref ID

The Campinas State University institutional review board reviewed and approved this prospective study. Patients were enrolled between June 2010 and November 2013, and all patients provided written informed consent.

All patients with thyroid nodules were underwent a fine needle aspiration biopsy. Patients with nodules classified Bethesda V (suspicious for PTC) or Bethesda VI (malignant)²⁵ on cytology exam without suspicion of LN metastases (cN0) by clinical and ultrasonography examinations were included.

All patients younger than 18 years who were pregnant; and/or who had undergone prior neck surgery, prior neck radiotherapy, and/or neck lymphadenopathy; and/or who had prior thyroid cancer were excluded from this study.

All patients were submitted to a preoperative cervical ultrasonography prior to thyroidectomy. All neck compartments were explored (central and lateral) in search of cervical LN metastases. The criteria used to identify metastatic LNs were alteration of morphology from oval to round shape; hyperechogenicity in relation to the adjacent muscles; intranodal cystic necrosis; calcifications and absence of an echogenic hilum.²⁶

Once the ultrasonography classified the patient as cN0, the patient was included in the study. All patients were submitted to a peritumoral injection of 7.4 MBq (0.2 mCi) of Technetium Tc 99mphytate in a 1 mL saline solution using a 23-gauge needle guided by ultrasonography prior to lymphoscintigraphy. A single injection was performed adjacent to the nodule in unifocal lesions. Two injections were performed in multifocal lesions regardless of the number of nodes in each lobe; 1 injection per lobe adjacent to the largest nodule. Only 2 injections were performed in multifocal lesions to avoid excessive radioactivity within the thyroid bed that could eventually affect the detection of adjacent radioactive SLNs.

After the injection, pressure was applied over the injection site for 10 minutes to avoid hematoma formation. The entire procedure lasted approximately 20 to 30 minutes.

All patients were submitted to SPECT/CT lymphoscintigraphy after the peritumoral injection of ^99mTc-phytate prior to total thyroidectomy. Images of the cervical region were acquired in all patients 15 minutes after the injection and approximately 2 hours prior to surgery using a SPECT/CT dual-head scintillation camera with low-energy, high-resolution collimators. The level, size, and depth (Figure 1) of the LNs detected by SPECT/CT images were recorded. Skin marks were drawn in the projection of the SLN prior to the surgery.

Intraoperative SLN localization was performed using a handheld gamma probe that was placed into a sterile surgical wrap and, before the incision, used to slowly scan from the injection site laterally to the jugular regions to examine any radioactivity marked on the patient’s skin.

The SLN detection was conducted after total thyroidectomy to avoid interference from the radioactivity of the primary tumor. The central neck compartment was scanned bilaterally with the handheld gamma probe for hot spot investigation. All radioactive nodes (regardless of counts), as well as nonradioactive nodes in the central compartment, were removed. Intraoperative scanning of the lateral compartments was also performed, especially in patients in whom the preoperative lymphatic mapping revealed SLNs located in these compartments. Data was recorded with respect to LN number and anatomical site. It is important to underline that at least one non-SNL was removed from the same level for each SLN biopsied.

The final assessment of all excised specimens (thyroid, sentinel lymph node, and nonsentinel lymph nodes) was performed with hematoxylin-eosin staining. Lymph nodes that were negative for metastases using hematoxylin-eosin staining were submitted to immunohistochemical staining using an anticytokeratin antibody and thyroglobulin.

All patients underwent iodine-131 whole body scans after total thyroidectomy and rSLNB. Patients received radioactive iodine ablation treatment (RAI) when necessary. The standard RAI dose was 3700 MBq (100 mCi) for patients with uptake solely in the thyroid bed. The RAI dose was increased to 5550 MBq (150 mCi) in patients with LN metastases.

The first RAI was performed within 2 to 3 months after thyroidectomy. Patients were required to withdrawal from levothyroxine for 4 weeks and undergo a low iodine diet for 2 weeks prior to RAI. Radioactive iodine ablation treatment was only performed in patients with thyroid-stimulating hormone levels above 30 mU/L.

The PTC were grouped into 2 tumor types for analysis: aggressive (solid, tall, and oxyphilic cells) and nonaggressive (classic form and follicular variant). Furthermore, correlations between the presence of LN metastasis and patient characteristics (age and sex), serum stimulated thyroglobulin levels, and tumor characteristics (size, multifocal lesions, angiolymphatic invasion, extrathyroidal extension, surgical margins, and histologic variant) were examined. Statistical analysis was first done including all patients (Table) and second with only patients T1 and T2 from TNM thyroid stage. A result was considered false negative when the perioperative analysis was negative (SLN not detected) but the definitive histological result was positive for PTC metastasis in non-SLNs.

