Introduction

Approximately one third of patients with differentiated thyroid cancer (DTC) have tumor recurrence; most are diagnosed within 10 years of initial treatment.^1–3 Locoregional recurrences may arise in the thyroid bed, the central or lateral neck, the mediastinum, or, rarely, in the trachea or the muscle overlying the thyroid bed. The mortality from locally recurrent disease in low-risk group patients (according to the Age, Metastases, Extent, and Size [AMES] prognostic index) with DTC is only 4%. However, high-risk patients, such as males and those older than 45 years of age, experience a significantly higher mortality rate of 27% with disease recurrence.⁴ Clinical or radiologic evidence of locally recurrent DTC is generally treated with surgical removal of the focus of disease and postoperative iodine-131 with some patients receiving postop radiotherapy (XRT) as well, depending on the tumor extent, exact histology, and completeness of resection. Among younger and low-risk patients (< 45 years), an increasingly preferred approach is to observe patients with nonbulky disease (less than 1 cm), after informed discussion with the patient, and to only recommend surgery with clear evidence of progressive disease.

Revision or reoperative thyroid surgery is often technically challenging because of anatomic changes and reparative fibrosis following primary surgery especially in the central neck (see Chapter 10, Reoperation for Benign Disease). Consequently, reoperative surgery may be associated with high rates of complications in inexperienced hands.^5–7 However, with experience and appropriate preparation, the risk of permanent hypoparathyroidism or recurrent laryngeal nerve (RLN) injury after reoperative surgery is reported to be low (less than 3% and 1%, respectively).^8,9 Surgeons contemplating revision thyroid surgery must possess the essential reoperative surgical skills and an intimate knowledge of regional anatomy to achieve such a low morbidity for what can often be a tedious and difficult procedure.

Indications for Revision Surgery

Reoperative thyroid surgery may be performed under a number of circumstances. A patient may have had a previous thyroid lobectomy and require a completion thyroidectomy for DTC. Recurrence of thyroid cancer may require reexploration of the thyroid bed and usually central compartment lymph node dissection sometimes combined with the lateral neck (see Chapters 38, Central Neck Dissection: Technique, and 40, Lateral Neck Dissection: Technique). Additionally, recurrence of cancer in the thyroid remnant, after a partially removed nodular or multinodular goiter for benign symptomatic thyroid disease, will also require reoperation.

Anatomic Changes Following Thyroidectomy

A thorough knowledge of the normal anatomy of the neck (and its common variations), in particular the central visceral compartment, provides the fundamental basis for considering both primary and reoperative surgery. However, changes in cervical anatomy often take place following primary thyroidectomy, which should be borne in mind. Scarring from prior surgery may cause increased difficulty in identifying and dissecting important structures. The carotid sheath is an important landmark in thyroid surgery as it provides the constant lateral boundary of dissection and is an important landmark when initially identifying the RLN.¹⁰ Realizing that the great vessels of the neck may medialize following thyroidectomy and may be directly adjacent to and often densely adherent to the trachea is important. If the strap muscles were excised because of tumor involvement, the great vessels may be superficial and medial. In circumstances where the strap muscles were previously resected, scarring in the central neck and reoperation is typically dense. Further, the internal jugular vein may become superficial and adherent to the undersurface of the anterior border of the sternocleidomastoid (SCM) muscle, and the brachiocephalic artery may be drawn high in the lower central neck as a result of scar contracture. These venous structures in reoperative surgery are perhaps the most difficult to manage, primarily because of their tendency to scar to adjacent musculature. When reoperating on the area of the ipsilateral thyroid remnant, the recurrent laryngeal nerve is often difficult to identify and dissect because it is buried in scar tissue. Scarring under the strap muscles may lead to a superficial RLN adherent to the undersurface of the muscles. On occasions RLNs have been described on the anterior wall of the trachea due to scarring and wound contracture. Typically during revision central neck dissection the recurrent laryngeal nerve is scarred immediately to the upper cervical trachea over a span of the first several tracheal rings and lateral cricoid and is most commonly injured in this segment (see Chapter 33, Surgical Anatomy and Monitoring of the Recurrent Laryngeal Nerve). The tracheoesophageal groove itself may be scarred and distorted, making the identification of the esophagus almost impossible. Placement of a bougie may help prevent inadvertent esophageal perforation during dissection. The location of preserved parathyroid glands will often be grossly distorted and requires a high index of suspicion of any potential parathyroid tissue during reexploration. Parathyroid and recurrent laryngeal nerve identification during revision central neck dissection require a bloodless field.

