European Journal of Nuclear Medicine and Molecular Imaging

, Volume 37, Issue 12, pp 2277–2285

18F-fluorodeoxyglucose positron emission tomography and computed tomography in anaplastic thyroid cancer

Authors

  • Thomas Poisson
    • Department of Nuclear Medicine and Endocrine OncologyInstitut Gustave Roussy and University Paris-Sud XI
    • Service de Médecine Nucléaire, Hôpital Bichat
  • Désirée Deandreis
    • Department of Nuclear Medicine and Endocrine OncologyInstitut Gustave Roussy and University Paris-Sud XI
  • Sophie Leboulleux
    • Department of Nuclear Medicine and Endocrine OncologyInstitut Gustave Roussy and University Paris-Sud XI
  • François Bidault
    • Department of RadiologyInstitut Gustave Roussy and University Paris-Sud XI
  • Guillaume Bonniaud
    • Department of Medical PhysicsInstitut Gustave Roussy and University Paris-Sud XI
  • Sylvain Baillot
    • Department of EpidemiologyInstitut Gustave Roussy and University Paris-Sud XI
  • Anne Aupérin
    • Department of EpidemiologyInstitut Gustave Roussy and University Paris-Sud XI
  • Abir Al Ghuzlan
    • Department of PathologyInstitut Gustave Roussy and University Paris-Sud XI
  • Jean-Paul Travagli
    • Department of Endocrine SurgeryInstitut Gustave Roussy and University Paris-Sud XI
  • Jean Lumbroso
    • Department of Nuclear Medicine and Endocrine OncologyInstitut Gustave Roussy and University Paris-Sud XI
  • Eric Baudin
    • Department of Nuclear Medicine and Endocrine OncologyInstitut Gustave Roussy and University Paris-Sud XI
    • Department of Nuclear Medicine and Endocrine OncologyInstitut Gustave Roussy and University Paris-Sud XI
    • Service de Médecine Nucléaire et de Cancérologie EndocrinienneInstitut Gustave Roussy
Original Article

DOI: 10.1007/s00259-010-1570-6

Cite this article as:
Poisson, T., Deandreis, D., Leboulleux, S. et al. Eur J Nucl Med Mol Imaging (2010) 37: 2277. doi:10.1007/s00259-010-1570-6

Abstract

Purpose

Our aim was to evaluate in anaplastic thyroid carcinoma (ATC) patients the value of 18F-FDG PET/CT compared with total body computed tomography (CT) using intravenous contrast material for initial staging, prognostic assessment, therapeutic monitoring and follow-up.

Methods

Twenty consecutive ATC patients underwent PET/CT for initial staging. PET/CT was performed again during follow-up. The gold standard was progression on imaging follow-up (CT or PET/CT) or confirmation with another imaging modality.

Results

A total of 265 lesions in 63 organs were depicted in 18 patients. Thirty-five per cent of involved organs were demonstrated only with PET/CT and one involved organ only with CT. In three patients, the extent of disease was significantly changed with PET/CT that demonstrated unknown metastases. Initial treatment modalities were modified by PET/CT findings in 25% of cases. The volume of FDG uptake (≥300 ml) and the intensity of FDG uptake (SUVmax ≥18) were significant prognostic factors for survival. PET/CT permitted an earlier assessment of tumour response to treatment than CT in 4 of the 11 patients in whom both examinations were performed. After treatment with combined radiotherapy and chemotherapy, only the two patients with a negative control PET/CT had a confirmed complete remission at 14 and 38 months; all eight patients who had persistent FDG uptake during treatment had a clinical recurrence and died.

Conclusion

FDG PET/CT appears to be the reference imaging modality for ATC at initial staging and seems promising in the early evaluation of treatment response and follow-up.

