ORIGINAL ARTICLE

Radio-guided surgery with the use of [99mTc-EDDA/HYNIC]octreotate in intra-operative detection of neuroendocrinetumours of the gastrointestinal tract

A. Hubalewska-Dydejczyk & J. Kulig & P. Szybinski &R. Mikolajczak & D. Pach & A. Sowa-Staszczak &

K. Fröss-Baron & B. Huszno

Received: 30 October 2006 /Accepted: 12 April 2007 / Published online: 25 July 2007# Springer-Verlag 2007

AbstractPurpose Radio-guided surgery (RGS) is an intra-operativelocalising technique which enables identification of tissue“marked” by a specific radiotracer injected before surgery.It is mainly used for sentinel node mapping and fordetection of parathyroid adenomas and other tumours,including neuroendocrine tumours of the gastrointestinaltract (GEP-NET). The aim of this study was to determinewhether intra-operative radio-detection with the use of[99mTc-EDDA/HYNIC]octreotate, a new somatostatin ana-logue, is able to reveal an unknown primary and secondarysites, thereby improving surgical treatment and the finaloutcome of GEP-NET.Methods The study group included nine patients withsuspected GEP-NET (four carcinoids, five pancreaticNET) localised with somatostatin receptor scintigraphy(with [99mTc-EDDA/HYNIC]octreotate), who had negative

results on other pre-operative imaging tests. At surgery,suspected tumours were measured in situ and ex vivo andprecise exploration of the abdominal cavity was performedwith the intra-operative scintillation detector (Navigator).Results Intra-operative gamma counting localised threecarcinoids. In one patient SRS was false positive (owingto inflammatory infiltration). Compared with SRS, RGSrevealed additional lymph node metastases in one case.RGS resulted in successful localisation of all pancreaticNET (the smallest lesion was 8 mm in diameter).Conclusion [99mTc-EDDA/HYNIC]octreotate SRS followedby RGS is a promising technique to improve the rate ofdetection and efficacy of treatment of GEP-NET, especiallyin the presence of occult endocrine tumours. The imagingproperties of [99mTc-EDDA/HYNIC]octreotate and the 1-dayimaging protocol offer opportunities for more widespreadapplication of this tracer followed by RGS in oncology.

Keywords Radio-guided surgery . Somatostatin analogues .

SRS . [99mTc-EDDA/HYNIC]octreotate .

Neuroendocrine tumours . Gastrointestinal tract

Introduction

The intra-operative use of radioisotopes in the detection ofoccult pathology and metastases of cancers is one of themost rapidly evolving areas of surgery [1–3]. Radio-guidedsurgery (RGS) is an intra-operative localising techniqueusing target-specific radiotracers that accumulate in tissues.Intra-operative gamma probes have been used for years forsentinel node mapping in oncology. Currently, the othermain clinical applications of RGS include the detection ofparathyroid adenomas, thyroid and colon cancers, bone

Eur J Nucl Med Mol Imaging (2007) 34:1545–1555DOI 10.1007/s00259-007-0476-4

A. Hubalewska-Dydejczyk :D. Pach :A. Sowa-Staszczak :K. Fröss-Baron :B. HusznoNuclear Medicine Unit of the Department of Endocrinology,Medical College at Jagiellonian University,31-501 Cracow, Poland

J. Kulig : P. SzybinskiThe Gastrointestinal and General Surgery Department,Medical College at Jagiellonian University,Cracow, Poland

R. MikolajczakRadioisotope Centre POLATOM,Otwock-Swierk, Poland

A. Hubalewska-Dydejczyk (*)Kopernika Str. 17,Krakow, Polande-mail: alahub@cm-uj.krakow.pl

tumours (osteoid osteoma) and gastroenteropancreaticneuroendocrine tumours (GEP-NET) [4–6]. For labellingsuitable ligands, 111In, 99mTc, 123I and 131I are mainly used.