Univariable analyses for associations between occult LN metastases and the patient clinical characteristics and the pathologic features of the primary lesion were performed using the Mann-Whitney U test, χ² test, t test and Fisher Exact test, and P values less than .05 were considered significant. All statistical analyses were performed using SPSS version 17 software (SPSS Inc).

Forty-two patients were prospectively enrolled, and 5 patients were excluded because final histopathology examination demonstrated that in 4 cases the nodules were benign and 1 case had uncertain features of malignacy. Thirty-seven patients (6 men; 31 women) with a mean (range) age of 47 (22-83) years were eligible for evaluation (Table).

Overall, SPECT/CT lymphoscintigraphy identified at least 1 SLN in 34 of 37 patients (92%). Among these 34 patients, 1 radioactive SLN was removed during surgery in 35% (n = 12), 2 SLNs in 38% (n = 13), and 3 or more SLNs in 27% of the patients (n = 9). The remaining 3 patients in which SPECT/CT was negative (did not detect SLN) a biopsy of LNs located in level VI was undertaken.

During surgery, a total of 72 SLNs were removed (mean per patient, 1.9). Among these 72 SLNs, 13 (18.0%) were located in level II, 11 (15.3%) in level III, 7 (9.7%) in level IV, 4 (5.6%) in level V, 36 (50%) in level VI, and 1 (1.4%) in level VII.

Among the 37 patients studied, histopathology revealed metastases in at least 1 SLN in 14 patients. On the other hand, histopathology revealed metastases in non-SLNs in an additional 3 patients (8.1%), which were classified as false-negative. Overall, metastases in cervical SLNs were detected in 14 of 37 patients (37.8%), resulting in a sensitivity of 82.3% (14 of 17) for the rSLNB technique.

Two of 3 false-negative cases presented tumors with extrathyroidal extension and positive surgical margins. No extrathyroidal extension finding was classified as extensive but only as minimal, which is invasion into immediate perithyroidal soft tissues or sternothyroid muscle typically detected only microscopically (T3 tumors). Thus CT or magnetic resonance imaging would not be able to detect this minimal invasion.

The primary tumor histology of the patients with metastatic SLNs and non-SLNs showed the following characteristics: 4 tumors (23.5%) were of aggressive variant; 6 tumors (35%) had positive surgical margins; 5 tumors (29%) had extrathyroid extensions; and 5 tumors (29%) presented with angiolymphatic invasion. Interestingly, metastases in the lateral compartments of the neck (all ipsilateral to the primary tumor) occurred in 3 of 17 (18%) of the patients with tumors in the lateral compartments of the neck. One patient with a 3.5-cm multifocal PTC, aggressive variant, extrathyroid extension, positive surgical margins, and angiolymphatic invasion presented a metastatic SLN in level V. Another patient with a 1.8-cm multifocal PTC and positive surgical margins presented a level III metastasis and metastatic SLN in level VI as well.

There was a significant association between the presence of LN metastases and tumor size (Figure 2), with a 6% increased odds of presenting LN metastases for each 1-mm increase in tumor size (odds ratio, 1.06; 95% CI, 1.00-1.13; P = .02).

All SLNs that were negative for metastases using hematoxylin-eosin staining were systematically examined by immunohistochemical staining using an anticytokeratin antibody and thyroglobulin that was able to detect metastases in 15 SLNs (21.4%). Frozen sections were not used to examine SLNs so that there was no influence on intraoperative decisions.

During follow-up, after total thyroidectomy, there were no relevant complications, such as hypocalcemia or recurrent laryngeal nerve paralysis, associated with the procedure. Three patients did not undergo RAI because their whole body scans were negative (no iodine-131 uptake in the thyroid bed in cervical LNs or distant sites). Seventeen patients without cervical LN metastases received the standard 3700 mBq (100mCi) of RAI. However, the RAI dose was increased to 5550 MBq (150mCi) in the remaining 17 patients (46%) that presented with occult LN metastases after SLNB.

To our knowledge, this study is the first to correlate the presence of SLN metastases in PTC to clinical and pathological features. In this cohort, LN metastases occurred in 46% of our patient population. This finding is in accordance with a former meta-analysis,²⁷ that found that among 843 patients submitted to SLNB (although mainly with blue-dye technique and without SPECT/CT), SLNB could potentially avoid central neck dissection in 57% of patients because it found 43% of metastatic LNs.²⁷ Furthermore, we noted a 35% probability of metastatic LNs in the central compartment, which is similar to the published rates (33%-47%)^28,29 of LN metastases in patients with clinically node-negative PTC undergoing prophylactic central node dissection.

