European Journal of Nuclear Medicine and Molecular Imaging

, Volume 33, Issue 11, pp 1273–1279

Improved tumour detection by gastrin receptor scintigraphy in patients with metastasised medullary thyroid carcinoma

Authors

    • Department of Nuclear MedicineRadboud University Nijmegen Medical Center
  • Martin P. Béhé
    • Department of Nuclear MedicinePhilipps-University of Marburg
  • Daniela Beuter
    • Department of Nuclear MedicinePhilipps-University of Marburg
  • Anke Battmann
    • Department of Diagnostic RadiologyPhilipps-University of Marburg
  • Artur Bauhofer
    • Institute of Theoretical SurgeryPhilipps-University of Marburg
  • Tino Schurrat
    • Department of Nuclear MedicinePhilipps-University of Marburg
  • Meike Schipper
    • Department of Nuclear MedicinePhilipps-University of Marburg
  • Halina Pollum
    • Department of Nuclear MedicinePhilipps-University of Marburg
  • Wim J. G. Oyen
    • Department of Nuclear MedicineRadboud University Nijmegen Medical Center
  • Thomas M. Behr
    • Department of Nuclear MedicinePhilipps-University of Marburg
Original article

DOI: 10.1007/s00259-006-0157-8

Cite this article as:
Gotthardt, M., Béhé, M.P., Beuter, D. et al. Eur J Nucl Med Mol Imaging (2006) 33: 1273. doi:10.1007/s00259-006-0157-8

Abstract

Purpose

Radiopeptide imaging is a valuable imaging method in the management of patients with neuroendocrine tumours (NET). To determine the clinical performance of gastrin receptor scintigraphy (GRS), it was compared with somatostatin receptor scintigraphy (SRS), computed tomography (CT) and 18F-FDG positron emission tomography (PET) in patients with metastasised/recurrent medullary thyroid carcinoma (MTC).

Methods

Twenty-seven consecutive patients underwent imaging with GRS, SRS (19 patients), CT and PET (26 patients). GRS and SRS were compared with respect to tumour detection and uptake. CT, PET, magnetic resonance imaging (MRI), ultrasound (US) and follow-up were used for verification of findings. In addition, GRS, CT and PET were directly compared with each other to determine which method performs best.

Results

Nineteen patients underwent both GRS and SRS. Among these, GRS showed a tumour detection rate of 94.2% as compared to 40.7% for SRS [mean number of tumour sites (±SD) and 95% confidence intervals (CI): GRS 4.3±3.1/2.8–5.7, SRS 1.8±1.6/1.1–2.6]. In 26 patients, GRS, CT and PET were compared. Here, GRS showed a tumour detection rate of 87.3% (CT 76.1%, PET 67.2%; mean number of tumour sites and 95% CI: GRS 4.5±4.0/2.9–6.1, CT 3.9±3.5/2.5–5.3, PET 3.5±3.3/2.1–4.8). If GRS and CT were combined, they were able to detect 96.7% of areas of tumour involvement.

Conclusion

GRS had a higher tumour detection rate than SRS and PET in our study. GRS in combination with CT was most effective in the detection of metastatic MTC.

Keywords

PeptideScintigraphyThyroid cancer

Introduction

Somatostatin receptor scintigraphy (SRS) is a standard procedure for the detection of neuroendocrine tumours (NET) and their metastases and is considered the diagnostic gold standard in several tumour entities [13]. The method is based on binding of radiopeptides to the somatostatin receptor over-expressed mainly by tumour cells [4]. As a result of the introduction of residualising labels [5], image quality could be improved significantly. Radiopeptide imaging may have an influence on patient management in as many as 25–50% of cases [6, 7]. In medullary thyroid carcinoma (MTC), SRS is less effective, in part due to an inverse relation of the grade of differentiation of MTC tumour cells and somatostatin receptor expression [8]. Unlike in NET of the gut, in MTC 18F-FDG positron emission tomography (18F-FDG PET), computed tomography (CT) and magnetic resonance imaging (MRI) show a better sensitivity than SRS [913].