Keywords

Anaplastic thyroid carcinoma18F-FDGPET/CT

Introduction

Anaplastic thyroid carcinoma (ATC) is a rare tumour, representing less than 5% of all thyroid carcinomas [16]. It occurs in elderly patients and is one of the most aggressive cancers in humans, with a median survival of 3 months in the absence of effective treatment [7]. Combined treatment modalities with surgery, chemotherapy and external radiation therapy permit local disease control in two thirds of patients, but a prolonged survival in less than 20% [721]. Death is more often related to distant metastases than to local disease progression. Favourable prognostic indicators for survival include age less than 60 years, small size of ATC, absence of tumour extension beyond the thyroid capsule and absence of distant metastases [5, 7, 8, 10, 11, 15, 2022]. In fact, disease control is obtained in patients in whom complete surgical resection of the tumour is feasible and who can tolerate a combination of chemotherapy and external radiation therapy. Distant metastases may occur at various sites, are poorly sensitive to cytotoxic chemotherapy and are associated with a poor outcome. Also, the combination of chemotherapy and external radiation therapy is highly toxic and should be given only when benefits are expected.

Therefore, there is a clinical need for complete initial staging and for compulsory follow-up, which is achieved with computed tomography (CT) scan, magnetic resonance imaging (MRI) and bone scintigraphy (BS) and may be more reliable with 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (FDG PET). Up to now, the benefit of FDG PET in ATC patients has been studied in only one series [23] and in a few case reports [2428]. The value of FDG PET has also been demonstrated in non-ATC for staging, prognosis and prediction of response to radioactive iodine treatment [29, 30].

Twenty patients treated at the Institut Gustave Roussy (IGR) for an ATC underwent PET/CT and total body injected CT at initial staging and subsequently during follow-up. In this study, we have retrospectively compared the respective value of these examinations for initial staging, prognosis and assessment of tumour response to treatment.

Materials and methods

Patients

From 2003 to 2008, 20 consecutive patients treated at the IGR for a histologically confirmed ATC underwent PET/CT and cervico-thoraco-abdomino-pelvic CT with contrast material during the initial staging of their disease. The mean interval of time between PET/CT and CT was 8 days.

The patients were treated with surgery when feasible and chemotherapy (six courses of a combined doxorubicin-cisplatin regimen) combined with bifractionated and accelerated radiation therapy to the neck and mediastinum performed after the second course of chemotherapy [18].

PET/CT and CT were performed again after two cycles of chemotherapy (in seven patients) and/or after radiation therapy (in six patients) and/or at the end of chemotherapy/radiation therapy (in five patients) and during subsequent follow-up (in four patients).

This study was approved by the Ethics and Scientific Review Boards at the IGR.

Imaging techniques

FDG PET/CT

All imaging and data acquisitions were performed on an integrated PET/CT Biograph LSO system (Siemens Medical Solutions, Erlangen, Germany). PET/CT scanning was performed 60–90 min after an injection of 5 MBq/kg 18F-FDG. All patients had fasted for at least 6 h and capillary glycaemia was normal in all (mean: 5.6 mmol/l, range: 3.7–7). During image acquisition, patients maintained their arms above their head if possible and no specific breathing instruction was given.

The PET elements of the system were based on a full-ring tomograph (ECAT ACCEL, CPS Innovation, Knoxville, TN, USA). Emission data were acquired for 4 min at each bed position from the top of the head to the mid-thighs. Three-dimensional mode was used for PET image acquisition. PET data were reconstructed on a 128 × 128 matrix, using an iterative algorithm [Fourier rebinning (FORE) and attenuation-weighted ordered subset expectation maximization (AWOSEM)] with 2 iterations, 8 subsets and a 5-mm full-width at half-maximum (FWHM) Gaussian postfilter. Attenuation-corrected and uncorrected data were both available for review. CT data were acquired with a single-slice spiral CT (Somatom Emotion, Siemens Medical Solutions, Erlangen, Germany) without intravenous contrast agent. CT parameters were set to 80 mAs and 110 kV, slice thickness to 5 mm and pitch to 1.5. CT data were reconstructed using filtered backprojection with a smooth filter on a 512 × 512 matrix; reconstructed slices had a thickness of 5 mm with a gap of 2.5 mm. Standard orthogonal views, as well as maximum intensity projections, were reviewed during scan interpretation.