Preoperative Workup

Serum thyroglobulin measurements and more recently high-resolution ultrasound scanning represent the standard method of surveillance for persistent and recurrent thyroid cancer following initial surgery. Alternative strategies must be used when thyroglobulin is not a valid tumor marker, such as in undifferentiated thyroid cancer or in patients with thyroglobulin antibodies. Radioiodine scanning (RAI) does not provide sufficiently accurate anatomic imaging data upon which to base surgery. In addition, iodine avidity is typically lost in a high percentage of patients with recurrent disease. In patients without radioiodine uptake, PET uptake may be present and is discussed later. It is not typically RAI or PET scanning but anatomic imaging (i.e., ultrasound and axially imaging including CT) that form the basic surgical map for revision surgery.

Subsequently, preoperative assessment aims to, first, confirm the diagnosis of persistent or recurrent malignancy and, second, to provide imaging of its anatomic site and extent to allow planning of surgical treatment. A fine-needle aspiration (FNA) biopsy performed under ultrasound guidance may be carried out on suspicious lesions found on physical examination, radiologic examination, or radioiodine scan. FNA should be sent for cytology as well as for thyroglobulin rinsing.

Serum Thyroglobulin Levels

Serum thyroglobulin (Tg) level measurement is the most widely used method of early detection and monitoring for persistent or recurrent thyroid cancer. Persistently elevated thyroglobulin after surgery or focal radioiodine avidity suggests residual tumor burden or residual functioning thyroid tissue. When thyroglobulin is elevated because of tumor burden, it may be due to residual or recurrent tumor, regional nodal disease, or distant metastases. Thyroglobulin doubling time of less than 1 year has recently been found to have significant prognostic importance.¹¹

When thyroid-stimulating hormone (TSH) is low (on levothyroxine [T4] therapy), basal serum Tg may be undetectable or only mildly elevated, and recombinant human thyrotropin (rhTSH) administration or thyroxine withdrawal may be useful, if clinical suspicion of disease exists, to increase serum Tg into the more significant and measurable range. The increase in Tg titer in response to rhTSH is an index of the tumor’s sensitivity to TSH and is calculated by dividing the rhTSH-stimulated Tg level by the basal Tg level. Normal thyroid remnant and differentiated thyroid cancer (DTC) display a greater (> 10-fold) serum Tg response to TSH stimulation compared with less well-differentiated tumors (< 3-fold).

Ultrasonography

Ultrasonography has been shown to be a highly sensitive and specific technique that can be used to monitor patients for recurrent thyroid carcinoma in the thyroid bed after total thyroidectomy. A major advantage is that it can detect recurrences that are noniodine avid and when Tg measurements are compromised by the presence of antibodies. One large study reported sensitivity, specificity, and positive predictive values for ultrasound (US) in reoperative patients of 90%, 79%, and 94%, respectively.¹² The ultrasound scan must encompass both the central compartment and lateral necks in a comprehensive fashion. High-resolution ultrasound offers detection limits as small as 3- to 4-mm metastatic or recurrent deposits, and thus allows for the accurate detection of cervical lymph node and soft tissue metastases that are not evident on physical examination.¹³ Preoperative high-resolution ultrasound mapping has been shown to improve the detection and assessment of lymph node metastasis in patients with persistent or recurrent papillary thyroid cancer (PTC).¹² Ultrasound becomes especially helpful in reoperative planning and determination of the need for further neck dissection. Stulak et al. reviewed almost 1000 patients who underwent preoperative US scan prior to thyroid surgery (primary and revision) and showed that in reoperative patients, nonpalpable disease was detected in 64% via US.¹² They also showed that even when disease was palpable in the neck, US assessment altered the extent of planned operation in 43% of reoperative cases. However, US imaging is relatively poor at the skull base, retromanubrial, retropharyngeal, and retrotracheal spaces. Also, it offers poor sensitivity for tracheal invasion and extranodal extension of disease. Computed tomography (CT) or magnetic resonance imaging (MRI) scanning is required for accurate assessment of these areas and features.