Although calcitonin sampling after intravenous application of pentagastrin is the most sensitive method for establishing recurrent MTC [14], this test does not localise metastases. Given the outstanding sensitivity of biochemical pentagastrin testing, it was hypothesised that a radiolabelled peptide binding to the same receptor as pentagastrin would be able to detect metastases of MTC. Subsequently, such a compound was developed and the results of preclinical and preliminary clinical testing indicated that a gastrin-based scintigraphic approach holds promise as a useful clinical imaging modality [15].

In this study, we compared the results of gastrin receptor scintigraphy (GRS), SRS, CT and 18F-FDG PET in a series of patients with metastasised MTC.

Materials and methods

Patients and study design

A series of consecutive patients who underwent staging for histologically verified MTC between July 2001 and July 2003 was included in the study. All patients had an elevated serum calcitonin level and all underwent GRS, CT and 18F-FDG PET. In addition, SRS, MRI and ultrasound (US) were performed to verify findings of one of the other imaging procedures that could not otherwise be confirmed. All imaging studies had to be performed within 21 days to assure comparability. No interventions or therapies were done between the scans. Single-photon emission computed tomography (SPECT) images were not included in the study as complete SPECT scanning of all patients (thorax, abdomen, pelvis) had not been performed. All patients had consented to participation in the study and to the use of their data for scientific purposes. The study had been approved by the local ethics committee of Philipps-University, Marburg.

Imaging protocols

Somatostatin receptor scanning and gastrin receptor scanning were both performed using the same protocol: 4 and 24 h after the injection of 150–200 MBq 111In-DPhe1-DTPA-octreotide (OctreoScan, Tyco Healthcare, Neustadt/Donau, Germany) or 111In-DTPA-DGlu1-minigastrin (Bachem, Weil am Rhein, Germany), anterior and posterior planar whole-body images were obtained using a Siemens dual-head gamma camera (E CAM, Siemens, Hoffman Estates, Illinois, USA) equipped with medium-energy parallel-hole collimators (scanning speed 7 cm per minute) [15]. The radiochemical purity of the labelled minigastrin was always >95% as evaluated by high-performance liquid chromatography with a specific activity of 55.5 GBq/μmol.

For PET scanning, a Siemens dedicated PET scanner (E CAT, Siemens, Hoffman Estates, Illinois, USA) was used. Patients were fasted for at least 4 h. Prior to the examination, patients were hydrated with 500 ml of water. Emission (scanning time 20 min) and transmission scans were obtained 60 min p.i. of 350 MBq 18F-FDG, including the skull base and the proximal lower extremities. The images were corrected for attenuation and reconstructed using the ordered subsets expectation maximisation (OSEM) algorithm. The slice thickness was 4 mm. The reconstructed images were displayed in coronal, transverse and sagittal planes.

For CT, a Siemens Somatom 2 was used. Slices of 3–8 mm were obtained at 4.5- or 12-mm increments from the skull base to the jugulum and from the lung apex to the pubic symphysis, respectively. Non-ionic contrast medium (Iopamidol 300; Solutrast, Bracco-Byk Gulden, Konstanz, Germany) was administered intravenously with an automated power injector (70/120 ml volume, flow rate 2–3 ml/s). Hard copies of the scans were optimised for visualisation of soft tissues (L 50–60, W 250–400) and lung parenchyma (L 600, W 1,600).

MRI scans were acquired in a 1.0-Tesla clinical scanner (Magnetom Expert, Siemens, Erlangen, Germany) with commercially available gradients capable of a 1200 μs rise time and 20 mT/m maximum gradient strength using a body array coil. We used a T2-weighted axial turbo spin echo sequence (TR/TE 2,730/138 ms, slice thickness 6 mm), an axial T1-weighted fast low angle shot gradient sequence (TR/TE 134/6 ms, flip angle 70°, slice thickness 6 mm) prior to and after injection of contrast medium [Gd-DTPA (Magnevist), 0.1 mmol/kg, flow 1 ml/s] and coronal true fast imaging with a steady state precession sequence (TR/TE 10.22/4.7 ms, slice thickness 5 mm).

