Bone and chest radiographs are no longer routinely obtained for patients with undetectable serum Tg concentrations. The reason is that virtually all patients with abnormal radiographs have readily detectable serum Tg concentrations.

Thyroglobulin is a glycoprotein that is produced only by normal or neoplastic thyroid follicular cells, and should not be detectable in patients who have had total thyroid ablation (Van Herle et al. 1973). Serum Tg antibodies may induce falsely negative results in immunometric assays and should be quantitatively assessed at each serum Tg determination (Spencer et al. 2005). In patients who are in complete remission after total thyroid ablation, serum Tg antibodies when initially detected will decline gradually to low or undetectable levels, with a median time of 3 years (Chiovato et al. 2003). Their persistence or their reappearance during follow-up should be considered suspicious for persistent or recurrent disease.

There is a close relationship between tumor burden and the Tg level, both during levothyroxine and following TSH stimulation; this explains why serum Tg may be undetectable in patients with isolated lymph node metastases in the neck or with small lung metastases not visible on X-rays (Bachelot et al. 2002). Also, serum Tg level may remain low in patients with recurrent disease from a poorly differentiated thyroid cancer (Volante et al. 2007).

The production of Tg by both normal and neoplastic thyroid tissue is in part TSH dependent (Schlumberger et al. 1980; Schneider et al. 1981). High serum TSH concentrations can be achieved by withdrawing levothyroxine for 4 weeks. However, the resulting hypothyroidism is poorly tolerated by some patients. This effect can be attenuated by substituting the more rapidly metabolized liothyronine for levothyroxine for 3 weeks and withdrawing it for 2 weeks (Goldman et al. 1980; Schlumberger et al. 1980; Schneider et al. 1981). The serum TSH concentration should be above an empirically determined value (>25–30 mU/L) in patients treated in this way; if it is not, withdrawal should be prolonged until it is. Intramuscular injections of rhTSH (0.9 mg for 2 consecutive days) are an alternative because levothyroxine treatment need not be discontinued; therefore, hypothyroidism is avoided, quality of life is maintained, and side effects are minimal (Haugen et al.1999; Schlumberger et al. 2007b); as shown by medico-economic evaluation, benefits of rhTSH largely outweight its cost (Borget et al. 2007). After rhTSH stimulation, the peak of serum Tg is usually obtained 3 days after the second injection. Although the magnitude of increase in serum Tg is frequently less important following rhTSH than withdrawal, the diagnostic efficiency of rhTSH for stimulating Tg production in the serum is comparable to that of levothyroxine withdrawal in most patients (Haugen et al. 1999; Schlumberger et al. 2007b; Eustatia-Rutten et al. 2004).

Modern Tg assays with a functional sensitivity of 0.1 ng/mL have an improved sensitivity for the detection of persistent disease during levothyroxine treatment RhTSH stimulation may be avoided in low risk patients with an undetectable or low serum Tg level on thyroxine treatment (< 0.3 ng ng/mL) who have no evidence of disease, but remains useful in the other patients (Schlumberger et al. 2007a; Snozek et al. 2007; Brassard et al. 2011).

The results of a ¹³¹I total-body scan depend on the ability of neoplastic thyroid tissue to take up ¹³¹I in the presence of high serum TSH concentrations, which are achieved by either withdrawing levothyroxine as previously reported, or by administering intramuscular injections of rhTSH (0.9 mg for 2 consecutive days, with radioiodine administration on the day following the second injection) (Haugen et al. 1999; Schlumberger et al. 2007b).

When ¹³¹I scanning is planned, patients should be instructed to avoid iodine-containing medications and iodine-rich foods, and urinary iodine should be measured in doubtful cases. Pregnancy must be excluded in women of childbearing age. For routine diagnostic scans, from 2 to 5 mCi (74–185 MBq) of ¹³¹I is given; higher activities may reduce the uptake of a subsequent therapeutic activity of ¹³¹I. In fact, stunning is a transient decrease of ¹³¹I uptake due to previous irradiation of thyroid cells, that may decrease NIS expression in irradiated cells (Lassman et al. 2004; Nordén et al. 2007). This suggests that exposure to a low diagnostic activity of ¹³¹I should be avoided in patients who are to be treated with ¹³¹I.

The scan is done and uptake, if any, is measured 48–72 h after the activity, preferably using a double-head gamma camera equipped with thick crystals and high-energy collimators. ¹³¹I SPECT/CT fusion imaging may provide superior lesion localization, in particular for foci located in the neck, chest, or pelvis (Wong et al. 2008; Aide et al. 2009). False positive results are rare and are usually easily recognized.

