Introduction
Thyroid cancer has an incidence of 3.9 per 100 000 women and 1.5 per 100 000 men in the UK.1 This figure is slowly increasing.1 Approximately 90 per cent of cases involve differentiated thyroid carcinoma, comprising follicular and papillary types with their various subtypes. The long term prognosis for differentiated thyroid cancer is good, with 80–90 per cent 10-year survival for middle-aged adults; only 9 per cent of patients diagnosed as having differentiated thyroid cancer die from their thyroid disease.Reference Mazzaferri2 This said, 5–20 per cent of patients will suffer loco-regional recurrence and 10–15 per cent will develop distant metastases.3 Early detection of recurrent disease is associated with improved long term survival. For this reason, patients with differentiated thyroid cancer should be followed up for life.4
Follow-up investigation to detect residual disease or loco-regional recurrence consists of clinical examination with or without ultrasound imaging of the neck, and thyroglobulin assay. Following total thyroidectomy, the first indication of differentiated thyroid cancer recurrence is often a rise in the serum thyroglobulin concentration.Reference Schlumberger, Pacini, Wiersinga, Toft, Smit and Sanchez Franco5, Reference Mazzaferri, Robbins, Spencer, Braverman, Pacini and Wartofsky6 Both normal and cancerous thyroid cells secrete thyroglobulin; therefore, meaningful interpretation of thyroglobulin assay requires the patient to have had a total or near-total thyroidectomy followed by 131iodine ablation. The possibility of autoantibodies to thyroglobulin should also be excluded.
The location of the thyroid cells then needs to be elucidated. Most are within the neck, so cervical ultrasound is typically the first investigation. To identify sites of recurrence other than the neck, the thyroid carcinoma cell's ability to concentrate iodine is utilised by way of whole body iodine scintigraphy. However, 20–30 per cent of differentiated thyroid cancers do not take up and retain iodine, resulting in a negative iodine uptake scan.Reference Mirallie, Guillan, Bridji, Resche, Rousseau and Ansquer7 As thyroid cancer becomes more advanced and less differentiated, expression of the sodium iodide symporter decreases, and the ability to concentrate 131iodine therefore decreases. In patients in whom there is a strong suspicion of differentiated thyroid cancer recurrence, identified by a raised or rising thyroglobulin, but a negative iodine scan, investigative options include computed tomography (CT), bone scintigraphy, 99 mTc-sestamibi scanning, and 111In-pentertreotide and 18fluoro-2-deoxyglucose (18FDG) positron emission tomography (PET).
In recent years 18FDG-PET has been increasingly used for the detection of many types of cancer recurrence. 18Fluoro-2-deoxyglucose is an analogue of glucose labelled with the positron-emitting isotope 18F. It is transported across cell membranes by glucose transporters, phosphorylated and trapped within the glucose-metabolising cell, where it accumulates. Positron emission tomography images can be fused with CT scans to obtain accurate anatomical detail of areas of increased metabolic activity. Such a combined PET-CT approach has been shown to reduce the false positive rate by identifying normal physiological FDG uptake sites more accurately.Reference Zimmer, McCook, Meltzer, Fukui, Bascom and Snyderman8, Reference Finkelstein, Grigsby, Siegel, Dehdashti, Moley and Hall9
18Fluoro-2-deoxyglucose uptake in thyroid cancer metastases was first described in 1987.Reference Joensuu and Ahonen10 Three patterns of uptake were noted when assessed alongside 131iodine uptake. Differentiated thyroid cancer metastases either concentrated 131iodine but not FDG, concentrated both 131iodine and FDG, or concentrated FDG but not 131iodine. As differentiated thyroid cancer becomes less differentiated, its metabolic activity (reflected in FDG uptake) increases.Reference Wang, Larson, Fazzari, Tickoo, Kolbert and Sgouros11 This leads to the ‘flip-flop’ phenomenon, whereby whole body 131iodine scintigraphy uptake scanning becomes negative as FDG-PET scanning becomes positive.Reference Joensuu and Ahonen10, Reference Khan, Oriuchi, Higuchi, Zhang and Endo12–Reference Feine, Lietzenmayer, Hanke, Wohrle and Muller-Schauenburg14
Numerous centres have assessed the use of FDG-PET scanning for the investigation of differentiated thyroid cancer.Reference Finkelstein, Grigsby, Siegel, Dehdashti, Moley and Hall9 These studies all have one or more severe limitations, including: patient cohorts with large or primary tumour bulk as opposed to small recurrences; patients whose recurrences or metastases had already been detected by clinical examination or whole body iodine scintigraphy or other anatomical imaging; small patient sample size; concordance with whole body iodine scintigraphy counting as a positive outcome; classification of negative whole body iodine scintigraphy and negative PET as ‘no disease’ despite elevated thyroglobulin levels; and lack of evidence that the site identified by FDG-PET was the true site of thyroid malignancy.