Evaluation of scans

SRS, GRS, CT and FDG PET were evaluated in a blinded manner. Other imaging procedures (MRI, US) and follow-up served for verification of findings. The standard for comparison of the different methods was defined as all sites of tumour involvement that could be visualised by at least two imaging methods (GRS, SRS, CT, PET, MRI, US) or follow-up according to the method described by Juweid and co-workers [16]. By comparing the results of GRS with those of all other imaging modalities, the chance of missing a true positive lesion owing to low sensitivity of only one imaging method as standard was kept low. Furthermore, the chance of false positive results was also limited by comparing GRS with established imaging modalities. Histology is the best standard for evaluation of the results of scans. However, it could not be performed in our study owing to the high number of lesions. Therefore, specificity is not given.

All images were evaluated by two independent observers who were experienced specialists in either nuclear medicine or diagnostic radiology. They were blinded with respect to patient identity, extent of disease and results of other imaging procedures. For all imaging techniques, the localisation of foci was regionally defined: skeleton, head, cervical region (left/right), supraclavicular area (left/right), lung (left/right), hilus (left/right), mediastinum, liver (right lobe/left lobe), peritoneal cavity, retroperitoneum and soft tissues of the pelvis. In cases of disagreement between the observers, the final judgment on the number of lesions was settled by consensus.

Octreotide uptake was defined using a modification of the score of Krenning and co-workers (1, lower than liver uptake; 2, equivalent to liver uptake; and 3, higher than liver uptake) [1]. As minigastrin uptake into the liver is absent and uptake into the thyroid is comparable to somatostatin uptake in the liver, the thyroid uptake served as the reference for a minigastrin score equivalent to the Krenning score for octreotide (1, lower than thyroid uptake; 2, equivalent to thyroid uptake; and 3, higher than thyroid uptake). If patients had previously been thyroidectomised, a standard scan was used for comparison. Uptake into the thyroid was (subjectively) determined in comparison to abdominal and thoracic background in the standard scan. This was compared with the background in the respective patient scans and subsequently with tumour uptake. To assess smaller differences in uptake, intermediate scores of 0.5, 1.5 and 2.5 were also allowed.

The studies were evaluated twice. First, an evaluation was done for direct comparison of GRS with the established method of SRS. In the second evaluation, knowing that SRS did not perform well, GRS, PET and CT were compared to determine the optimal imaging protocol for metastatic MTC.
  1. 1.

    Comparison of GRS and SRS: Besides assessment of the localisation and number of abnormalities average, uptake scores for GRS and SRS were derived from the number of positive regions multiplied by the uptake score. The standard for comparison was all sites of tumour involvement positive on GRS or SRS that could be confirmed by any other imaging method.

     
  2. 2.

    Comparison of GRS with CT and FDG PET: In addition to assessment of the number and localisation of sites of lesions as described above, regional localisation of bone and bone marrow metastases were more precisely described (skull, upper limbs, cervical/thoracic/lumbar spine, pelvis, lower limbs). The standard with which GRS, CT and PET were compared was all sites of tumour involvement detected by at least two of the methods as described above.

     

Statistical evaluation

Descriptive statistics were used. A tumour detection rate is given for direct comparison of the different imaging techniques together with 95% confidence intervals. This approach was chosen as no technique needs to be compared with an inferior standard and the results are reliable even without histological evaluation of every single tumour [16]. The correlation of the results of scintigraphy (positive/negative) with calcitonin and carcinoembryonic antigen (CEA) levels was determined by Pearson’s correlation coefficient.