Assuming equivalent fractional uptake after administration of either a diagnostic or a therapeutic activity of ¹³¹I, an uptake too low to be detected with 2–5 mCi (74–185 MBq) may be detectable after the administration of 100 mCi (3,700 MBq). Therefore, a TBS is performed 3–7 days after the post-operative radioiodine administration (Cooper et al. 2009), and if this TBS is informative (uptake in thyroid remnants <1 %) and does not show any abnormal focus of uptake outside the thyroid bed, there is no reason to perform a less sensitive diagnostic control TBS during the subsequent follow-up, in the absence of any other abnormality (Cailleux et al. 2000; Pacini et al. 2002; Cooper at al. 2009). This is also the rationale for administering a large activity of ¹³¹I in patients with persistently elevated or increasing Tg levels (>10 ng/mL in the absence of levothyroxine treatment), even if the diagnostic scan is negative (Sherman et al. 1994; Schlumberger et al. 1997).

These tests should be performed only in selected cases and may include spiral CT (with injection of contrast medium) or MR imaging of the neck and chest, bone scintigraphy, MRI of bones, and PET scanning using ¹⁸F-FDG. Indeed, no iodinated contrast medium should be injected during the 2 months preceding radioiodine administration.

FDG-PET scanning is performed both for diagnosing remote neoplastic foci and for prognostic assessment (Leboulleux et al. 2007b; Cooper et al. 2009). The sensitivity of FDG-PET scan may be improved by TSH stimulation, obtained following levothyroxine withdrawal or injections of rhTSH (Leboulleux et al. 2009). Low-risk patients are very unlikely to require FDG-PET scanning as part of initial staging or follow-up. FDG-PET scanning in thyroid cancer is used

Inflammatory lymph nodes, suture granulomas, and increased muscle activity are common causes of false positive FDG-PET findings. Also, asymmetrical laryngeal uptake is frequently observed in patients with vocal cord paralysis. Therefore, cytologic or histologic confirmation is required before one can be certain that an FDG positive lesion represents metastatic disease.

If the total-body scan performed after the administration of ¹³¹I to destroy the thyroid remnants is informative (when uptake in normal thyroid remnants is low, < 1 %) and does not show any abnormal focus of uptake outside the thyroid bed, physical examination is performed and serum TSH and Tg are measured during levothyroxine treatment 3 months later (Table 1) (Schlumberger et al. 2004; Cooper et al. 2009). In most centers, at 9–12 months after initial treatment, a neck ultrasonography is performed and the serum Tg level is measured after thyroid hormone withdrawal or rhTSH stimulation. A diagnostic ¹³¹I total body scan is no longer routinely performed in low-risk patients with undetectable serum Tg, because the vast majority of patients with ¹³¹I uptake in their metastases also have detectable serum Tg levels (Cailleux et al. 2000; Pacini et al. 2002). Thus, ablation is currently established by an undetectable serum Tg level following TSH stimulation. Furthermore, neck ultrasonography can identify lymph node metastases that are too small to produce detectable serum Tg levels and that may be not seen on ¹³¹I total body scan either because of their small size or of the absence of radioiodine uptake as observed in a third of these patients (Frasoldati et al. 2003; Pacini et al. 2003; Torlontano et al. 2004). Tests performed at 9–12 months permit a reassessment of initial prognostication: patients with undetectable serum Tg following TSH stimulation and with normal neck ultrasonography are considered cured, the risk of long-term recurrence being less than 1 % (Cailleux et al. 2000; Pacini et al. 2002). They can be reassured and the dose of levothyroxine is decreased to maintain a serum TSH concentration within the normal range, between 0.5 and 2.5 mU/L. Clinical and biochemical evaluations with serum TSH and Tg determination on levothyroxine treatment are then performed annually; neck ultrasonography is performed in case of doubt or in high-risk patients, but any other testing is unnecessary as long as the patient’s serum Tg concentration is undetectable on levothyroxine treatment and the patient does not produce an interfering anti-Tg autoantibody; there is no indication to obtain another rhTSH stimulated Tg in these patients, at least for 5 years (Castagna et al. 2008). The use of sensitive assays for serum Tg determination permit to avoid the initial rhTSH testing at 9–12 months, in low risk patients who have undetectable or low (<0.3 ng/mL) serum Tg during levothyroxine treatment (Schlumberger et al. 2007a, b, c; Brassard et al. 2011). Such assays permit early diagnosis of persistent or recurrent disease and should indeed be used during follow-up (Snozek et al. 2007).

Suspicious abnormalities at neck ultrasonography should be submitted to FNAB, even if serum Tg remains undetectable following TSH stimulation. Lymph node metastases are indeed treated, but currently there is no evidence that treatment at a very early stage (when lymph nodes measure less than 5 mm in diameter) may improve the outcome as compared to treatment at a stage when they measure 5–10 mm in diameter. This is why only lymph node with a smaller diameter larger than 5–7 mm should be submitted to FNAB, and smaller lymph nodes should be carefully followed up.