As FDG-PET is relatively costly, it is a second line investigation in the UK. In clinical practice, it is usually only requested in cases in which recurrence is thought likely due to rising thyroglobulin levels but the location has not been elucidated by ultrasonography or whole body iodine scintigraphy. It is in this cohort of patients that we assessed the role of FDG-PET in the follow up of differentiated thyroid cancer with presumed recurrence but negative whole body iodine scintigraphy scanning.
Materials and methods
This retrospective study was completed using data collected from two sites over a six-year period (February 2002 to February 2008). All combined PET-CT scans from The Hammersmith and Charing Cross Hospitals, London (Imperial College Hospitals National Health Service (NHS) Trust) were carried out at the Hammersmith Hospital. Positron emission tomography scans requested by St Mary's Hospital, London (NHS Trust) were carried out by Alliance Medical Imaging Centres, London. Scan reports containing the words thyroid, papillary, Hürthle, follicular or anaplastic were retrieved, along with all other histology reports, radiology reports and clinical notes for these patients. We included in our analysis data from patients in whom the indication for PET scan was follow up of differentiated thyroid cancer. Patients from both the involved sites had all been discussed at the same thyroid multidisciplinary team meetings.
A PET scan was deemed positive if it identified a site of increased metabolic activity consistent with differentiated thyroid cancer recurrence or metastasis, and negative if activity was reported as being within normal physiological limits. A rising thyroglobulin concentration was taken to be the ‘gold standard’ marker for differentiated thyroid cancer recurrence. From this data, sensitivity values were calculated. The scans were all conducted in patients with strong evidence of differentiated thyroid cancer recurrence. This precluded the calculation of specificity, positive predictive value and negative predictive value, as there were no scans performed on patients without the disease.
A comparison was made with other studies, which were identified by a search of the PubMed electronic database using the keywords ‘PET’ or ‘positron emission tomography’ and ‘thyroid’. The search was carried out in June 2008. Review articles and reference lists were used to identify other studies. Eligible studies were those purporting to assess the use of 18FDG-PET in patients with recurrent thyroid cancer and a negative iodine uptake.
Results and analysis
Eighteen 18FDG-PET scans on 13 female and five male patients were identified as being undertaken for the follow up of differentiated thyroid cancer. Two patients had follicular carcinoma and 16 had papillary carcinoma. The age range at the time of scanning was 19–77 years (mean 53 years). All scans were conducted to elucidate the site of differentiated thyroid cancer recurrence, as evidenced by rising thyroglobulin concentrations with negative thyroglobulin autoantibodies. Fourteen of the patients had had negative iodine whole body iodine scintigraphy. The remaining four patients had undergone whole body iodine scintigraphy scans which were felt to be equivocal in the clinical setting, and so further imaging in the form of PET had been carried out. The results are shown in Figure 1.
Number of patients with positive and negative whole body iodine scintigraphy (WBS) scan and positive and negative 18fluoro-2-deoxyglucose positron emission tomography (PET) scan results.
Eight of the 18 18FDG-PET scans (44.4 per cent) showed positive uptake, whilst 10 (55.6 per cent) were negative. Of the 14 patients with a negative whole body iodine scintigraphy scan and rising thyroglobulin levels, six had a positive 18FDG-PET scan, giving a sensitivity of 42.9 per cent. Two of the four patients with equivocal whole body iodine scintigraphy had positive 18FDG-PET scans.
Histological analysis results were only available in four patients who had undergone surgery for whole body iodine scintigraphy negative, 18FDG-PET positive disease. In all four cases, the resection specimen contained differentiated thyroid cancer, and in all cases the thyroglobulin level fell post-operatively. In the remaining two cases of whole body iodine scintigraphy negative, 18FDG-PET positive disease, a multidisciplinary team meeting concluded that surgery was either not possible or not in the patient's best interest. For this reason, definitive histological results were not available.
The first 18FDG-PET positive case with equivocal whole body iodine scintigraphy results had possible superior mediastinal recurrence in addition to iodine uptake in lung nodules. 18Fluoro-2-deoxyglucose positron emission tomography confirmed increased metabolic activity in the mediastinum but showed no suspicious pulmonary areas. Superior mediastinal clearance showed differentiated thyroid cancer on histological analysis. Post-operatively, the thyroglobulin concentration fell to an undetectable level.
The second patient (a 20-year-old woman) underwent surgery to clear presumed pathology from the superior mediastinum, on the basis of rising thyroglobulin levels, iodine uptake and positive PET of that area. Histology of the specimen showed normal thymic tissue and no evidence of differentiated thyroid cancer recurrence.
No surgery was carried out, and hence no histological analysis was undertaken, in patients with whole body iodine scintigraphy negative, 18FDG-PET negative disease.