The maximum standardized uptake value (SUVmax) and FDG uptake volume were determined for each patient. The SUVmax was automatically semi-quantified using e.Soft tools (Siemens Medical Solutions, Erlangen, Germany) dedicated to image analysis. The FDG uptake-related tumour volumes were estimated by a semiautomated image intensity threshold volume calculated through a segmentation method, using an e.Soft-dedicated three-dimensional segmentation tool (Siemens Medical Solutions, Erlangen, Germany), as previously described [31]. The largest sizes on axial slices were measured on combined CT only for morphologically visible lesions.

CT

Imaging acquisitions were performed on a multislice helical CT (LightSpeed 16, General Electric Healthcare, Buc, France) using intravenous contrast material in the absence of contraindication. The protocol was: acquisition on the thorax (arterial contrast enhancement), then on the abdomen and pelvis (contrast enhancement: portal venous in upper abdomen, venous in pelvis) and finally on the head and neck (late venous contrast enhancement with a bolus). The parameters were set to automated dose modulation, 120 kV, slice thickness 2.5 mm and pitch 1.375. CT data were reconstructed using filtered backprojection with smooth and sharp filters on a 512 × 512 matrix.

BS

BS was performed in 11 patients 2–4 h after intravenous injection of 500–700 MBq 99mTc-hydroxymethylene diphosphonate, using a large field of view dual-head gamma camera equipped with a low-energy and high-resolution collimator. Each BS was reviewed by one nuclear medicine physician (TP).

Image quantitative analysis

PET/CT images were reviewed by two nuclear medicine physicians (TP and DD) and CT images by a radiologist (FB). The review of each imaging technique was performed blindly and independently. The two nuclear medicine physicians compared their interpretations, and in cases of discrepancy a consensus was reached. An abnormal PET/CT was defined by non-physiological FDG uptake in at least one tumour site or by the detection on the attenuation-corrected CT of at least one tumour site without FDG uptake. An abnormal CT was defined by the detection of at least one tumour site. Tumour sites were considered as true positive in cases of progression, persistence on imaging follow-up (CT or PET/CT) or confirmation with another imaging modality, such as MRI. Numbers of involved organs and of lesions on PET/CT and CT were recorded. Thyroid mass, cervical lymph nodes, mediastinal lymph nodes, lungs, liver, adrenal glands, bones and abdominal lymph nodes were each considered as an “organ”. When more than 15 lesions were depicted in a given organ, 16 were considered for statistical analysis. PET/CT and CT results were compared per patient, per involved organ and per number of detected lesions. Modifications of therapeutic options based on PET/CT results were recorded.

Assessment of response to treatment with either PET/CT or CT was analysed independently. European Organization for Research and Treatment of Cancer (EORTC) recommendations and Response Evaluation Criteria In Solid Tumors (RECIST) were used for response assessment on PET/CT [32] and CT [33], respectively.

Statistics

Sensitivities of PET/CT and CT and their exact 95% confidence intervals (95% CI) were estimated per patient, per organ and per lesion. In the per organ and per lesion analyses, sensitivities of PET/CT and CT were compared using the McNemar test extended to clustered data [34]. Survival was estimated with the Kaplan-Meier method. The log-rank test was used to test the prognostic value of tumour characteristics (number of involved organs, SUVmax, functional volume) on survival in univariate analysis and Cox’s model for multivariate analysis. All reported p values are two sided. The statistical analysis was performed using SAS version 9.1 software (SAS Institute Inc., Cary, NC, USA).

