Introduction

Active surveillance of small low-risk papillary thyroid cancers (PTCs), previously an isolated management approach popularized by two centers in Japan (1,2), is now endorsed by the American Thyroid Association (ATA) guidelines (3) and validated in multiple publications (4–10). While initially restricted to papillary microcarcinomas (<1 cm in maximal dimension), current observational management programs will consider active surveillance in tumors as large as 1.5 cm (5,8–10). Successful implementation of active surveillance requires careful consideration of three major domains: tumor/imaging characteristics, patient characteristics, and medical team characteristics (11–14). Thoughtful integration of the variables contained within each of these domains can be used to classify patients with PTC as either ideal, appropriate, or inappropriate for an active surveillance monitoring program.

Primary tumors located in a subcapsular position immediately adjacent to the trachea or a recurrent laryngeal nerve (RLN) are considered to be inappropriate for active surveillance (12,13). While several ultrasonographic features are highly suggestive of tumor invasion into or through the thyroid capsule, this imaging is imprecise. Indeed, it is difficult to rule out minor extrathyroidal extension when a primary tumor is in contact with the thyroid capsule (15,16). The 8th edition of the American Joint Committee on Cancer/tumor-node-metastasis (AJCC/TNM) Cancer Staging System included RLN invasion as a major prognostic factor (17), but it is controversial whether isolated gross invasion of an RLN is associated with disease-specific mortality (18–20).

However, even if it is not a major independent predictor of disease-specific mortality, gross RLN invasion with corresponding irreversible vocal cord dysfunction is an undesirable outcome in active surveillance, and primary tumors in a subcapsular location immediately adjacent to the expected course of an RLN should be excluded from this management approach.

While the location of the trachea relative to the intact thyroid gland is readily apparent on imaging studies, neither ultrasonography nor other cross-sectional imaging can reliably identify the entire course of the RLN as it passes deep and medial to the thyroid gland to enter the larynx. Therefore, an understanding of the anatomy of the RLNs is required to identify specific thyroid subcapsular locations likely to be associated with the greatest risk of direct nerve invasion (Fig. 1). The left RLN arises from the left vagus nerve, loops around the aortic arch and ligamentum arteriosum, ascends parallel to the left tracheoesophageal groove, passes the thyroid gland in the space bounded by the medial thyroid capsule and the trachea, and then proceeds just lateral to the ligament of Berry to enter the larynx.

The right RLN arises from the right vagus nerve, loops around the right subclavian artery more laterally, ascends immediately behind the right lower pole of the thyroid gland, moves from the lateral aspect to the medial aspect of the thyroid gland, and then proceeds superiorly along the trachea and adjacent to the medial aspect of the thyroid gland to pass the ligament of Berry to enter the larynx. The left RLN is, therefore, at risk of local invasion from a primary tumor located anywhere along the medial aspect of the left lobe of the thyroid gland and along the tracheoesophageal grove. The right RLN is at risk of invasion from primary tumors located either along the medial aspect of the right lobe of the thyroid gland (adjacent to the trachea) or in the posterior aspect of the right lower pole of the thyroid gland in areas not immediately adjacent to the trachea.

To better understand the subcapsular location within the thyroid gland and the minimum tumor size that was associated with invasion of the RLNs from small PTC tumors, we retrospectively identified 30 patients with preoperative and/or intraoperative evidence for gross invasion of an RLN from a primary tumor ≤2 cm in maximal diameter. A better understanding of the anatomic locations associated with gross invasion of the RLN will allow clinicians to identify specific subcapsular tumor locations within the thyroid gland that should exclude patients from an active surveillance monitoring program.

Materials and Methods

After obtaining Institutional Review Board (IRB) approval for retrospective data review, we used the Memorial Sloan Kettering Cancer Center (MSK) disease management team surgical database to identify 123 patients with PTC who were treated at MSK between 1986 and 2015, and had gross extrathyroidal extension of tumor involving either the right or left RLN identified either preoperatively or intraoperatively. During this time period, 6259 patients had initial thyroid surgery at MSK, including 4334 cases of PTC tumors ≤2 cm. Of the 123 cases with gross extrathyroidal extension, 39 (31.7%) had a maximum tumor diameter of ≤2 cm. Four patients were excluded because the RLN was invaded by cervical lymph node metastases rather than by the primary tumor, and five were excluded because presurgery images were not available for review.

Since this study focused on the single active surveillance eligibility criteria of extracapsular extension, we did not exclude patients based on preoperative evidence of gross extrathyroidal extension or lymph node metastasis even though these other criteria would have excluded them from an observational management approach. Therefore, 30 patients with direct invasion of an RLN from a primary tumor ≤2 cm in diameter were included in this analysis. For this retrospective review, the IRB deemed waiver of consent to participate is ethically acceptable. Research-related procedures adhered to the IRB-approved protocol and all ethical requirements, including the Declaration of Helsinki.

Electronic medical records were reviewed to assess for clinical or radiological evidence of RLN invasion before surgery. Operative reports were reviewed to determine the location of gross invasion of the RLN, which was then correlated with the results of clinical presurgical neck ultrasound and/or computed tomography (CT) neck scans.

Data are presented as mean ± standard deviation, and medians and ranges are reported when appropriate. All statistical analyses were performed using IBM SPSS Statistics Version 24.0 for Windows (IBM Corp., Armonk, NY).

Results

Gross invasion of an RLN by tumors ≤2 cm is a rare event seen in only 0.8% (35/4334) of patients with PTC who underwent their initial thyroid surgery at MSK between 1986 and 2015. This study included 30 patients with well-differentiated PTC (≤2 cm diameter), as well as preoperative imaging available for review, whose primary tumor demonstrated gross invasion of an RLN at the time of surgery. The median age at diagnosis was 55.1 years (range 15–76 years), median tumor diameter was 1.6 cm (range 0.9–2.0 cm), and 76.7% of patients were female (Table 1).

