Reoperative Thyroid Surgery

Definitions

In 2015 the ATA recommended a risk-adapted stratification model for recurrence risk based on response to initial therapy. Patients are initially stratified using traditional low-, intermediate-, and high-risk groups based on clinicopathologic features known during the initial treatment (presence of locoregional or distant metastases, invasion, and aggressiveness of histologic subtype). This initial stratification is critical in guiding primary treatment and early surveillance decisions, but it is a static measurement of recurrence risk, which does not reflect the patient’s response to treatment, a significant factor affecting recurrence risk.

Initially proposed by Tuttle and modified in the 2015 ATA guidelines, the “response-to-treatment restratification model” acknowledges that recurrence risk is on a continuum and employs a dynamic model based on an individual’s actual response to therapy during surveillance. After definitive treatment, patients are categorized into one of four groups based on biochemical and structural response to therapy. “Excellent response” indicates no clinical, biochemical, or structural evidence of recurrence. “Biochemically incomplete response” describes a patient with rising Tg or stable anti-Tg autoantibody (TgAb) levels in a patient without evidence of structural recurrence. “Structurally incomplete response” denotes a patient with a persistent or newly identified locoregional or distant metastasis. “Indeterminate response” is used in patients with nonspecific imaging findings and equivocal biochemical profiles. By definition, “recurrence” is reserved for patients previously staged as an “excellent response” without biochemical or structural evidence of disease after initial treatment; otherwise the patients disease would be termed persistent.

Recurrence Rates and Patterns

Approximately 30% to 40% of patients with DTC will have persistent disease and 1.2% to 6.8% of patients will have structural tumor recurrences during their surveillance, as determined by Tg levels and high-resolution US. Long-term survival in patients with only biochemical evidence of disease is nearly 100%, and decreases to 85% for patients with evidence of structural recurrence during surveillance. Given that only a fraction of these patients will develop adverse outcomes related to their persistent or recurrent disease, it is important to understand structural disease progression and its clinical implications.

Studies looking at metastasis rates in prophylactic lateral and central neck dissections in patients with normal preoperative ultrasonography (cN0) demonstrated that up to 90% of patients with PTC < 1 cm have metastatic central neck nodes and up to 40% have metastatic lateral neck lymph nodes. Despite the high incidence of occult lymph node metastasis, the observed clinical locoregional recurrence rate for patients with DTC is significantly lower. Although most patients with DTC present with clinically negative (cN0) nodal disease by examination and imaging, approximately 35% of patients will have clinically evident disease at presentation. In these patients, high-risk features associated with increased rates of recurrence and persistent disease include lymph node metastases > 3 cm (27% to 32% risk of recurrence), extranodal extension, and metastases in > 5 lymph nodes (19% risk of recurrence).

In one of the largest prospective cohort studies looking at sites of recurrences, Mazzaferri et al. found that most recurrences localized to the cervical and mediastinal lymph nodes (74%), whereas 20% involved the thyroid remnant and 6% involved the trachea and adjacent muscle. In addition, 21% of patients had distant metastases, with 63% isolated to the lungs alone. In this study, the time interval from detection of distant metastasis and death was less than 5 years in 49% of cases, 5 to 9 years in 38%, 10 to 14 years in 20%, and 15 years or more in 8% of cases.

Shah and colleagues retroactively reviewed 86 patients who underwent revision central compartment surgery with or without lateral neck dissection. 43% of patients had recurrent disease confined to the central neck, whereas 12% had disease limited to the lateral neck and 35% of patients had involvement of both the central and lateral neck compartments.

A large study by Merdad et al. looked at patterns of nodal involvement in 185 patients undergoing 248 selective neck dissections of levels II-Vb for biopsy-proven DTC metastatic to cervical lymph nodes. 73% of patients had metastatic disease in multiple levels, with 9% demonstrating skip lesions. The patterns of nodal involvement were as follows: level II, 49%; level III, 76%; level IV, 61%; and level Vb, 29%. A subsequent meta-analysis by Eskander et al. looked at 18 studies in 1145 patients undergoing 1298 neck dissections for well-differentiated thyroid carcinoma to determine patterns of disease spread within the lateral neck with similar rates of nodal spread, leading the authors to recommend that a comprehensive selective neck dissection for recurrent DTC should encompass levels II, III, IV, and Vb.

