Surveillance for differentiated thyroid cancer can be performed under unstimulated or stimulated conditions. Unstimulated testing refers to testing that is performed while the patient continues to take thyroid hormone at replacement or TSH-suppressive doses. Hence, unstimulated testing is also known as testing “on suppression.” Stimulated testing refers to testing that is performed after thyroid hormone withdrawal (THW) for a time period sufficient to raise endogenous serum TSH > 30 mIU/L¹ [17, 18] or after administration of exogenous recombinant human TSH (rhTSH or Thyrogen^®) [19]. These methods of stimulation are known as withdrawal-based testing and rhTSH-based (or Thyrogen^®-based) testing, respectively. The purpose of stimulated testing is to increase the sensitivity of the surveillance method in question, i.e., the likelihood of finding residual or recurrent disease.

Of the surveillance methods described below, whole-body iodine scanning is the only test performed solely under stimulated conditions. The sodium-iodine symporter responsible for uptake of iodine into thyroid follicular or epithelial cells is directly stimulated by TSH [20]. Residual thyroid cells—normal or cancerous—are therefore unlikely to take up radiolabeled iodine unless TSH is raised endogenously or administered exogenously. Moreover, it is also recommended that a period of dietary iodine restriction with low-iodine diet (LID) be undertaken prior to whole-body iodine scanning, again to facilitate uptake of radiolabeled iodine into any remaining thyroid cells. Data for LID is less robust than that for raising the TSH, but the potential for enhancing sensitivity of WBS is felt to outweigh the relatively limited harms (primarily inconvenience and unpalatable diet) of dietary iodine restriction [21].

Serum thyroglobulin can be measured on suppression or after stimulation. The sensitivity of thyroglobulin for detection of disease is greater with stimulation, since TSH stimulates the release of the thyroglobulin protein from thyroid epithelial or follicular cells into the bloodstream [4]. With the advent of “ultrasensitive” thyroglobulin assays, stimulated thyroglobulin testing can be deferred in many low-risk patients with no clinical or radiographic evidence of disease with undetectable thyroglobulin on suppression [22]. But for patients at intermediate or high risk for thyroid cancer recurrence, or those with a moderate to high pretest probability for recurrence, stimulated thyroglobulin measurement remains a powerful tool for detection [4].

Anatomic imaging studies—which include neck ultrasound and cross-sectional imaging (CT, MRI)—do not change based on serum TSH. Hence, there is no need to perform these tests under stimulated conditions. Positron emission tomography (PET) scanning, a functional imaging technique that visualizes uptake of a radiolabeled analog of glucose (¹⁸F-FDG) into areas with high metabolic activity, may demonstrate improved sensitivity under stimulated conditions. Elevated TSH is known to induce expression of glucose transporters in thyroid follicular cells, resulting in greater visualization on PET [23]. This effect, however, is not as pronounced as the effect of elevated TSH on the sodium-iodine symporter. Thus, raising serum TSH is not essential for PET, in contrast to the conditions required for whole-body iodine scanning [7]. The decision to recommend stimulation with THW or rhTSH prior to PET should be based on the individual clinical scenario. Patients who are already undergoing THW or rhTSH in preparation for stimulated thyroglobulin and WBS, and in whom PET is also needed, would not experience any additional logistical or preparation burden for a stimulated PET and might benefit from improved diagnostic yield. Likewise, patients in whom disease was previously detected on stimulated PET may need to be followed with repeat stimulated PET to adequately assess response to therapy on an analogous exam.

Figures 22.1 and 22.2 provide detailed timelines for preparation of DTC patients for stimulated testing with THW or rhTSH, respectively.

Evidence for THW in general, as well as for various withdrawal protocols, comes largely from patients being prepared for treatment with radioiodine rather than from patients being prepared specifically for diagnostic or surveillance testing. In early observational studies, endogenous elevation of TSH to >30 mIU/L appeared necessary for thyroid remnants to significantly concentrate administered radioiodine [17]. More recently, two randomized controlled trials reported that either directly stopping levothyroxine for 4–5 weeks or substituting levothyroxine with liothyronine for the initial 2–3 weeks resulted in similar short-term quality of life and hypothyroidism symptom scores; and in patients being prepared for treatment, rates of successful ablation were similar [18, 24]. One of the RCTs included patients being prepared for diagnostic scanning, but the number of patients was very small, and the ability to detect residual or recurrent disease or persistent thyroid tissue was not specifically studied [24].