Preincision or intraoperative ultrasonographic localization during surgery can also be used when thick scar tissue hinders localization of recurrent disease (see the “Adjunctive Techniques in Revision Surgery” section). Intraoperative ultrasonographic localization can be difficult in the detection of small deposits of tumor in this setting of an irregular surgical field, which can lead to difficulties with probe placement.

Despite the clear-cut utility of US in these revision patients, high-resolution CT imaging of the thyroid and neck (obtained from the skull base to the midthoracic area) and conventional CT of the chest should be obtained in the evaluation of these patients.

CT Axial Imaging with Contrast

Contrast enhanced CT imaging is commonly used in the head and neck in the evaluation of nodal metastases. CT is not operator dependent and is done in a highly repeatable fashion and performed with fine cuts that are several millimeters in size. CT images are familiar to surgeons and can give detailed anatomic information. Additionally, CT is used as the test of choice for evaluating laryngeal or tracheal cartilage invasion in thyroid cancer. CT performs better than MRI in the determination of positive lymph nodes.¹⁴ Characteristics of lymph node metastases on CT scans include cystic change, enhancement, calcification, and enlarged size.

The downside of CT imaging is that when the CT is performed with iodinated contrast, the patient will need to wait approximately 2 months before receiving radioactive iodine. Recent data have shown, however, that even significant delays in radioactive iodine do not have an adverse effect.¹⁵ Generally, radioactive iodine treatments are not given until 6 to 8 weeks following surgery; therefore, the delay resulting from the CT contrast imaging is minimal. The benefits of contrast on localizing lymph node metastases allowing for effective surgery outweigh the short wait for radioactive iodine while the patient recovers from surgery. Further, it has been our experience that the majority of patients with recurrent central neck disease no longer maintain radioactive iodine avidity, a fact confirmed by others. Clayman et al. found that 82% of those with recurrent disease scanned failed to take up radioactive iodine.¹⁶

Evidence is now emerging that the combination of CT and ultrasound will give the highest yield in preoperative nodal planning (see Figures 14-1, 14-2, and 14-3 in Chapter 14, Preoperative Radiographic Mapping of Nodal Disease for Papillary Thyroid Carcinoma). The combination approach has been found to be superior to ultrasound alone.^17,18 Kim and others found the combination approach increased the sensitivity of detection of central lymph node metastases when compared to ultrasound alone.

Recently, in a study examining the efficacy of lymph node detection by physical exam, ultrasound, and contrast-enhanced CT, the greatest sensitivity in imaging of the central and lateral neck was demonstrated to be with a combination of CT and ultrasound.¹⁸ Primary and revision patients with papillary thyroid cancer benefited from imaging with both modalities. The sensitivity of CT scanning was found to be superior to US in the central neck of patients undergoing primary surgery for thyroid cancer (50% versus 26%). The addition of the CT scan gave valuable data in the search for macroscopic disease in the central neck that would have been missed by ultrasound alone. In both the central and lateral neck, the greatest sensitivity is delivered by the combination of US and CT compared to physical exam or ultrasound alone. Importantly, in 26% of patients overall, dissection based on a positive CT scan but negative ultrasound scanning yielded positive lymph nodes on pathology that would not have been removed with ultrasound alone; 25% of patients undergoing primary surgery and 27% of patients undergoing revision surgery had lymph nodes removed because of CT findings that would not have been removed based on ultrasound alone. This represents a significant number of patients who may have been spared from further required surgeries.