At 9–12 months, serum Tg becomes detectable following TSH stimulation in 14–20 % of patients with no other evidence of disease. The discovery of any other abnormality will dictate specific treatments. In the absence of any other abnormalities, suppressive thyroxine treatment is maintained and another rhTSH stimulated serum Tg is obtained some months or years later, depending on serum Tg level and the clinical context. This second TSH stimulated serum Tg determination will be lower than the first one or even undetectable in the absence of any further treatment in two-thirds of patients who then will be considered cured. In fact, persistent thyroid cells may produce Tg for several months after ¹³¹I treatment and may disappear during the subsequent months. Serum Tg will remain elevated or will increase with time in the remaining third of patients who should then be submitted to an extensive work-up, which may include neck ultrasonography, chest spiral CT, bone scintigraphy, FDG–PET scanning and administration of a large activity of ¹³¹I with a TBS 3–5 days later, and clinical recurrence will likely be found in time. These data demonstrate that the trend in serum Tg level is probably more relevant than the actual serum Tg level by itself (Baudin et al. 2003).

In very low-risk papillary thyroid cancer patients who have had a total thyroidectomy but who were not given ¹³¹I postoperatively (Hay et al. 2002), the intensity of the follow-up strategy depends largely on the serum Tg level that becomes undetectable during thyroxine treatment in 95% of patients (Durante et al. 2012) and on neck ultrasonography. If despite adequate TSH suppression, the Tg is readily detectable and increases with time, an ablative ¹³¹I treatment may be necessary with a post therapy scan some days later. The follow-up protocol previously described is then applied.

In very low-risk patients who have initially undergone only a unilateral lobectomy for a single papillary thyroid microcarcinoma (Hay et al. 2008), yearly follow-up should consist of a careful neck examination and serum Tg determination during levothyroxine treatment. With time, ultrasonography is likely to show focal nodular abnormalities in the remaining lobe in most patients with detectable Tg concentrations. Usually, biopsies of these lesions can be performed under sonographic guidance, and most prove to be cytologically benign. However, if recurrent papillary thyroid carcinoma is found on biopsy, a completion thyroidectomy should be performed.

Local and regional recurrences occur in 5–20 % of patients with DTC. They are more frequently observed in patients with extensive neck disease including larger primary thyroid tumors, extension beyond the thyroid capsule, extensive lymph node involvement, and with poorly differentiated carcinomas or with particular histological features (Leboulleux et al. 2005). Recurrences in lymph nodes are more frequently observed in patients with papillary thyroid cancer; recurrences in the thyroid bed and in soft tissue are more frequently observed in patients with follicular and poorly differentiated thyroid carcinoma, and usually announce a dismal prognosis. More than a third of reoperations for persistent or recurrent disease are related to inadequate initial thyroid surgery (Kouvaraki et al. 2004). Small lymph node metastases with initial radioiodine uptake may be treated with radioiodine, but their persistence after 2 or 3 treatments should lead to surgery (Pacini et al. 1994a, b) , and it may be difficult to assess cure in these patients after several treatments with radioiodine (Bachelot et al. 2005). A recurrence that is palpable or easily visualized with ultrasonography, CT or PET scanning should be excised. Compartmental dissection of previously unexplored compartments with clinically significant persistent/recurrent disease while sparing vital structures is performed because microscopic lymph node metastases are commonly more extensive than would appear from imaging studies alone. Conversely, compartmental surgical dissections may not be feasible in the setting of compartments that have been previously explored due to extensive scarring and only a more limited or targeted lymph node resection may be possible (Cooper et al. 2009).

When radioiodine uptake is present in neoplastic foci, complete excision may be facilitated by total-body scanning 3–5 days after administration of 100 mCi (3,700 MBq) of ¹³¹I because additional tissue that should be excised may be identified. In some selected centers, surgery is performed 1 day after the total-body scan, typically using an intraoperative probe. The completeness of resection is verified 1–2 days after surgery by another total-body scan, and in one series this was achieved in 92 % of cases (Travagli et al. 1998). In the absence of detectable uptake, other methods may be used to facilitate the excision of small neoplastic foci located in scar tissue or in sites that are difficult to re-dissect, such as intra-operative ultrasound or pre-operative ultrasound-guided charcoal tattooing (Hartl et al. 2009). External radiotherapy is indicated in patients with soft tissue recurrences that cannot be completely excised and that do not take up ¹³¹I (Tubiana et al. 1985; Terezakis et al. 2008).

Some patients with papillary thyroid cancer who were not eligible for further surgery or ¹³¹I therapy have been treated for regional nodal recurrence with ultrasound-guided radiofrequency ablation or percutaneous ethanol injections (Hay et al. 2005). For tumors that invade the upper aero-digestive tract, patient outcome is related to complete resection of all gross disease, with techniques ranging from shaving tumor off the trachea or esophagus for superficial invasion with the preservation of function to more aggressive techniques in case of direct intraluminal invasion, including tracheal resection and anastomosis or laryngopharyngoesophagectomy (Gaissert et al. 2007; McCaffrey 2000). Surgery is usually combined with ¹³¹I and/or external beam radiation.