The extremely favorable prognosis of most cases of DTC, coupled with the rapidly increasing incidence and detection of low-risk tumors since the early 1990s, intensified interest in surveillance protocols that were safe and better preserved quality of life. Specifically, there was a need for methods to raise serum TSH without inducing clinical hypothyroidism. Bovine TSH proved effective but was abandoned due to the development of allergic reactions as well as TSH-neutralizing antibodies [25]. Reduction of daily levothyroxine to half the usual dose was proposed but not confirmed in large series. Moreover, this approach attenuated but did not entirely avoid the consequences of clinical hypothyroidism [26].

The large-scale production of recombinant human TSH became possible after the β-subunit coding of the TSH gene was cloned, allowing for overexpression of the encoded TSH protein in a cell system [27–29]. (TSH is a pituitary glycoprotein composed of an α-subunit shared by the pituitary gonadotropins FSH and LH and a hormone-specific β-subunit.) Recombinant human TSH has the same biologic properties as native TSH and an identical amino acid sequence, but is less glycosylated and has more sialic acid. While these differences result in a lower affinity of rhTSH to the TSH receptor compared to endogenous TSH [27], preclinical in vitro and animal studies demonstrated that rhTSH increased serum T4 and T3 and stimulated radioiodine uptake into thyroid cells [30, 31]. Additional studies showed that rhTSH was a potent stimulus for T4, T3, and thyroglobulin release in normal human subjects [32, 33].

Initial clinical trials of rhTSH in DTC patients included several phase I/II studies for dose-finding and pharmacokinetics [34], followed by two phase III studies which established its safety and efficacy in this population [19, 35]. The first phase I/II study was completed in 1994 in 19 patients following thyroidectomy. It showed that rhTSH administration at doses of 0.9–3.6 mg for 1–3 days produced iodine scans of equal quality and with similar number of abnormal uptake sites in 63% of subjects. Serum thyroglobulin more than doubled in 73% after rhTSH, though the absolute rise in TSH was lower after rhTSH compared to THW in most subjects [34].

The first phase III study was conducted between 1992 and 1995 across multiple centers in the USA. One hundred twenty-seven thyroid cancer patients underwent two ¹³¹I whole-body scans, the first after rhTSH (0.9 mg IM for two consecutive days) while remaining on levothyroxine and the second after withdrawal from thyroid hormone. Compared to THW, preparation with rhTSH was associated with significantly fewer symptoms of hypothyroidism, but in 29% of patients was less sensitive for detecting residual or recurrent disease by ¹³¹I whole-body scan. The study may have been limited by the high percentage of negative scans (51% of patients) and scans with uptake in the thyroid bed only (72% of those with a positive scan), as well as by variation in the diagnostic dose of ¹³¹I and a lack of inclusion of serum thyroglobulin as an endpoint [35].

The second phase III study, performed across multiple centers in the USA and Europe, addressed some of these limitations. It was designed to compare two different dose regimens of rhTSH, as well as compare rhTSH-stimulated results to THW-based results. A total of 229 thyroid cancer patients were randomized; those in arm I received two injections of 0.9 mg on two consecutive days, and those in arm II received three injections of 0.9 mg 3 days apart. Serum thyroglobulin was included as an endpoint, both alone and in combination with WBS. The dose of ¹³¹I was fixed at 4 mCi and scanning procedures were standardized. Peak serum TSH concentrations occurred 24 h after the last rhTSH injection in both groups, though elevation of TSH persisted for longer in arm II than in arm I (9 days versus 4 days). Peak thyroglobulin levels occurred 3 days after the last rhTSH injection in arm I and 1–3 days after the last rhTSH injection in arm II. In both groups, the peak thyroglobulin was lower after rhTSH than after THW. Overall, 89% of patients had concordant scans, 16% had superior scans after THW, and 4% had superior scans after rhTSH. There was no difference between the two dosage regimens of rhTSH, suggesting the more convenient two-dose regimen could be used. The combination of serum thyroglobulin and WBS after rhTSH identified 100% of patients with metastatic disease and 93% with uptake in the thyroid bed. Moreover, a thyroglobulin cutoff of 2 ng/mL predicted thyroid bed uptake in 52% of patients and metastatic disease in 100%. Similar to prior studies, rhTSH preparation resulted in a much better quality of life, and it did so without causing significant side effects or anti-rhTSH antibodies [19].