Results

Patients

From July 2001 until July 2003, 27 consecutive patients (12 female, 15 male; mean age 45±14 years) with MTC were included. Five patients had occult disease while in all others, metastases were already known. A subgroup of 19 patients (nine female, ten male; mean age 46.4±11 years) also underwent SRS.

Side-effects after injection of the radiolabelled minigastrin were similar to those observed in pentagastrin testing (i.e. nausea, flush, tickling, hypotension). Usually, these symptoms were mild. In only one patient, more severe side-effects were observed (severe flush and nausea). This patient recovered quickly.

Direct comparison of somatostatin and gastrin scans

GRS detected 94.2% of sites of tumour involvement, with an average tumour uptake score of 1.9±0.8, while SRS detected 40.7%, with a tumour uptake score of 0.6±0.5. The mean number of tumour sites, the standard deviation and the 95% confidence intervals (CI) were 4.3±3.1/2.8–5.7 for GRS and 1.8±1.6/1.1–2.6 for SRS. There was no overlap in CI of SRS and GRS. In all patients but one, GRS achieved a higher overall score than SRS (Table 1). For GRS, in 37% of the sites of tumour involvement the observers had to reach a consensus; in 18.8% of these cases this concerned the localisation. A patient example is shown in Fig. 1.
Table 1

Patient data, calcitonin and CEA levels and results of GRS and SRS

Pat. no.

Age (yrs)

Sex

Calcitonin (pg/ml)

CEA (ng/ml)

GRS: no. of regions

GRS: avg. uptake

SRS: no. of regions

SRS: avg. uptake

Overall score: GRS/SRS

1

36

m

44,746

523

7

2.6

3

0.6

18.2/1.8

2

48

f

216,700

1,181

0

0.0

5

2.0

0/10

3

48

m

3,447

33

2

1.0

0

0.0

2/0

4

51

m

1,871

802

12

1.7

0

0.0

20.4/0

5

57

m

1,478

1

6

1.7

3

0.5

10.2/1.5

6

33

f

186,449

401

4

2.5

2

0.5

10/1

7

32

f

8,275

288

3

1.3

1

0.3

3.9/0.3

8

62

m

26,300

871

4

2.5

2

0.5

10/1

9

35

m

1,123

2,248

9

3.0

5

0.6

27/3

10

61

f

3,420

6,069

5

1.8

1

0.3

9/0.3

11

43

f

8,239

224

1

0.7

0

0.0

0.7/0

12

63

m

1,985

3,276

1

2.0

1

1.0

2.0/1.0

13

39

m

15,668

510

3

2.7

0

0.0

8.1/0

14

47

f

396

4

2

2.0

2

1.0

4.0/2.0

15

58

m

118,700

3

7

2.6

3

0.7

18.2/2.1

16

39

f

8,960

36

5

2.0

3

1.2

10.0/3.6

17

59

f

2,400

138

3

2.3

1

0.7

6.9/0.7

18

37

m

56,548

376

6

2.2

3

0.7

13.2/2.1

19

34

f

210

2

1

2.0

0

0.0

2.0/0

m male, f female

https://static-content.springer.com/image/art%3A10.1007%2Fs00259-006-0157-8/MediaObjects/259_2006_157_Fig1_HTML.gif
Fig. 1

Patient with metastases of MTC (calcitonin 210 pg/ml, CEA 2 ng/ml). CT and PET were negative. GRS shows two soft tissue metastases (a, arrows), while SRS shows no abnormalities (b)

One patient (no. 2 in Table 1) who was negative on gastrin scanning also had a negative pentagastrin test despite a basal calcitonin level of 216,700 pg/ml. Pentagastrin testing had also been negative ever since the diagnosis of MTC, even after tumour masses had been detected. Furthermore, this patient had never shown the usual side-effects of pentagastrin injection, such as nausea etc. As shown in Fig. 2, SRS was positive.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-006-0157-8/MediaObjects/259_2006_157_Fig2_HTML.jpg
Fig. 2

Negative GRS (a) and clearly positive SRS (b) in a patient with proven metastatic MTC despite negative pentagastrin testing, indicating that CCK2 receptors are missing or unable to bind the gastrin agonists

No correlation could be established between minigastrin uptake and either calcitonin level (r=−0.198, p=0.43) or CEA level (r=0.032, p=0.89). Somatostatin uptake showed an intermediate correlation with calcitonin level (r=0.519, p=0.027) while there was no correlation with CEA level (r=−0.02, p=0.938) [17].