Results

Patients

The clinical characteristics of the 20 patients are reported in Table 1. They were 9 men and 11 women with a mean age at diagnosis of 66 years (39–84 years). At initial staging, 16 patients had anaplastic cancer in the thyroid bed (absence of surgery in 12 cases and incomplete surgery in 4); the remaining 4 patients had undergone a complete surgical resection of the thyroid tumour 140, 28, 44 and 106 days, respectively, before the first PET/CT. Before the availability of PET/CT results, 15 patients had known distant metastases: in the lungs (12 patients), mediastinum (5), liver (3) and adrenals (1); the remaining 5 patients had local disease with thyroid tumour and involved cervical lymph nodes.
Table 1

Patients: clinical characteristics and extent of disease according to CT and PET/CT

Patient No.

Sex

Age (years)

Stage based on TNM classification according to:

Treatment

CT

PET/CT

1

M

61

IVC

IVC

Chemoradiation

2

F

66

IVC

IVC

Surgery+chemoradiation

3

F

68

IVC

IVC

Surgery+chemotherapy

4

F

64

IVC

IVC

Chemotherapy

5

M

78

IVC

IVC

Chemotherapy

6

M

63

IVC

IVC

Surgery+chemotherapy

7

M

64

IVC

IVC

Surgery+chemoradiation

8

F

68

IVC

IVC

Surgery+chemotherapy

9

F

73

NA

IVC

Radiotherapy

10

M

58

IVC

IVC

Chemoradiation

11

M

84

IVC

IVC

Chemoradiation

12

F

69

IVB

IVB

Chemoradiation

13

M

55

IVC

IVC

Chemoradiation

14

M

39

IVC

IVC

Surgery+chemoradiation

15

M

73

IVA

IVC

Surgery+chemoradiation

16

F

68

IVB

IVC

Chemoradiation

17

F

74

IVC

IVC

Chemoradiation

18

F

58

IVB

IVB

Chemoradiation

19

F

61

IVC

IVC

Surgery+chemotherapy

20

F

75

IVA

IVC

Surgery+chemoradiation

M male, F female, IVA tumour localized in the thyroid bed, IVB tumour with neck extension beyond the thyroid gland, IVC presence of distant metastases

The median follow-up after diagnosis was 21 months (6–38 months). At the end of the study, four patients were still alive, with two in complete remission at 14 and 38 months after initial diagnosis and two with metastatic disease. Sixteen patients had died after a median time of 6 months after diagnosis, 15 from disease progression and 1 from treatment-related toxicity.

Imaging

The SUVmax and volume of FDG uptake are reported in Table 2. Initial PET/CT and CT were compared in 18 patients. Two patients were excluded from this analysis because no CT was performed in one and external beam radiation therapy was initiated between PET/CT and CT in another patient.
Table 2

PET/CT data

  

Initial PET

Patient No.

Indication for the PET study

Volume of FDG uptake (ml)

SUVmax

Location SUVmax

1

IS

587

42.8

Primary tumour

2

IS

39

10.1

Mediastinal lymph nodes

3

IS

350

29.6

SVC

4

IS

343

13.2

Primary tumour

5

IS

731

38.4

Primary tumour

6

IS

57

11.9

Bone

7

IS

406

39.4

Cervical lymph nodes

8

IS

3

7.3

Cervical lymph nodes

9

IS (CT NA)

671

18.7

Liver

10

IS+R1

496

39.7

Primary tumour

11

IS+R2

392

37.6

Mediastinal lymph nodes

12

IS+R4

243

14.2

Primary tumour

13

IS+R1+R2

155

16.5

Primary tumour

14

IS+R1+R2

361

28.8

Mediastinal lymph nodes

15

IS+R1+R2+R4

43

9.4

Bone

16

IS+R1+R2+R3+R4

199

55.8

Primary tumour

17

IS+R1+R2+R3+FU+R4

309

50.9

Primary tumour

18

IS+R2+R3+FU+R4×3

145

16.4

Primary tumour

19

IS+R1+R3+FU×2

41

8.9

Lung

20

IS+R2+R3+FU×2

1

5.6

Mediastinal lymph nodes

Mean ± SD

 

279 ± 225

24.8 ± 15.8

 