On the basis of these studies, rhTSH was approved in the USA in 1998 and Europe in 2001 as a diagnostic tool in patients being tested for persistent or recurrent well-differentiated thyroid cancer [36]. Since its approval, multiple published studies have confirmed its efficacy and safety. As compared to THW-based regimens, rhTSH-based preparation has been associated with significant improvements in multiple quality of life measures [37, 38] and in some cases (particularly the elderly and those with underlying medical illness that are exacerbated by hypothyroidism) avoidance of the significant medical consequences of THW [39]. Another advantage of rhTSH compared to THW is the shorter duration of exposure of any remaining cancer cells to high levels of TSH, with potential to minimize the risk of residual tumor growth [40]. Indeed, suppression of TSH with supraphysiologic doses of thyroid hormone is one of the mainstays of long-term management of DTC. It thus follows that limiting exposure to elevated TSH levels might be beneficial.

Given similar rates of disease detection in most patients, and clearly improved quality of life with use of rhTSH compared to THW, under what circumstances might THW be selected instead of rhTSH to prepare a patient for stimulated surveillance studies? There have been several reports of acute swelling of known thyroid cancer lesions after rhTSH [41, 42], presumably due to the rapid surge in TSH that resulted in clinical compromise. Thus, THW might be considered if there is a strong suspicion for residual or recurrent disease in or near critical structures. Examples include metastases to the brain or near the spinal cord or skeletal metastases in weight-bearing bones susceptible to fracture. However, swelling at such sites, particularly in patients with a history of large-volume or bulky disease, is not unique to rhTSH stimulation. It is well known to occur after endogenous TSH stimulation with THW as well. Fortunately, serum thyroglobulin and imaging studies on levothyroxine suppression are usually already positive in patients with significant residual disease, or the clinical history and past treatment record indicate patients who are at higher risk. In such cases, administration of glucocorticoids prior to stimulation with rhTSH or THW may prevent or mitigate tumor expansion [43].

Financial or logistical circumstances may favor THW over rhTSH in select situations. On a societal level, the high cost of rhTSH may be balanced or offset by the decrease in morbidity and increase in productivity (e.g., fewer missed work days) associated with avoiding prolonged hypothyroidism. Several but not all studies examining this issue have concluded that rhTSH for these reasons represents good value for money, with the benefit to patients and society obtained at modest net cost [38, 44–46]. On an individual level, however, rhTSH is not affordable to everyone. rhTSH is expensive, with costs and insurance reimbursement rates varying widely among different countries. In the USA, for example, patients without insurance or those with high out-of-pocket costs of medication may be unable to afford rhTSH injections [45].

From a logistical perspective, the extra clinic or hospital visits required for rhTSH-based preparation may not be feasible, particularly if the patient’s primary residence is located far from the site of rhTSH administration. Patients, who tolerate THW well but cannot attend four visits within a single week for rhTSH injections, WBS, and thyroglobulin measurement, may opt for THW since the scan and stimulated thyroglobulin can be obtained with only two visits [8]. Finally, while not a reason in itself to select THW over rhTSH, it should be noted that rhTSH-based surveillance is more time sensitive than THW-based withdrawal. Diagnostic dose of ¹³¹I should be given 24 h after the second rhTSH injection (with images obtained 48 h after the ¹³¹I), and measurement of stimulated thyroglobulin should be done approximately 72 h after the second rhTSH injection [8]. If the patient misses any of these critical appointment times, rhTSH must be readministered (which may not be feasible due to cost) before obtaining WBS and stimulated thyroglobulin. By contrast, patients withdrawn from thyroid hormone may undergo WBS and serum thyroglobulin measurement any time after the serum TSH reaches >30 mIU/L, allowing for greater flexibility in case of unanticipated delays. Similarly, if there is a high suspicion for residual disease necessitating a therapeutic dose of ¹³¹I, preparation with THW may be preferred, since the therapeutic dose could be administered soon after the diagnostic WBS and thyroglobulin results are available. For patients prepared with rhTSH, an additional two injections of rhTSH (or subsequent THW) would be needed to treat any disease detected on surveillance testing.

Overall, rhTSH-based preparation for stimulated surveillance studies is preferred over THW in most situations. A good understanding of the specific advantages and limitations to each approach allows for identification of patients who may benefit (or are unlikely to suffer significant harm) from THW, thus facilitating a patient-centered approach that includes informed decision making.

It is recommended that patients follow a low-iodine diet (LID) prior to whole-body radioiodine scanning for treatment or diagnostic purposes, to facilitate increased uptake of radioiodine into any remaining thyroid cells. This recommendation is based primarily on observational studies of patients who followed an LID in preparation for remnant ablation or treatment with ¹³¹I [21]. No studies have specifically examined the efficacy or safety of LID in patients undergoing surveillance WBS.