Our results suggest that occult metastases are unobserved and underestimated on initial staging owing to the low detection rate by ultrasonography and the lack of proper intraoperative LN dissection, especially in the lateral compartments. To that effect, cervical nodal metastasis in PTC may be more frequent than expected. As these are slow-growing tumors, there is the possibility of an incorrect down-staging at initial diagnosis. Thus, the presence of future locoregional recurrence in these patients may actually be the initial occurrence of micro-metastases and not recurrence. Furthermore, prior studies have shown a 15%²⁷ occurrence of metastases solely in the lateral compartments of the neck; in our study, we found 18% of metastases in the lateral compartments, and therefore these patients would most likely be under-staged if submitted to a standard surgical procedure. Recurrence rates have been shown to be significantly higher in patients with positive LNs in the lateral compartment vs patients with LN metastasis in the central compartment (60% vs 30%; P = .007). Disease-free survival and mean time to recurrence have also been shown to be significantly shorter (30 vs 52 months; P = .035 and 7 vs 44 months; P = .004, respectively).³⁰

In our study, the management was changed in 46% of the patients due to the detection of LN metastases. Although the role of prophylactic central neck LN dissection (CNLD) in the treatment of PTC is controversial, many surgeons perform routine prophylactic CNLD. The presence of positive LNs are associated with higher postoperative RAI to decrease the risk of recurrence in patients undergoing CNLD.²⁸ Previous studies³¹ have shown that postoperative ablation of thyroid bed remnants activities range from 1.11 to 3.7 GBq (30-100 mCi) depending on the radioiodine uptake measurement and amount of residual functioning tissue present. On the other hand, activities between 5.55 and 7.40 GBq (150-200 mCi) are administered for treatment of PTC with cervical or mediastinal LN metastases.³¹

Sentinel lymph node biopsy has become the most accurate investigation tool for proper staging of many cancer types, mainly with the use of SPECT/CT lymphoscintigraphy prior to surgery. In our study, the technique reduced the surgical time and improved surgical planning and incision based on the prior identification of LNs, especially in the lateral neck compartment. SPECT/CT lymphoscintigraphy and rSLNB is a procedure that is feasible and easy to perform, although slightly more costly and time-consuming, as well as not readily available in many centers. Perhaps for these reasons rSLNB studies have not been extensively published in comparison to SLNB with the blue-dye technique in PTC.

On the other hand, the power of our study was to inject the radiotracer preoperatively. This strategy eliminates the potential disruption of lymphatic channels during the initial dissection that normally occurs with SLNB using the blue-dye technique. Additionally, rSLNB has the advantage of detecting SLNs (potentially metastatic) outside the central neck compartment, properly selecting patients who would benefit from complete neck dissection and/or optimized RAI.

The ATA guidelines suggest that thyroidectomy without prophylactic central neck dissection may be is appropriate for small (T1 or T2), noninvasive, clinically node-negative PTC (cN0) and for most follicular cancers.¹⁵ In our series, among the 17 metastatic SLN and non-SLN, 7 were classified T1, 5 as T2, and an additional 5 as T3. Therefore, if these patients (T1 and T2) were submitted to a surgical procedure following the ATA guidelines, 12 of 17 (70%) of the metastatic LNs would be undetected.

Prophylactic central neck dissection remains a controversial topic in those patients without evidence of LN involvement. Some studies^32,33 have shown that it does not change survival but does improve locoregional control. Moreover, many authors^4,34,35 have shown that compartment-oriented LN dissection decreases locoregional recurrence rates and improves overall survival. On the other hand, extensive nodal dissection increases morbidity, hospitalization, costs, and is associated with higher complications rates, including permanent hypoparathyroidism and vocal cord paresis, compared with total thyroidectomy alone. Therefore, rSLNB is currently an interesting option to perform a more accurate selection of patients that would benefit from a more extensive procedure.

There was a significant association between the presence of LN metastases and tumor size (Figure 2). Additionally, a higher probability of SLN metastases occurred in younger patients with larger tumors. Prior reports have shown that the incidence of metastases increases with tumor size (38% in tumors <1 cm vs 79% in tumors >4 cm)³⁶ and recent studies^37,38 have shown that larger tumors (especially >1 cm) are more likely to have ipsilateral and contralateral metastases.

One weakness of our study was that it was not possible to measure the true false-negative rate for SLN because central neck dissection was not routinely performed in all patients. Our institutional review board deemed the performance of radical cervical neck dissection in these patients unethical based on international guidelines.^15,39 The option was to remove LNs from the same compartment where the rSLNB was detected.