Discussion
Positron emission tomography is an established tool in the follow up of other cancer subtypes. A recent meta-analysis of the role of PET in the follow up of head and neck squamous cell carcinoma showed that it had a sensitivity of 94 per cent (95 per cent confidence interval (CI) 87–97 per cent) and a specificity of 82 per cent (95 per cent CI 76–86 per cent).Reference Isles, McConkey and Mehanna15 The sensitivity of 18FDG-PET for detecting differentiated thyroid cancer recurrence varies widely, from 45 per cent in one study to 100 per cent in the current case series (Table I).Reference Alnafisi, Driedger, Coates, Moote and Raphael16–Reference Wang, Macapinlac, Larson, Yeh, Akhurst and Finn32 A recent review of PET scans by the UK NHS Health Technology Assessment agency concluded that 18FDG-PET had a specificity of 25–83 per cent in detecting differentiated thyroid cancer recurrence.Reference Facey, Bradbury, Laking and Payne33 Their assessment of combined PET-CT scans, limited to two studies, calculated the sensitivity to be approximately 66 per cent.
Studies of 18FDG-PET sensitivity for detection of differentiated thyroid cancer recurrence in thyroglobulin-positive, whole body 131iodine scintigraphy negative patients
18FDG-PET = 18fluoro-2-deoxyglucose positron emission tomography; pts = patients
Calculation of sensitivity requires that the presence of disease be confirmed by a gold standard. Typically, researchers have used a rising thyroglobulin level, positive iodine whole body scan, or tumour bulk found on clinical or radiological examination. Overall, all the series are limited by a lack of histological proof. The actual method of confirmation of disease is rarely mentioned in the literature. As previously mentioned, the case mix is heterogeneous in most series, with both whole body iodine scintigraphy positive and negative cases being included in the same series.
Another common assumption is that patients who are scan-negative (i.e. whole body iodine scintigraphy and 18FDG-PET) despite an elevated thyroglobulin level, should be characterised as having no disease recurrence. The lack of an identified source of thyroglobulin cannot exclude differentiated thyroid cancer recurrence. The indolent course of differentiated thyroid cancer prevents the conclusion that an elevated thyroglobulin concentration in the absence of identifiable recurrence is a true negative scan; the recurrence may manifest itself a few months later.
It is noticeable that the studies suggesting a very high sensitivity of 18FDG-PET were all of small case series published in the late 1990s, when PET was first becoming available. It is likely that, since this time, 18FDG-PET scans have become more widely used, and have been employed in patients with smaller disease volume. For these reasons, combined with the natural regression towards the mean, the sensitivity has decreased to a more realistic level for the actual clinical scenario for which it is employed.
A retrospective study is usually thought of as inferior to a comparable, prospective series. The main reason for this is the lack of availability of case notes and results, and the fact that retrospective studies are subject to selection bias. It is this potential selection bias which allows us to assess the use of 18FDG-PET in the ‘real world’. It is still an expensive investigation, which in the clinical setting is used only when other imaging modalities (such as ultrasound, CT and whole body iodine scintigraphy) fail to elucidate the site of differentiated thyroid cancer recurrence. It is this cohort which interests us, not patients randomly assigned to undergo 18FDG-PET or not.
Histological proof of either differentiated thyroid cancer recurrence or lack thereof would allow for a more robust assessment of 18FDG-PET as a tool in the detection of differentiated thyroid cancer recurrence. This was not feasible, and is unlikely ever to be so. The indolent course of the disease means that recurrence in surgically inaccessible sites such as vertebrae or the inferior mediastinum do not warrant excision in all cases. Proving the absence of disease is even harder. Post-mortems would have to include a cell-by-cell analysis of all the tissue – an unlikely prospect.
• Most recurrences of differentiated thyroid carcinoma are detected by rising thyroglobulin levels and can be diagnosed by clinical examination, anatomical imaging or iodine uptake scans
• However, a significant cohort of patients develop recurrence but do not concentrate iodine
• This group of iodine-negative patients may benefit from 18fluoro-2-deoxyglucose positron emission tomography (18FDG-PET) scanning to elucidate the site of recurrence
• Previous studies have shown 18FDG-PET to have a high sensitivity in detecting the site of thyroid carcinoma recurrence; however, these studies often included cases of iodine-positive disease – not a true clinical picture
• This study, the largest reported from the UK, showed 18FDG-PET to have a sensitivity of 43 per cent in detecting differentiated thyroid carcinoma recurrence, as evidenced by rising thyroglobulin levels, in iodine uptake scan negative patients
Specificity has been harder to quantify, as 18FDG-PET (with or without CT) is used in a cohort of patients who are all thought to have disease; there are few patients with no evidence of differentiated thyroid cancer recurrence who undergo a 18FDG-PET scan.
This study assessed a very distinct group of patients with suspected recurrence of differentiated thyroid cancer, identified by rising thyroglobulin levels, and a negative or equivocal whole body iodine scintigraphy scan. 18Fluoro-2-deoxyglucose positron emission tomography proved a useful tool in the localisation of differentiated thyroid cancer recurrence, leading to a successful operation and normalisation of thyroglobulin levels in five out of 18 cases (27.8 per cent). The sensitivity of 18FDG-PET (with or without CT) in our series (42.9 per cent) was lower than in previously published series. This was mainly due to 18FDG-PET only being requested in patients in whom other imaging modalities were negative. The lower sensitivity is likely to give a more realistic indication of the ability of 18FDG-PET to detect differentiated thyroid cancer recurrence in this more clinically relevant cohort.