Of 8 studies examined in a systematic review, in most cases, patients were restricted to < 50 mcg of dietary iodine per day for 1–2 weeks, though the duration of the LID varied from 4 days to 4 weeks. All LIDs studied were associated with significantly lower urinary iodine excretion, as well as increase in ¹³¹I uptake into residual thyroid or tumor, compared to diets free of iodine restriction [21]. In two of the studies included in the review, using an LID for 2 weeks resulted in an approximate 50% reduction in mean urinary excretion compared to the same diet for 1 week [47, 48]. (A subsequent study published after the review paper showed that a 3-week LID did not result in any further reduction in mean urinary excretion compared to a 2-week LID [49].) Only one study in the review paper examined LID in combination with rhTSH administration; the remainder used LID in conjunction with THW [21, 47]. While urinary iodine excretion significantly decreased after 2-week LID with rhTSH, the reduction was not as robust as the same 2-week LID with THW [47]. The addition of diuretic therapy to LID did not appear to further lower urinary iodine measurements in the one study that measured urinary iodine with and without concurrent ethacrynic acid [50].

No investigators have specifically studied whether using an LID for remnant ablation or treatment improves long-term disease recurrence or mortality. There is some evidence that remnant ablation is more successful with LID, but data is conflicting and for the most part derived from retrospective analyses that used historical controls [51, 52]. In terms of safety, only one study included in the systematic review reported on side effects of LID, noting that the only complaint from participants was the “boring” nature of the diet [53]. No other adverse effects were noted. However, multiple cases of severe, potentially life-threatening hyponatremia following an LID have been reported [54, 55]. Most such cases involve elderly patients who were withdrawn from thyroid hormone, frequently in the presence of metastases to the lung or brain and in some instances treated simultaneously with thiazide diuretics. Duration of diet was also greater than 1 week in the majority of cases where hyponatremia developed [56].

In summary, despite the inconvenience of LIDs, the demonstrated reduction in urinary iodine excretion and increase in radioiodine uptake supports their use for 1–2 weeks prior to surveillance WBS. For elderly patients, particularly those with known thyroid cancer metastases, measures should be undertaken to minimize the risk of dangerous hyponatremia, including avoidance of co-administered diuretics, possible limitation of LID to 1 week, and consideration of stimulation with rhTSH instead of THW. It is also important to communicate to patients that “low-iodine” diet does not imply a “low-sodium” diet; non-iodized salt is permitted and can mitigate the risk of hyponatremia. Descriptions of low-iodine diets and general instructions can be found at several websites, including that of the American Thyroid Association (http://​www.​thyroid.​org/​low-iodine-diet/​) and the Thyroid Cancer Survivors’ Association (http://​www.​thyca.​org/​pap-fol/​lowiodinediet/​).

Finally, ascertaining exposure to high amounts of iodine in the weeks to months preceding WBS—through vitamins, dietary supplements, or iodinated contrast medium including that given during contrast-enhanced CT scan or cardiac catheterization—is particularly important. If such exposures have occurred, an LID for a few weeks will be insufficient for adequate radioiodine uptake. The WBS would need to be delayed until the exogenous iodine load has been excreted, which could be assessed by serial measurements of urinary iodine excretion or waiting at least 4–6 weeks. Although studies of patients with intact thyroid glands have shown that total body iodine stores are increased for several months following iodinated contrast exposure, a study conducted in thyroid cancer patients following thyroidectomy demonstrated that the urinary iodine excretion returned to normal by 4 weeks [57].

The cervical lymph nodes are the most common site of spread outside the thyroid gland for papillary thyroid cancer. The majority of patients with PTC have cervical node involvement at the time of diagnosis, though in many cases these are not apparent preoperatively or at the time of initial surgery and often do not change the overall patient prognosis. This is particularly true of nodal micrometastases (those <2 mm) [87–89]. Likewise, ~90% of all recurrences of PTC occur in the cervical lymph nodes; and cervical lymph node recurrences have been reported in ~30% of patients. The central neck lymph nodes are more commonly involved than those in the lateral neck, but both central and lateral neck lymph node involvements are far more common than distant metastases in PTC [90]. The prevalence of cervical lymph node metastases is much lower in follicular thyroid cancer, on the order of 10%, and more commonly with the Hürthle cell subtype of FTC [91]. As papillary cancer comprises the great majority of DTC, an imaging modality that accurately identifies central and lateral lymph node metastases is essential in the surveillance of thyroid cancer patients.