Detailed radiographic evaluation is important, as central neck recurrences are typically small. Farrag et al., in a study of 33 patients, noted average lesional nodal size was only 1.4 cm.¹⁹ Others recommend at least one lesion of 1 cm and greatest diameter and evidence of progression of disease greater than a 50% increase in size over a 6-month period of observation.¹⁶ Rondeau et al. studied a cohort of patients with thyroid bed nodularity identified on postoperative surveillance ultrasonography. The average age of these patients was 40 years, and 84% had been treated with radioactive iodine ablation postoperatively. Thyroid bed nodules were all less than 11 mm, with a mean size of 5.7 mm. In only 9% of patients did the nodules significantly increase in size, on average 1.3 mm per year. An increase in size was less likely if there were no other worrisome cervical lymphadenopathies, no worrisome intralesional ultrasonographic findings, and a stable thyroglobulin profile.²⁰ We must keep in mind that not all thyroid bed nodules identified on surveillance ultrasonography are malignant. This is especially true if patients have not been treated with post total thyroidectomy iodine ablation where benign thyroid remnant nodularity may persist.

The combination approach of US and CT ensures that all compartments are viewed adequately prior to surgery, and the two modalities complement each other well. Ultrasound gives specific information about the lymph nodes including hilar anatomy, length-to-width ratio, calcifications, and cystic change. CT imaging provides localization of the lymph nodes in relation to surrounding viscera, which the surgeon can easily use. It also images the lymph node regions that are less accessible or less accurately seen by ultrasound, such as the central neck, mediastinum, and retropharynx. This combined US and CT approach is extremely useful in surgical mapping. An expert head and neck radiologist can evaluate lymph nodes with the surgeon on both modalities to map out the location of suspected metastases. During the surgery, the CT scan, with its full image of the neck and anatomic landmarks, can be used to guide the surgeon to the precise location of the previously mapped lymph nodes. CT scanning is also helpful in evaluating the extent of the invasion of recurrent disease.

CT scan also remains the most sensitive modality for detection of multiple small lung deposits from thyroid cancer that cannot be detected with US and is often missed on whole-body scintigraphy (WBS) and PET/CT. There should of course always be good communication between surgery and endocrinology staff when patients are administered iodine containing contrast so appropriate modifications to any planned RAI scanning or treatment can be made.

Of paramount importance, a surgical planning map, amalgamating all the preoperative localization studies utilized, is formulated to minimize the risk for persistent and recurrent disease. It is primarily on the basis of anatomic studies blending high-resolution cervical ultrasound and axial CT scanning upon which a three-dimensional cervical map of nodal targets is created, which then guides surgery.

Other Imaging Modalities

Preoperative Details/Documentation

Careful review of the primary surgery operative notes must be made to aid in planning the revision surgery. We have found when initial cases are associated with strap muscle resection and significant jugular vein dissection, reoperative surgery can be substantially more difficult. This is also true when initial surgery involved cervical work in conjunction with sternotomy. This substantially distorts the venous anatomy of the neck base dangerously. Of course, patients subjected to external beam radiation therapy make for difficult reoperation subjects. In particular, the status of the RLNs and parathyroid glands must be discerned. Preoperatively, indirect laryngoscopy or video stroboscopy should always be performed to assess and document vocal cord function and any prior RLN damage. Unrecognized RLN injuries occur after thyroid surgery and clinical symptoms alone are not reliable for documenting RLN function. Lo et al. showed a 6.6% incidence of RLN palsy on direct laryngoscopy postoperatively, whereas only 1% had recognized nerve damage during the operation.³¹ Further, Echternach et al. reported a 31% incidence of vocal cord injury on stroboscopic evaluation postoperatively with only 6% of problems being due to paresis secondary to RLN injury.³² Recognizing RLN palsy preoperatively also allows both the surgeon and the patient to be prepared for the possible need for a tracheostomy after the revision thyroid surgery if the contralateral RLN is injured.