IS initial staging, CT NA computed tomography with contrast material not available, R1 evaluation after two courses of doxorubicin-cisplatin, R2 evaluation after two courses of doxorubicin-cisplatin and radiotherapy, R3 evaluation at the end of treatment, FU post-therapeutic follow-up (after chemoradiation) (×2 two times), R4 evaluation of treatment during a second-line chemotherapy (×3 three times), SUVmax maximum standardized uptake value, SVC superior vena cava, SD standard deviation

Per patient analysis

All of the 18 patients had one or more lesions on PET/CT and/or on CT. Both examinations were abnormal in 17 patients. In the remaining patient with a normal CT, PET/CT demonstrated bone lesions that were also seen on BS and on MRI, but not on CT with bone window. Furthermore, in two patients CT showed only local disease, but distant lesions were seen on PET/CT (one patient in the mediastinum and one patient in the liver and bones) (Table 1). Therefore, PET/CT showed distant lesions in three patients that were not visible on CT. CT and PET/CT did not show any distant metastasis in only two patients who had local disease only.

Per organ analysis

The total number of involved organs was 63. PET/CT detected lesions in 62 (98%) of the involved organs (only 1 hepatic involvement was not seen on PET/CT) and CT in 41 (65%) of the involved organs (p < 0.001). The sensitivity of PET/CT was 100% for the diagnosis of neck (95% CI: 77–100), mediastinal (95% CI: 66–100), lung (95% CI: 74–100) and bone involvement (95% CI: 59–100). The sensitivity of CT was 57% (95% CI: 29–82) for neck (p < 0.02), 56% (95% CI: 21–86) for mediastinal (p < 0.06), 83% (95% CI: 52–100) for lung (p < 0.35) and 0% (95% CI: 0–41) for bone involvement (p < 0.01). The sensitivity of PET/CT for other lesions (liver, adrenals, abdominal lymph nodes or cardiac) was 88% (95% CI: 47–100) and that of CT was 50% (95% CI: 16–84) (p < 0.2). PET/CT detected 22 organ involvements in 18 patients that were not detected with CT (6 in bones, 6 in neck, 4 in mediastinum, 2 in lungs, 1 in liver, 1 in adrenals, 1 in spleen and 1 in heart). The involvement of these organs only detected with PET/CT was confirmed by progression in 11 cases, persistence on a subsequent PET/CT in 4 cases and by another imaging study in 7 cases.

Per lesion analysis

The total number of lesions detected in the 18 patients was 265 (Table 3). PET/CT detected 264 (99.6%) lesions (only 1 hepatic lesion was not seen on PET/CT) and CT only 165 (62%) (p < 0.002). Sensitivity of PET/CT was 100% for thyroid tumour, cervical and mediastinal lymph nodes, lung and bone metastases. Sensitivity of CT was 83% (95% CI: 59–96) for thyroid tumour (p < 0.2), 44% (95% CI: 33–55) for cervical lymph nodes (p < 0.025), 33% (95% CI: 18–52) for mediastinal lymph nodes (p < 0.03), 96% (95% CI: 90–99) for lung metastases (p < 0.23) and 0% (95% CI: 0–15) for bone metastases (p < 0.015). Sensitivity of PET/CT for other lesions (liver, adrenals, abdominal lymph nodes or cardiac) was 88% (95% CI: 47–100) and that of CT was 50% (95% CI: 16–84) (p < 0.12).
Table 3

Imaging procedures and detection of lesions according to the involved organs

Involved organ (n)

Total number of lesions

Number of lesions detected with

PET/CT and CT (%)

PET/CT only (%)

CT only (%)

Thyroid (14)

18

15 (83)

3 (17)

0

Cervical lymph nodes (14)

84

37 (44)

47 (56)

0

Mediastinal lymph nodes (9)

33

11 (33)

22 (77)

0

Lungs (12)

102

98 (96)

4 (4)

0

Bone (6)

20

0 (0)