Serum Thyroglobulin Levels

Serum Tg measurement is the most widely used method for early detection and monitoring of persistent or recurrent thyroid cancer. Tg is produced by thyroid epithelial cells in response to thyroid-stimulating hormone (TSH) stimulation. As such, it is highly specific for the presence of thyroid epithelial cells, but serum Tg is only a reliable marker in the setting of a total thyroidectomy. It has a half-life of 65 hours and should be undetectable within weeks of a total thyroidectomy without any remnant thyroid tissue. Persistently elevated Tg after surgery suggests residual tumor burden (residual or recurrent tumor, regional nodal disease, or distant metastases) or residual functioning benign thyroid tissue. By following the trends of serum Tg levels, one can determine the doubling time of Tg, and by inference, thyroid/cancer burden. A doubling time of < 1 year has been shown to portend a poorer prognosis for disease recurrence and possibly aggressiveness.

Older Tg tests were less sensitive in detecting small increases of Tg, especially in patients undergoing TSH-suppression therapy with levothyroxine [T4] therapy. These patients usually have undetectable or very low basal Tg levels. However, upon levothyroxine withdrawal (hypothyroid state), or administration of recombinant thyrotropin (rhTSH), serum Tg increases 10-fold. Furthermore, comparing stimulated-Tg titer to suppressed-Tg titer estimates the tumor’s sensitivity to TSH. Normal thyroid and well-differentiated thyroid carcinomas should display > 10-fold increase in the stimulated-Tg titer, whereas poorly differentiated thyroid carcinomas only modestly respond to TSH stimulation (less than threefold increase in Tg). Although a useful adjunct in complex cases, stimulated-Tg testing has some disadvantages; levothyroxine withdrawal can take weeks and is associated with decreased quality of life in patients, and rhTSH is expensive and may be less sensitive than levothyroxine-withdrawal.

Newer second-generation, high-sensitivity immunometric assays (Tg ^2G IMA) have detection sensitivities < 0.1 ng/mL, a full magnitude of power more sensitive than the first-generation tests. Previously “undetectable” Tg levels may now be detectable, necessitating a new definition of biochemical cure, suppressed-Tg < 0.2 ng/mL or stimulated-Tg < 1 ng/mL. The increased sensitivity of the second-generation tests has obviated the need for routine TSH-stimulated Tg testing.

Approximately one-third of patients with DTC have detectable levels of anti-TgAb or heterophilic antibodies (most commonly HAMA) at some point during treatment, with 12.7% of patients clearing the antibodies after successful treatment (positive to negative). Tg antibodies invariably interfere with the measurement of Tg to some degree, regardless of testing methodology. Therefore current guidelines mandate concurrent testing of TgAb titers to validate each Tg test. To variable extents, failure of TgAb levels to fall, rising TgAb titers, and de novo TgAb appearances are all highly predictive of an increased risk of recurrence.

One important caveat to trending Tg and TgAb levels during long-term surveillance is that levels can only be compared if the same testing methodology is used, and ideally within the same laboratory on the same machine. Each manufacturer uses different immunoassays to test various epitopes with different reagents, making between-method testing variability too high to compare results.

Ultrasonography

Ultrasonography is a highly sensitive and specific imaging modality used to monitor patients for recurrent thyroid carcinoma. With detection limits as small as 3 to 4 mm, preoperative US mapping improves the detection and assessment of occult lymph node metastasis. Stulak et al. reviewed almost 1000 patients who underwent preoperative US scans before thyroid surgery (primary and revision) and showed that in reoperative patients, nonpalpable disease was detected with US in 64% of patients. In patients with palpable neck disease, US altered the extent of planned operation in 43% of reoperative cases, with reported sensitivity, specificity, and positive predictive values of 90%, 79%, and 94%, respectively.

US’s advantages include its portability, ease of use, lack of ionizing radiation, and physiologic evaluation with Doppler mode. It has a key role in the standard posttreatment surveillance algorithm, and can detect recurrences in nonradioiodine-avid lesions and when Tg measurements are compromised by the presence of autoantibodies. All pretreatment and surveillance scans should encompass both the central compartment and lateral necks in a comprehensive fashion. Intraoperative US can be used during revision surgery when thick scar tissue hinders localization of recurrent disease (see “Adjunctive Techniques in Revision Surgery” section).