Evaluation of GRS versus CT and FDG PET

A total of 132 tumour sites were detected in the 26 patients with MTC. The number of tumour sites differed in respect to the first part of the study because of the different standard used (all sites of tumour involvement detected by two methods versus sites that could be detected by SRS or GRS only, confirmed by any other method). Patient no. 19 had not been examined by PET and was thus not considered in this part of the study. GRS detected 115 (87.3%), CT detected 100 (76.1%) and PET detected 90 (67.2%) sites of tumour involvement. The combination of CT and GRS revealed 96.7% of all known sites of tumour involvement (Table 2). The mean number of tumour sites detected by the methods, the standard deviation and the 95% CI were: GRS 4.5±4.0/2.9–6.1, CT 3.9±3.5/2.5–5.3, PET 3.5±3.3/2.1–4.8. An example is shown in Fig. 3.
Table 2

Results of GRS, CT and 18F-FDG PET

No.

Type of MTC

Occult

Total no. of regions

GRS

PET

CT

1

Sporadic

No

4

4

3

2

2

Sporadic

No

5

0

5

5

3

Sporadic

No

2

2

0

2

4

Sporadic

No

10

7

5

7

5

Sporadic

No

7

7

3

6

6

Sporadic

No

4

4

3

4

7

Sporadic

No

5

3

1

3

8

Sporadic

No

6

6

4

5

9

MEN

No

16

16

12

15

10

Sporadic

No

12

10

11

5

11

Sporadic

No

0

0

0

0

12

Sporadic

No

9

9

5

9

13

Sporadic

No

5

3

2

4

14

Sporadic

No

5

5

5

3

15

Sporadic

No

4

4

3

4

16

Sporadic

No

3

2

2

1

17

Sporadic

Yes

8

8

4

6

19

Sporadic

No

1

1

0

0

20

Sporadic

No

5

5

2

3

21

MEN

Yes

0

0

0

0

22

Sporadic

No

11

10

9

9

23

MEN

No

6

6

6

6

24

MEN

Yes

0

0

0

0

25

MEN

Yes

0

0

0

0

26

MEN

Yes

0

0

0

0

27

Sporadic

No

6

5

5

3

Total:

134

117

90

102

Percentage:

 

87.3%

67.2%

76.1%

MEN multiple endocrine neoplasia

https://static-content.springer.com/image/art%3A10.1007%2Fs00259-006-0157-8/MediaObjects/259_2006_157_Fig3_HTML.gif
Fig. 3

Patient with metastatic MTC. While GRS (a) showed high uptake into a metastasis in the thoracic spine and several other metastases (arrows), FDG uptake was low (b). Several GRS-positive metastases were negative on PET. The CT scan (c) showed the metastasis in a thoracic vertebral body. SRS (d) also showed uptake in the metastasis in the spine. However, the other lesions were not visualised despite the fact that the image was adjusted to high contrast and set very dark

A number of tumours that could not be confirmed otherwise were detected by only one of the methods (single lesions: GRS 80, CT 51, PET 33). These lesions were not taken into account for the evaluation of data.

Discussion

Owing to its significant impact on patient management, imaging with radiolabelled octreotide is a very valuable diagnostic technique in clinical practice [6, 7]. However, in patients with MTC, the sensitivity of SRS is limited. Since the successful introduction of SRS with 111In-DTPA-DPhe1-octreotide into clinical practice, no other radiopeptides binding to receptors other than somatostatin receptors have demonstrated clear advantages over 111In-octreotide. Several peptides binding to other receptors have undergone preclinical and clinical evaluation, but none of these have found their way into clinical practice so far [1823].