One must weigh the benefits and risks of reoperative surgery. We feel only lesions that are of sufficient size greater than 8 mm and can be clearly localized with combined ultrasound and CT scan are rational surgical targets. Targeting lesser lesions will likely increase the rate of negative explorations. The progression of a lesion is also a reasonable parameter to consider, and surgeons should be less willing to operate if the lesion has been proved over time to be stable. Reoperation on only functioning nerve also must be considered with great caution and after thorough discussion with both the referring endocrinologist and the patient. Under no circumstances should an exploration be undertaken if a targetable surgical lesion cannot be localized on anatomic imaging of ultrasound and CT scan.

Informed Consent

Reoperation can be associated with significantly increased risks, in particular to the RLNs and parathyroid gland function. Damage to these structures from the initial surgery will have more significant implications and should be discussed with the patient thoroughly, including the remote risk of the need for tracheostomy.

However, reoperative surgery has been shown to have low morbidity in experienced hands, and many studies have reported rates ranging from 0.4% to 4% of permanent hypocalcemia or RLN injury (see the “Complications” section, presented later in the chapter). Each surgeon should know his or her individual operative risk rate, and if the volume of reoperative surgery is low, then referral to a more experienced surgeon may be in the patient’s best interest.

Surgical Therapy

The two main indications for reoperation are completion thyroidectomy for thyroid cancer diagnosed after initial lobectomy, and reexploration of the central compartment (+/− lateral neck) and superior mediastinal dissection for recurrent disease after prior total thyroidectomy. The surgical principles and technical approach to these two scenarios will be discussed separately.

The use of surgical loupes and a headlight may be of benefit in the identification of critical structures and aid in safe disease removal. Many surgeons have routinely used intraoperative neural monitoring (IONM) techniques in an attempt to provide a functional assessment of the RLN during surgery. Although not considered as a substitute for detailed knowledge of anatomy, the use of IONM can be valuable in maintaining the functional integrity of the RLN, particularly in reoperative thyroid surgery, where the location of the RLN is less constant and embedded in scar tissue (see Chapter 33, Surgical Anatomy and Monitoring of the Recurrent Laryngeal Nerve).^33,34 A large study evaluating the use of IONM on 1000 RLN-at-risk patients showed that the postoperative RLN palsy rate for reoperative patients for reoperative patients was much higher in the control group compared with the neuromonitored group (19% versus 7.8%).³⁵ However, many other studies have failed to demonstrate definitive evidence for IONM in reducing the rate of RLN injury during thyroid surgery, and the authors recommend direct localization and visualization as the optimal method of nerve function preservation.³⁶ With the use of neural monitoring in reoperative surgery, international standards should be reviewed in order to provide optimized neural monitoring.³⁷

For hemostasis, both bipolar electrocautery and vascular suture ligation are utilized. Surgical clips may result in scatter artifact, which may complicate subsequent cross-sectional imaging (CT or MRI) during follow-up. An important surgical technique in the reoperative setting is to first work in areas more distant from the prior dissection to safely identify neural and vascular structures and then to safely transition into the previously dissected areas. Also as mentioned previously, the carotid artery is the key landmark that allow initial global orientation to the neck reoperative field.

Reoperative Surgery: Completion Thyroidectomy

Because the best way to minimize the complication risk of revision surgery is to avoid it altogether, thyroid surgeons should be aware of several important ways to decrease the need for completion thyroidectomy. Surgeons should also take care to ensure that any unavoidable need for reoperation can be facilitated by employing good surgical principles in the initial surgery in order to reduce associated morbidity.