Introduction

¹⁸F-deoxyglucose(FDG) is an analogue of glucose and is distributed in the glucose-utilizing cells by membrane glucose transporters which are usually overexpressed in cancer cells. Unlike glucose, FDG does not undergo further metabolism and is trapped in the cell, which permits visualization by positron emission tomography(PET). FDG-PET is regarded as a very powerful tool in cancer imaging, including thyroid cancer.

The role of FDG-PET in postoperative follow-up of thyroid cancer
Up to 20% of patients with differentiated thyroid cancer will develop recurrences, and rarely some of them will eventually die from their disease. Differentiated thyroid cancer runs a relatively indolent course compared to other malignancies, nevertheless, it is a lethal illness in some cases that requires a vigorous search for recurrence. Traditionally radioiodine has been used to detect recurrence or distant metastasis of well-differentiated thyroid cancer. High dose radioiodine was used to treat recurrent or metastatic disease when the lesion showed active iodine uptake on the diagnostic whole body scan (WBS). This approach is a still valid and very helpful diagnostic and therapeutic modality, especially in high risk patients with functioning residual disease after surgery. However, recent findings, namely that the stimulated thyroglobulin (Tg) is the most important tumor marker and an undetectable Tg after surgery and remnant ablation is the accepted criteria for cure, have highlighted the diagnostic dilemma of patients with elevated Tg with negative WBS (1).
Pacini et al found that when stimulated Tg measurement was combined with neck ultrasonography, the diagnostic sensitivity was 96.3% and the negative predictive value (NPV) was 99.5%; however, for stimulated Tg and iodine WBS, the sensitivity was only 21% with a NPV of 89% (2). It is certain that modern ultrasonographic techniques are extremely sensitive and can detect small lesion at a reasonable cost. Moreover ultrasonography-guided aspiration can be done simultaneously, facilitating cytological confirmation. In this regard, ultrasonography is considered as the first line imaging modality in the follow up of low risk patients with differentiated thyroid cancer. However, ultrasonography can detect only loco-regional recurrences. Distant metastases, which occur up to 15% of the patients, are not to be found with ultrasonography and another systemic imaging modality is required which can detect lesions regardless of the site of recurrence.
Moreover during carcinogenesis, dedifferentiation occurs and thyroid follicular cells may lose their ability to concentrate iodine. In this case, iodine WBS is not helpful to detect the non-functioning lesions compounding the need for other imaging modalities. Various imaging techniques have been applied in this condition and recently FDG-PET was very successfully utilized in patients with elevated serum Tg and negative WBS (3, 4), with a very high sensitivity and specificity ranging 75 to 95%. The sensitivity and specificity of FDG-PET was superior to MIBI SPECT and I-131 post treatment scan (5) and tetrofosmin SPECT (6). FDG-PET plays a complementary role with the conventional radioiodine WBS due to the flip-flop phenomenon, which is uptake of radioiodine with no FDG uptake and vice versa (7). This means that FDG uptake is relatively increased in less well differentiated thyroid cancer which has lost the normal characteristics of thyroid follicular cells, namely iodine uptake, during cancer development and is therefore not visualized by conventional WBS. In this regard, it seems quite natural that FDG-avid metastatic thyroid cancers are resistant to treatment with high-dose radioactive iodine (8).
FDG-PET is a very sensitive and specific method to detect recurrent/metastatic disease in patient with thyroid cancer, however, it is not clear whether it should be the first-line diagnostic method during the follow up of the patients, because of concerns over the very high costs and limited availability. Recurrent disease is usually diagnosed by the finding of an elevated Tg and after the diagnosis of recurrence localization of the recurrent disease is necessary for further therapy. Usually ultrasonography is done as the first line technique in low risk patients. In low risk patients FDG-PET may be performed to detect recurrent disease elsewhere in the body when neck ultrasonography is negative or to exclude the possibility of distant metastasis when loco-regional disease is considered for surgical management.
FDG, although a good tracer, is not perfect and there is a physiologic uptake in muscle, adipose tissue, and lymphoid tissue. Moreover increased uptake may occur in various benign conditions such as inflammation. In this regard, false positive cases may occur. Most of the physiologic FDG uptake in the neck occurs in the laryngeal muscles, eye muscles, tongue and various muscles in the neck due to tension. This can be avoided by relaxing the patient or administration of a benzodiazepine before study (9). Postoperative changes including abscess and lymphoid hyperplasia or second malignancy can also be a source of false FDG uptake and any FDG uptake should be confirmed by biopsy (10, 11).
An interesting finding is that with increasing Tg levels, the detection rate of the remaining disease by FDG-PET is increased. As shown in Table 1, the scintigraphic sensitivity of FDG-PET increased as serum Tg levels increased (12, 13). Tg levels roughly correlate with tumor burden in patients with differentiated thyroid cancer (14) and it seems that FDG-PET may not detect very small tumors. Menzel et al showed that a stimulated Tg of 1.9 ng/ml was a cut-off value, above which patients had detectable lesions by FDG-PET (15). It seems clear that like many other diagnostic modalities, FDG-PET has the highest value when tumor burden is high. From the above findings, FDG-PET examination is not recommended in patients with a Tg below 10 ng/ml as an initial evaluation, due to unacceptably low sensitivity. Thyrotropin stimulation, either endogenous after thyroid hormone withdrawal, or exogenous with recombinant human thyrotropin injection, may yield better results in patients with relatively low Tg levels (16). Thyrotropin stimulation is usually recommended for higher yield, however, in patient with Tg levels above 100 ng/ml, thyrotropin stimulation may not be necessary (17).
Thyrotropin stimulation, either endogenous with thyroid hormone withdrawal, or exogenous with recombinant human thyrotropin injection, may yield better result in patients with relatively low Tg levels (16). Although some investigators reported no difference of diagnostic performance of FDG-PET according to serum thyrotropin level (4, 11), most investigators agree that thyrotropin stimulation has some influence on the FDG uptake by tumor cells (16, 18). Thyrotropin stimulates FDG uptake by human thyroid cells in vitro (19). Thyrotropin stimulation may be recommended for a higher yield in selected patients with very low Tg levels; however, in these patients with relatively low tumor burden, it has to be determined whether the added cost of recombinant human TSH or discomfort of thyroxine withdrawal is justifiable rather than adopting a wait and see policy with yearly US examination.