Cervical ultrasonography has emerged as the most sensitive imaging method for detecting cervical lymph node metastases. Indeed, it has become the primary structural imaging modality for patients with differentiated thyroid cancer [13, 15, 16]. Using a high-resolution probe, with frequency of ≥10 MHz, abnormal lymph nodes as small as 2–3 mm can be detected [6]. No specific patient preparation in terms of diet and laboratory values is needed. Additional advantages compared to other imaging techniques include its noninvasive nature, lack of associated exposure to radiation, and permitted use in patient populations in whom radiation-emitting or contrast-based studies may be contraindicated (including pregnant and breastfeeding patients or those with advanced kidney disease). Moreover, it is the only imaging modality that can readily be coupled with fine-needle aspiration of any identified suspicious lesions [92].

If abnormal small cervical lymph nodes <5–7 mm in short axis are detected, close observation is generally recommended, as such lymph nodes may remain stable for long periods of time [93, 94]. Further, surgical resection may fail to produce a biochemical remission in up to 73% of patients [95]. If abnormal cervical lymph nodes measuring ≥8–10 mm in size are detected, fine-needle aspiration is general recommended, assuming that confirmation of disease would lead to change in management. In 80% of cases of cervical lymph node metastases, cytology is confirmatory. The diagnostic yield is improved when measurement of thyroglobulin in the aspirate fluid (Tg-FNA) is performed in addition to cytology [96, 97]. Tg-FNA samples are typically obtained by washing 1 mL of normal saline through the needle used for fine-needle aspiration into a sterile plain tube or container, after the contents of the biopsy needle have been placed onto a slide for cytology. The ability to measure thyroglobulin in the aspirate fluid is especially advantageous in patients with circulating thyroglobulin antibodies. Unlike serum thyroglobulin measurement, the measurement of thyroglobulin in cervical lymph node fluid is not affected by circulating thyroglobulin antibodies [98]. A thyroglobulin level of 1–10 ng/mL in aspirate fluid is considered suspicious for malignancy, and a value of >10 ng/mL is highly likely to represent DTC that has spread to the aspirated lymph node [99, 100].

Features concerning for recurrence in the thyroid bed include ovoid shape of a lesion in the longitudinal plane (and taller than wide in the transverse plane), hypoechogenicity, microcalcifications, irregular borders, and increased vascularization [101]. Features most concerning for cervical lymph node metastases include cystic appearance, microcalcifications (hyperechoic punctuations), and peripheral vascularization [15]. These lymph node findings have specificities in the range of 80–100%, thereby justifying fine-needle aspiration when present individually or in combination in a lymph node measuring 8 mm or larger [102]. Lymph nodes that have lost their normal hyperechoic fatty hilum, and those that are round in shape or hypoechoic, are considered suspicious, but in the absence of other concerning features are less specific for malignancy. Thus any one of these latter findings, if present in isolation, does not automatically warrant fine-needle aspiration [13, 15, 103].

It is critical that abnormalities on neck ultrasound be interpreted in the context of the patient’s clinical picture and pretest probability of disease. In low- and intermediate-risk patients with an undetectable serum thyroglobulin, the risk of lymph node recurrence is <2% [15]. Minor abnormalities are more likely to represent false-positive findings than true disease in this population, and even in the case of true disease, the clinical significance of small recurrences <8–10 mm is unclear [104]. A substantial portion of patients in this category may experience stability or regression of the abnormal cervical nodes over time, bringing into question the benefit of intervention, and supporting a strategy of continued observation. On the other hand, abnormal lymph nodes in patients with detectable or elevated thyroglobulin are more likely to grow or become clinically significant over time, warranting at minimum attempts at diagnostic aspiration [105].

The main limitation of cervical ultrasound is its dependence on the skill of the operator. A thorough ultrasound exam must include evaluation of the thyroid bed as well as the central and lateral cervical lymph node compartments. Individuals performing ultrasound should be specifically trained and experienced in thyroid disease and pathology, including imaging of the postoperative neck, in order to approximate the high sensitivity of neck ultrasound reported in the literature [92]. Traditionally, dedicated sonographers within radiology departments performed most neck ultrasounds. However, with the rising use of neck ultrasound in thyroid cancer surveillance, and the growing recognition that abnormal findings may be best interpreted in the context of the patient’s entire clinical picture, endocrinologists and surgeons who care for thyroid cancer patients are increasingly performing their own office-based neck ultrasounds. Office-based ultrasound also allows for “real-time” manipulation of the probe, with improved ability to identify abnormalities that are concerning for recurrence and distinguish them from less concerning findings such as postoperative scar or suture-related granuloma [106]. If abnormal findings are reported from a center with less thyroid disease experience, or normal findings are reported in the setting of a high clinical suspicion for recurrence, the images should be reviewed by a center or provider with specific training in thyroid cancer and a high volume of DTC patients.