20 (100)

0

Other (8)

8

3 (37.5)

4 (50)

1 (12.5)

Total

265

164

100

1

In the 63 involved organs, the number of lesions detected by PET/CT was superior, equal or inferior to that detected by CT in 35 cases (56%), 27 cases (43%) and 1 case (2%), respectively. Seventeen per cent of thyroid lesions, 56% of cervical lymph nodes, 77% of mediastinal lymph nodes, 4% of lung metastases, 100% of bone metastases, 33% of liver metastases and 50% of adrenal metastases were revealed only by PET/CT. Furthermore, cardiac lesions and spleen metastases were visualized only on PET/CT.

The mean size (measured on the CT of PET/CT) of all morphologically visible and metabolically active lesions in the 20 patients was 18 mm (SD: 17, range: 3–115) and the mean SUVmax was 12.2 (SD: 11.4, range: 0.5–55.8). Larger size and higher uptake on PET/CT were significantly correlated (p < 0.0001).

Among a total of 102 lung lesions, 6 with a mean size of 4 mm (SD: 0.7 mm) were seen on CT of the PET/CT but no FDG uptake was present.

Comparison of PET/CT and BS

Among the 11 patients who initially underwent BS, 3 had bone metastases. All lesions were seen on PET/CT. BS performed 2 and 17 days before PET/CT was normal in two of the three patients with bone lesions. No bone lesion was seen on CT.

Impact on therapy

PET/CT induced treatment changes in five patients (25%), indicating external radiation therapy on bone lesions in two patients and an increase of target volume for radiation therapy to the mediastinum and pulmonary hila in three patients.

Prognostic significance of initial FDG PET

The median SUVmax (higher SUVmax for each patient) and FDG uptake volume for the 20 patients were 18 (range: 5.6–55.8) and 300 ml (range: 1–731), respectively. The intensity of uptake was strongly related with functional volume (p < 0.0002). Univariate analysis showed that both SUVmax and functional volume were prognostic for survival (p < 0.015 and 0.004, respectively) (Fig. 1): 80% of patients with an SUVmax higher than 18 and only 20% of patients with an SUVmax lower than 18 died within 6 months after diagnosis; 90% of patients with an FDG uptake volume greater than 300 ml and only 10% of patients with an FDG uptake volume smaller than 300 ml died within 6 months. The number of involved organs on PET/CT had no prognostic role (p < 0.24). In a bivariate analysis, only the functional volume remained prognostic (p < 0.004).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-010-1570-6/MediaObjects/259_2010_1570_Fig1_HTML.gif
Fig. 1

Survival according to functional volume (p = 0.003) (a) and SUVmax (p = 0.014) (b)

Response to therapy and follow-up

One to four PET/CT and CT were performed again during treatment in 11 patients, the mean interval of time between PET/CT and CT being 6 days (SD: 8.5 days, range: 0–36 days). No other PET/CT was performed in the other patients who died from their disease. Results of PET/CT and CT were concordant in six patients, with partial response, stabilization and progression in one, four and one patients, respectively. These results were discordant in the other five patients: in two patients, after two courses of chemotherapy and then after radiation therapy, PET/CT demonstrated a complete metabolic response that was later confirmed, whereas CT showed only a partial response; in one patient, after two courses of chemotherapy and external radiation therapy, PET/CT showed disease progression that was confirmed by a rapid progression and death 2 months later, whereas CT showed stable disease; in one patient, PET/CT performed after two courses of chemotherapy and then after radiation therapy showed a partial metabolic response and CT stable disease, and the partial response was subsequently confirmed by CT, and the patient survived 17.5 months after initial diagnosis; finally in one patient (Fig. 2), PET/CT performed at the end of radiation therapy showed a complete metabolic response and CT a partial response, but metastases occurred 5 months later; however, this patient is still alive with disease at 20 months after the discovery of the disease.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-010-1570-6/MediaObjects/259_2010_1570_Fig2_HTML.gif
Fig. 2

Treatment monitoring with PET for patient 18

PET/CT was performed during follow-up in four patients (patients 17, 18, 19 and 20) at 4, 6, 5 and 4 months, respectively, after the end of the combined protocol of chemotherapy and radiation therapy. PET/CT was normal in the two patients who are still alive and in complete remission at 29 and 7 months after the end of treatment. PET/CT was abnormal in two patients: one who died from ATC progression 12 months after the end of treatment and one who is still alive with metastatic disease 16 months after the end of treatment.