Despite the clear-cut utility of surveillance US imaging in patients with thyroid cancer, it has a few specific limitations when working up patients with suspected recurrent disease. US is limited in evaluating retropharyngeal, retrotracheal, and mediastinal disease, and has poor sensitivity for tracheal invasion and extranodal extension of disease. Any patients with suspected recurrent disease should be further evaluated using either contrast-enhanced CT or MRI for accurate assessment of these areas and features.

CT Axial Imaging With Contrast

Contrast-enhanced CT imaging is commonly used in the head and neck in the evaluation of nodal metastases (see Chapter 14 , Preoperative Radiographic Mapping of Nodal Disease for Papillary Thyroid Carcinoma). CT is not operator dependent but is highly repeatable, obtained with 1.5 mm axial cuts, and has a very fast acquisition time. CT images are familiar to surgeons and give detailed anatomic information. They image lymph node regions that are less accessible or less accurately seen by US, such as the central neck, mediastinum, and retropharynx, and is the test of choice for evaluating laryngeal or tracheal cartilage invasion in thyroid cancer. CT imaging also remains the most sensitive modality for detection of multiple small lung deposits from thyroid cancer, which are often missed on whole body scan (WBS) and PET/CT.

Although it had previously been recommended to avoid iodinated contrast because of the delay in RAI ablation by 8 weeks, this is no longer recommended. RAI ablation is not typically given until 6 to 8 weeks after surgery, and any delay resulting from iodinated contrast is minimal. Furthermore, data has shown that even significant delays in RAI administration do not have an adverse effect. Given the small size of most residual or recurrent thyroid masses, the benefit of contrast on localizing locoregional disease and evaluating soft tissue invasion outweighs the delay for RAI ablation. Communication is important when patients are administered iodinated contrast so that appropriate modifications to any planned RAI scanning or ablation can be made.

The combined use of US and contrast-enhanced CT is extremely useful in surgical planning to map out the location of suspected metastases and structural invasion. The surgical planning map amalgamates the preoperative localization studies allowing the surgeon to create a three-dimensional map of nodal targets to minimize the risk for persistent and recurrent disease.

Other Imaging Modalities

Fine-Needle Aspiration and Thyroglobulin Washout

The ATA guidelines have strongly recommended that suspicious lesions should be at least 8 mm in the central neck or 10 mm in the lateral neck before performing a biopsy or surgical resection. Although these measurements are typically measured in the short axis, some authors have recommended that a threshold of > 8 mm or > 10 mm in any diameter can signify disease and is sufficiently large enough to be targetable for surgical localization. Lesions that are smaller than 8 mm/10 mm should be closely followed for progression, but generally should not be biopsied until they reach the threshold size limits.

US-guided FNA allows for an accurate evaluation of suspicious lesions within the thyroid bed. In one study, the reported sensitivity of US-guided FNA for diagnosing recurrent carcinoma in the thyroid bed after total thyroidectomy was 100%, with a specificity of 85.7%. However, inconclusive and false-negative results occur. By evaluating Tg measurements in fine-needle aspirate washouts (FNAC-Tg) of suspected metastatic lymph nodes and disease recurrence, the diagnostic accuracy and sensitivity of FNA increased significantly, with Cuhna et al. reporting 100% sensitivity and accuracy. Furthermore, this relatively simple and inexpensive assay is valid even in the presence of TgAbs.

Active Surveillance

Active surveillance with close serial follow-up using Tg, US, and possibly CT/MRI is a valid option for patients with low-volume disease. A study by Tuttle et al. found that small (< 11 mm) postoperative thyroid bed nodules occur in approximately one-third of patients undergoing total thyroidectomy, with and without adjuvant RAI ablation. Of these, less than 10% of these nodules were found to be malignant lymph nodes, and even fewer demonstrated progression over time. A second study by Robenshtok et al. showed that lateral neck lymph nodes also demonstrated a low potential for disease progression, even among nodes with imaging suspicious for malignancy. Over a median of 3.5 years during surveillance, only 9% of nodes increased by more than 5 mm. In both studies, delay of surgical resection until disease progression was not associated with local invasion or distant metastases. In the absence of suspicious US characteristics and stable Tg levels, the majority of subcentimeter thyroid nodular disease will remain stable over a 5-year period. In contrast, a Tg doubling time of less than 1 year is a negative prognostic indicator that should prompt discussion for possible intervention. Therefore in the setting of low-volume disease, active surveillance may serve as a viable option. Many patients will be pleased with this option to avoid further surgery if they have an understanding that this is a safe and reasonable alternative.