In our study, 111In-DTPA-DGlu1-minigastrin was directly compared with 111In-DTPA-DPhe1-octreotide in patients with MTC. GRS showed a higher tumour detection rate than SRS. One might argue that SRS is at a disadvantage because most of the patients suffered from metastatic disease, probably associated with tumour dedifferentiation. However, GRS even performed better in patients with high calcitonin levels, which are associated with better performance of SRS [17]. In MTC, CCK2 receptors therefore seem to be preserved even in cases of tumour dedifferentiation. Thus, in patients with MTC, GRS may become the scintigraphic imaging modality of choice. The combination of CT and GRS was the most effective approach to the localisation of metastatic MTC in this study.

In only one of six patients with occult MTC, a previously unknown metastasis could be detected by GRS. Thus, it is questionable whether GRS will contribute to the localisation of occult metastases from MTC suspected on the basis of positive pentagastrin testing. If diffusely spread tumour cell clusters are responsible for positive pentagastrin testing, it is in any case questionable whether those patients can be cured. Thus, GRS may only be helpful to identify those patients with localised recurrence, who might profit from therapy. SPECT imaging might increase the sensitivity of GRS. Complete SPECT imaging from the pelvis to the skull base was not performed on a regular basis in our patients, which reflects clinical practice, where SPECT is usually done only in the event of indeterminate results (also due to time restrictions). Therefore, in a future study, the performance of GRS with SPECT in patients with occult MTC needs to be determined.

A correlation between calcitonin level, CEA level and outcome has been described in the literature [17]. Here, no correlation between gastrin uptake and either calcitonin or CEA level could be established, while somatostatin uptake showed an intermediate correlation with calcitonin level but no correlation with CEA level.

One advantage of minigastrin is its potential specificity. Octreotide shows uptake in inflammatory disease, granulomatous disease and neovasculature [2426]. SRS has also been reported to possibly be false positive in patients with MTC after external beam irradiation of the chest [27]. To address the issue of a possibly higher specificity of GRS under various conditions, further studies are necessary.

In previous studies, other nuclear medicine imaging techniques for the detection of MTC have been evaluated. 99 mmTc-labelled methoxyisobutylisonitrile (MIBI) and dimercaptosuccinic acid (DMSA) have been evaluated, as well as 201Tl and 123I-metaiodomethylguanidine (MIBG). These methods did not show advantages over anatomical imaging or SRS [2830]. Therefore, these imaging modalities were not included in our study.

As tumour uptake is high, minigastrin may also offer a therapeutic option if labelled with β-emitters such as 90Y or 177Lu [15]. Thus, it may also offer the possibility of optimising peptide receptor radiotherapy (PRRT) by choice of the most appropriate receptor to target or by use of a peptide cocktail targeting different receptors to increase the efficacy of PRRT [31]. The problem of high kidney uptake may be solved by co-administration of amino acids [32]. In respect to therapy, 131I-MIBG may also play a role in MTC. In view of its low sensitivity in imaging, however, its use will probably be limited to selected patients in combination with octreotide or minigastrin.

In conclusion, this clinical study indicates that GRS may become a valuable new method in the management of patients suffering from MTC. GRS in combination with CT showed high detection rates for metastatic lesions. Therefore, DGlu1-minigastrin seems to broaden the indication for radiopeptide imaging as an alternative to DPhe1-octreotide in patients with MTC. Further studies will have to be performed to address the role of GRS in localising occult MTC and to investigate the possibly higher specificity of GRS in comparison with SRS. In the future, development of a minigastrin derivative labelled with a positron emitter such as 68Ga may further increase the sensitivity of the method [33].

Copyright information

© Springer-Verlag 2006