Discussion

In the present series, 18 of the 20 ATC patients (90%) had initial distant metastases. This is higher than the 20% in our previous report [18] and may explain the lower rate of complete remission, 10 versus 23%. In fact, the presence of initial distant metastases is the main prognostic factor [5, 8, 15, 20].

Many studies have clearly shown that FDG uptake is high in patients with poorly differentiated thyroid carcinoma [29, 30, 35, 36]; however, limited data are available for ATC, and only one series [23] and a few case reports [2428] have shown that FDG uptake is high in this disease. Our study clearly confirms the high uptake of FDG in ATC, in the primary thyroid tumour, in cervical and mediastinal lymph nodes and in distant metastases.

Due to this high FDG uptake, PET/CT provides a better staging than CT and BS. PET/CT has a higher sensitivity than CT for the detection of cervical and mediastinal lymph nodes. PET/CT and CT have a similar sensitivity for the detection of lung metastases, because lung micronodules are easily visualized on CT and when they are small they frequently do not have detectable FDG uptake. In the present series, the initial extent of the disease was changed in three patients (17%) with PET/CT that demonstrated distant metastases not observed on CT. Furthermore, only one liver lesion not seen on PET/CT was seen on CT, but this patient had already other known metastases. Thus, CT did not provide any further significant information for disease staging. Cervico-thoracic CT stays however essential for the precise preoperative location and extent of the tumour regarding vessels, and for the diagnosis of the frequent venous thromboses.

Also PET/CT appears to be more sensitive than BS, and among the three patients with bone metastases who had both BS and PET/CT, all had abnormal PET/CT but two had normal BS. Thus, when PET/CT has been performed, there is no indication to routinely perform BS.

PET/CT had an important impact on therapy that was modified in 25% of our patients. This is less than in a previous series [23] in which 50% of changes were induced by FDG PET, probably because only patients with normal results of conventional imaging were selected.

The main limitation of this study, like in many other studies, is the absence of histopathological characterization of lesions seen on PET/CT and CT; however, the importance of uptake is a strong indicator of neoplastic lesions.

The intensity of uptake and the FDG uptake volume were both highly prognostic for survival in ATC patients, as already shown in differentiated thyroid cancer patients [35].

PET/CT permitted an earlier assessment of tumour response to treatment in 4 of the 11 patients in whom it was performed, and this suggests that PET/CT should be performed to monitor therapy, as this is the case for many other cancer types [37]. PET/CT has also an important prognostic value during and after treatment, as this is the case for other cancer types [3840]. The three patients with a normal PET/CT during treatment were still alive at the end of the study, 14, 20 and 38 months after the discovery of the disease, with two patients in complete remission. All eight patients who had an abnormal PET/CT during treatment had a clinical recurrence and died. This is also in close agreement with a previous report [23].

In conclusion, PET/CT provides a better staging of the disease than CT and BS and should be routinely performed during initial staging of ATC patients. When PET/CT has been performed, there is no indication to routinely perform BS or abdomino-pelvic CT, but preoperative cervico-thoracic CT stays important. Both the FDG uptake volume and the intensity of FDG uptake are prognostic for survival. PET/CT may permit an early assessment of tumour response to therapy and a reliable follow-up of patients after therapy.

Acknowledgments

This work was partially supported by the Programme Hospitalier de Recherche Clinique 2002, AOM 02 118.

Conflicts of interest

None.

Copyright information

© Springer-Verlag 2010