Using this strategy, we found that the false-negative rate was 3 of 37 patients (8.1%), which was lower than previously reported (16%).²⁷ The 3 metastatic non-SLNs were located in level VI, and all patients presented multifocal tumors. False-negative cases may have been owing to the following reasons: the number of injections, the location of the injection site, tumor-blocked lymphatic channels, and inability to detect through lymphoscintigraphy. Many injections may pose a difficulty as the radioactivity adjacent to the nodule may mask LN uptake. Peritumoral injections might cause spillage of radioactive colloid into the surrounding tissue, rendering the detection of rSLN in the central neck compartment a challenging task. However, this was not a major concern of our surgical team because the thyroid is removed intraoperatively first hand. Tumor-blocked lymphatic channels may cause false-negative nodes because of the inability to take up the tracer due to tumor invasion.

Fischer et al⁴⁰ found that the use of ancillary techniques, including immunohistochemistry and molecular analysis, can significantly improve diagnosis. Tumors with follicular architecture that lack the unequivocal features of malignancy may pose difficulties in the distinction of benign and malignant conditions. However, a single antibody is usually suboptimal in terms of sensitivity and specificity to diagnose metastases. Panels with 2 or more antibodies are usually more effective and improve diagnostic accuracy in fine-needle aspirates and paraffin-embedded tissue. In our study, the additional immunohitochemical staining was able to detect the micrometastases in 15 SLNs (21.4%).

The strength of this study is that it is a prospective cohort with the potential introduction into clinical thyroid practice of the rSLNB technique that involved only highly trained specialists in the field and submitted all patients to the same procedures. Our study proposed a better selection of patients who would benefit from compartment-oriented nodal dissection, thus reducing unnecessary surgery and morbidity and maintaining accurate locoregional LN staging. Furthermore, patients with metastatic disease could be better selected allowing optimized RAI and avoidance of RAI in patients with low-risk tumors and negative SLNs. The use of SPECT/CT provided clearer images, better planning for operations, and identification of LN metastases outside the central compartment. The main limitation related to the rSLNB is the possibility that LN metastases can be found in non-SLNs. The study is limitated by the small sample size and that reduces the analysis power and enlarges confidence intervals. Another concern is regarding the external validity that might be diminished because the study was conducted at a single institution with a high-volume of thyroid cancer surgery and could make replicating this technique difficult for lower-volume institutions. To evaluate the true false-negative rate of SLNB, the recurrence rate will be analyzed in the future.

SPECT/CT lymphoscintigraphy and rSLNB detects occult neck LN metastases in patients with PTC. We observed that 37.8% of the patients would have been under-staged if only submitted to clinical evaluation and ultrasonography for primary staging, especially in the 18% of patients that had metastases in the lateral compartments. Sentinel lymph node biopsy avoided central compartment lymphadenectomy in 44.6% of patients. The size of primary tumors appears to be a variable to guide the necessity and extent of prophylactic central neck dissection because the possibility of a metastatic SLN is increased in larger tumors. Radioguided sentinel lymph node biopsy led to management change in 37.8% of the patients. According to ATA guidelines, if these patients had not been diagnosed with metastases they would have been considered low-risk and thus received less RAI and less agressive clinical surveillance. Our findings may help guide future patient staging and surveillance. Further prospective multicenter studies for longer periods and with a larger number of SLNBs are needed to corroborate these findings.

Corresponding Author: Raquel Novas Cabrera, MD, Division of Nuclear Medicine of the Department of Radiology, University of Campinas (UNICAMP), Rua Vital Brasil, 251, Campinas, Brasil, CEP 13083-888 (raquelcabrera@yahoo.com.br).

Accepted for Publication: May 5, 2016.

Published Online: June 30, 2016. doi:10.1001/jamaoto.2016.1227.

Author Contributions: Drs Cabrera and Etchebehere had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Cabrera, Chone, Crespo, Etchebehere.

Acquisition, analysis, or interpretation of data: Cabrera, Zantut-Wittmann, Matos, Ferreira, Pereira, Ribeiro, Santos, Ramos, Etchebehere.

Drafting of the manuscript: Cabrera, Ribeiro, Etchebehere.

Critical revision of the manuscript for important intellectual content: Cabrera, Chone, Zantut-Wittmann, Matos, Ferreira, Pereira, Santos, Ramos, Crespo, Etchebehere.

Statistical analysis: Cabrera.

Administrative, technical, or material support: Cabrera, Pereira, Ribeiro, Ramos, Crespo, Etchebehere.

Study supervision: Cabrera, Chone, Zantut-Wittmann, Matos, Ferreira, Etchebehere.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Additional Contributions: We would like to thank the Clesnan Mendes-Rodrigues, PhD, for his valuable statistical contribution.