Table 1. Positive rate of FDG-PET uptake according serum thyroglobulin levels in differentiated thyroid carcinoma with negative radioiodine uptake

*according to reclassified thyroglobulin levels from published raw data

PET/CT
The combination of PET and CT imaging equipment seems to be very helpful for evaluation of anatomical lesions in the neck region and yielded sensitive result at low Tg levels. In one recent study, the sensitivities of FDG-PET/CT at serum Tg levels of less than 5, 5-10, and more than 10ng/ml was 60%, 63% and 72% respectively (11). In another study, PET/CT findings modified the original PET diagnoses in 77% of the patients (18). FDG-PET/CT seems promising, and, PET/CT using other tracers, such as I-123, may be also utilized, permitting individual dosimetry (17, 20).

Thyroid PET incidentaloma
Widespread use of FDG-PET in various malignant and benign diseases has disclosed some patients who have unexpected FDG uptake in the thyroid, the so-called “thyroid PET incidentaloma”. Several groups have reviewed the nature of thyroid PET incidentalomas from a large number of PET examinations (21-24). Thyroid PET incidentalomas were found in approximately 2% of subjects in all series, and the rate of malignancy ranged from 14% to 50%. Most of these were primary thyroid malignancies and some were metastatic tumors to the thyroid. Diffuse thyroid uptake was noted in patients with thyroiditis. The prevalence of malignancy in thyroid PET incidentaloma is high. Therefore, every focal FDG uptake in the thyroid has to be properly investigated and often operated on. When FDG-PET/CT was applied, focal thyroid uptake was noted in 4% of subjects and 36.7% of the cases with focal uptake were malignant (25).

The role of FDG-PET in preoperative diagnosis and staging of thyroid cancer
Some investigators reported that FDG standardized uptake values (SUVs) of PET imaging might predict malignancy and malignant tumors had higher SUVs (22, 23, 25). However, others have found that maximum SUVs in benign thyroid nodules were as high as in malignant tumors and SUVs alone were not helpful in differentiating malignancy from benign nodules in thyroid PET incidentalomas (21, 24). There are two recent prospective studies concerning the value of maximum SUVs predicting malignancy in cytologically undetermined thyroid nodules. In one, the subjects had thyroid nodules with inconclusive cytology (26) and in the other the subjects had a cytological diagnosis of follicular neoplasm (27). In both studies, SUVs were not helpful in differentiating malignant from benign nodules. However, in one study all the malignant nodules accumulated FDG, whereas only about one third of the benign nodules had FDG uptake and the remaining two thirds did not have any FDG uptake. This finding may help in differentiation between malignancy and benign nodules in patients with cytologically inconclusive results (26). In another study, all patients with a cytological diagnosis of follicular neoplasm had vivid FDG uptake and the maximum SUVs between malignant and benign nodules were similarly high (27), indicating that FDG-PET does not help to differentiate malignancy from benign thyroid nodules.
The efficacy of FDG-PET in preoperative staging of thyroid cancer has also been evaluated (28, 29). For thyroid papillary microcarcinomas, SUVs from FDG-PET imaging had a correlation with tumor size only and did not predict extrathyroidal extension: ultrasonographic findings gave better information (28). Diagnostic accuracy of FDG-PET/CT for the preoperative evaluation of the cervical lymph nodes was compared with that of ultrasonography and conventional contrast-enhanced CT (CECT). The overall sensitivity, specificity and diagnostic accuracy of FDG-PET/CT was 30%, 96% and 87%. The corresponding values for US and CECT were 41%, 97%, 89% (US) and 35%, 96%, 87% (CECT) respectively. The diagnostic value of FDG-PET/CT, ultrasonography and CECT were similar and PET/CT did not provide any additional benefit for the preoperative evaluation of patients with thyroid cancer (29).
FDG-PET may be performed as an initial work up in high risk patients. PET may alter the therapeutic approach. In one study, initial preoperative FDG-PET result was an independent prognostic factor for survival (30).

Summary
FDG-PET is a very useful imaging modality in the postoperative follow-up of the patients with differentiated thyroid cancer. When tumor recurrence is diagnosed with elevated Tg levels and diagnostic iodine WBS is negative, FDG-PET can localize recurrent/metastatic disease in 70 to 90% of the patients. However, considering the very high cost of FDG-PET examination and its availability only at some referral centers, neck ultrasonography should be done first. If ultrasonography is negative, FDG-PET or PET/CT may be the diagnostic procedure of choice. In the case of those with localized recurrent disease, FDG-PET may help excluding distant metastases. Some patients show incidental FDG uptake in the thyroid during FDG-PET or PET/CT. Since many of these thyroid PET incidentalomas are malignant, they need to be thoroughly investigated and properly managed by surgery. However, maximum SUVs by FDG-PET do not differentiate malignancy from benign thyroid nodules. FDG-PET does not have any beneficial effect in the initial preoperative evaluation of the patients with thyroid cancer, especially for the assessment of extrathyroidal extension and regional lymph node metastasis.