Medical Therapies

There is growing evidence in the safe and effective use of selective treatment such as US-guided percutaneous ethanol ablation, radiofrequency ablation (RFA), and laser ablation therapies in select patients with persistent/recurrent disease (see Chapter 51 , Nonsurgical Treatment of Thyroid Cysts, Nodules, Thyroid Cancer Nodal Metastases, and Hyperparathyroidism: The Role of Percutaneous Ethanol Injection). The Mayo Clinic has more than 25 years of experience with percutaneous ethanol ablation and has reported excellent outcomes in lymph node response rates based on size and Doppler blood flow. In one study of 109 patients, Lewis et al. found that 93% of patients demonstrated decreased nodal volume, and they reported successful treatment of nodal disease up to 3 cm as well as treatment of lesions in close proximity to the IJV, carotid artery, and RLN. A second study by Heilo et al. looked at 64 patients who received ethanol ablation and found that 84% of lymph nodes responded completely to ethanol ablation based on imaging. These techniques often require multiple administrations to obtain satisfactory results but may confer significant health system cost savings compared with surgery and postoperative hospitalizations.

RFA under US guidance is a similar nodule-directed technique for nonsurgical ablation. RFA typically uses a 1 to 2 cm active tip probe placed under local anesthesia within the center of a nodule and heated to 90° C for 2 minutes, creating a coagulative necrosis of the thyroid nodule. Some have argued that the heat generated may cause damage to the RLN, parathyroid gland, carotid, or IJV, though many studies have demonstrated the safety and efficacy, albeit in limited cohorts of patients. Currently, the ATA guidelines state a general consensus, but no formal recommendation, that both ethanol ablation and RFA should be considered in patients who are poor surgical candidates. Most studies that describe this technique are limited in sample size, had a relatively short follow-up, and consisted of lymph node disease smaller than a centimeter.

Although surgical resection continues to be favored in the ATA guidelines for recurrent structural disease management, RAI may be an option in certain situations, including patients with rising Tg levels in the absence of radiographic disease, those with surgically unresectable disease, such as disease involving a single functioning RLN, and distant metastatic disease where surgical resection may not provide a cure. However, the ATA guidelines recommend against the use of RAI in cases with iodine-refractory masses on scintigraphy or persistently elevated Tg levels despite RAI administration. Important side effects and complications of RAI, including xerophthalmia, xerostomia, and radiation-induced malignancy to breast, bone, and stomach, must be discussed with the patient and weighed into the risk-benefit analysis (see Chapter 48 , Postoperative Radioactive Iodine Ablation and Treatment of Differentiated Thyroid Cancer).

Preoperative Documentation

Careful review of the primary surgery operative note (and subsequent operative notes, if any) and all pathology reports is essential for planning revision surgery. Factors associated with increased surgical difficulty include the need for resection of the strap muscles, extensive IJV dissection, and cases associated with sternotomy, as these scenarios cause potentially dangerous distortion of venous anatomy in the lower neck. The status of the parathyroid glands may be discerned from the operative note, but in many cases, it can be more helpful to look for the presence of parathyroid tissue in the pathology reports as it is not uncommon to have unrecognized parathyroid gland removal.

The status of the RLN should be noted in the initial operative report, and its functional status should always be tested and documented preoperatively with the use of indirect laryngoscopy or stroboscopy. Unrecognized RLN injuries occur during thyroid surgery, and clinical symptoms alone are not reliable for documenting RLN function. Recognizing RLN dysfunction preoperatively is essential for surgical planning, especially in circumstances where there is risk of injury to the last functioning RLN, as this may alter the extent of surgical resection or increase the likelihood of a tracheostomy postoperatively. Reoperation on the last functioning nerve should only be performed after a thorough discussion with both the patient and the referring endocrinologist because this can dramatically alter the quality of life after surgery, and the surgical risk may be higher than other nonsurgical options.