Most patients underwent a total thyroidectomy (93.3%), and all patients had evidence during surgery of gross extrathyroidal extension to a variety of locations, including the RLN (100%), trachea (50%), esophagus (40%), fibroadipose tissue (13%), strap muscles (10%), and larynx (10%). Half of the patients (50%) had classical variant PTC, 30% had tall-cell variant PTC (defined as >30% tall cells with height twice their width), and 6.7% had follicular variant PTC, while 13.3% had no specific description other than PTC.

Tumor imaging before surgery was performed using ultrasound in 17 patients (56.7%) and using CT with contrast scan in 13 patients (43.3%). Imaging studies performed before surgery reported possible or probable gross extrathyroidal extension into an RLN in only 8 patients (26.7%). Direct visualization of the vocal cords by direct laryngoscopy or mirror examination before surgery was documented in 28 patients (93.3%), of whom 9 (32%) had a unilateral cord paresis/paralysis and 19 (68%) had normal vocal cord function despite gross invasion of the RLN. The smallest tumor with gross invasion of an RLN was 0.9 cm in diameter (Fig. 2), and the majority of tumors ranged from 1.5 cm to 2.0 cm (n = 20; 66.7%).

Gross invasion of the RLN was associated with a subcapsular primary tumor that invaded the medial aspect of the thyroid capsule in the paratracheal region in 96.7% (29/30) of the cases: right paratracheal (n = 18; 60%) or left paratracheal (n = 11; 36.7%) (Fig. 3). Only 1 patient (1/30, 3.3%), a 41-year old female who presented with normal vocal cord examination and palpable biopsy-proven lateral neck lymph metastases demonstrated unexpected intraoperative findings of gross invasion of the right RLN from a 1.5 cm well-differentiated PTC tumor positioned in the right lateral posterior lobe that on preoperative ultrasound appeared to be abutting, but not invading, the thyroid capsule. Representative images of ultrasonographic findings in each location are presented in Figure 4.

Discussion

Patients who had a PTC primary tumor ≤2 cm in diameter with direct gross extrathyroidal extension into the RLN at the time of surgery were selected for inclusion in this study. All tumors were located at the right paratracheal, left paratracheal, and right lateral posterior lobe subcapsular positions in the thyroid gland. As expected, these locations correspond to the anatomically predicted course of the RLN.

Tumors involving the thyroid capsule at the paratracheal interface between the capsule and the trachea are of concern, as are tumors in the right posterior lower lobe. The latter is of particular interest as it is important to recognize that invasion of the right RLN may occur in posterior locations in the right lower pole in regions where the nerve passes by the thyroid gland on its way to its expected midline paratracheal location to enter the larynx. Unfortunately, the inability to precisely identify either RLN by ultrasonography coupled with the inherent variability in the course of the nerves for any individual patient makes it impossible to more precisely specify areas of the posterior capsule or paratracheal locations where the nerve may be at risk for direct invasion (21).

Therefore, tumors that are in contact with the thyroid capsule (no normal tissue visible between the tumor and the thyroid capsule) in either the paratracheal region or in the posterior right lower lobe locations should be considered poor candidates for active surveillance.

In addition, invasion of the RLN was not observed for primary tumors <0.9 cm in diameter, regardless of tumor location, among the 123 patients treated at our center between 1986 and 2015. This encouraging finding is consistent with previous studies, which document the smallest known tumor to invade the RLN as 0.7 cm in diameter (21). However, while active surveillance is currently considered for tumors ≤1.5 cm (5, 8–10), our findings suggest that this management strategy should be used with caution for tumors in the range of 1–2 cm in diameter. The size and location of the tumor should be considered as distinct but related, and equally important, variables when selecting patients for active surveillance.

Importantly, preoperative radiological and clinical findings were often not predictive of nerve involvement at the time of surgery. Only a quarter of the patients had imaging findings suggestive of extrathyroidal extension, and only 30% had clinically apparent vocal paresis/paralysis on preoperative examination. However, the patients evaluated in this retrospective series were not being specifically evaluated as potential candidates for active surveillance and, therefore, may not have had the same careful evaluation of the tumor-thyroid capsule interface that is currently recommended.

For example, Ito et al demonstrated that in papillary thyroid microcarcinomas being considered for observation, the presence of a clearly visible normal rim of thyroid tissue between the tumor and the thyroid capsule or an acute angle between the tumor and the tracheal wall ruled out the presence of gross invasion of the trachea and was associated with minimal invasion into peritracheal connective tissue in only 2% of the cases (20). Therefore, even in the absence of radiological or clinical findings of RLN invasion, tumors in these subcapsular vulnerable locations should generally be considered as inappropriate for active surveillance (21).

Active surveillance is gaining wide acceptance as a favorable management strategy for patients with low-risk PTC. The criteria for selection of appropriate patients are well established (12,13); however, our data provide evidence that additional tumor characteristics are important in the successful implementation of active surveillance. Patients with well-differentiated PTC tumors of ≥0.9 cm in subcapsular locations, particularly in the paratracheal region or the right posterior lobe of the thyroid gland are at risk for direct invasion of the primary tumor into an RLN even if there is no evidence of extrathyroidal extension on imaging or clinically apparent vocal cord paresis/paralysis. Thus, patient selection for an effective and safe active surveillance management program should include evaluation not only of the size and growth rate of a tumor but also of its specific location in relation to the expected course of the RLN.