GEP-NET are usually well-differentiated neoplasms(according to the WHO classification [7]) with thecapability of hormone secretion and with malignantpotential irrespective of their small size and benignappearance. The exception to this rule is insulinoma, whichin 90% of cases turns out to be a benign tumour. Despite awide spectrum of modern diagnostic imaging methods,GEP-NET remain undetectable for years, up to the time ofdissemination of the neoplasm. On the other hand, in recentyears a significant increase in the incidence of GEP-NEThas been reported in the literature owing to a greaterawareness of them and hence increased diagnosis [8].

The effectiveness of surgical treatment depends on theradical removal of the tumour mass. Remaining tumourtissue, imperceptible to the surgeon (microscopic and occultdisease), leads to unsuccessful treatment and shortening ofthe patient’s life.

Somatostatin receptor scintigraphy (SRS) followed byRGS offers to detect occult tumours and improve theeffectiveness of surgical treatment of patients sufferingfrom GEP-NET expressing somatostatin receptors (sstr) andtaking up somatostatin analogues. The sensitivity of SRS inlocating GEP-NET varies from 51.2% to 96% according todifferent studies and depending on the tracer used [9–11].For PET examinations, the sensitivity is higher, and it canreach 95% according to Orlefors et al. [12]. In a few studiesbased on examination of small groups of patients, thesensitivity of scintigraphy followed by RGS in detectingGEP-NET reached 90% [1, 13, 14], as a result of thepossibility of localising a lesion up to 5 mm in diameterwhen a scintillation detector is used [13]. Some authorshave concluded that intra-operative detection of GEP-NETafter administration of 111In-OctreoScan is more sensitivethan SRS and intra-operative palpation, allowing moreextensive surgical eradication of the disease [13, 15, 16].

[111In-DTPA]octreotide is a well-established radiotracerwhich has been used for many years for pre-operativelocalisation of GEP-NET, and it is also used successfullyfor intra-operative detection of these tumours. In recentyears, new somatostatin analogues for 99mTc labelling havebeen developed to optimise the imaging diagnosis of GEP-NET [17–19]. Tyr3-octreotate labelled with 99mTc, [99mTc-EDDA/HYNIC]octreotate, a somatostatin analogue withhigher affinity to somatostatin receptor subtype 2 (sstr2)than octreotide, was applied in the study of Reubi et al.[19]. Our earlier experiences showed that [99mTc-EDDA/HYNIC]octreotate has a similar biodistribution to 111In-OctreoScan, with a somewhat higher target/non-target ratio,and that it helps in staging and determining the SRS statusof well-differentiated NET (WHO groups 1 and 2), and in

the detection of a primary focus not detectable with otherimaging methods [20].

The aim of this study was to determine whether intra-operative radio-detection using the new somatostatinanalogue [99mTc-EDDA/HYNIC]octreotate is able to revealunknown primary and secondary sites, thereby improvingsurgical treatment and the final outcome of GEP-NET.

Materials and methods

Patients

The study group included nine patients (mean age 48±20,range 18–68 years, seven males and two females) withGEP-NET diagnosed with [99mTc-EDDA/HYNIC]octreo-tate SRS, all of whom had negative results on pre-operativeimaging tests.

The first subgroup consisted of four patients suspected ofhaving carcinoid tumours based on the clinical symptomsand/or abnormal biochemical tests [elevated urinary5-hydroxyindoleacetic acid (5-OHIAA) and serum chro-mogranin A (CgA) levels]. Only one patient had histopa-thologically confirmed carcinoid in a liver metastasis (theprimary focus of the disease was unknown). Carcinoid livermetastasis was resected in the year 2000, followed by6 months of chemotherapy (Zanosar + 5-fluorouracil). Thepatient did not turn up for a control until 2005. The meanvalue of 5-OHIAA in this subgroup of patients was107.5 μmol/24 h, with a range of 52–238 (normal range5–42 μmol/24 h); the mean value of CgA was 67.5 U/l,with a range of 22–187 U/l; (normal range 2–18 U/l). A siteof focal uptake of [99mTc-EDDA/HYNIC]octreotate duringSRS suggested a small intestinal tumour in two cases, andcaecal GEP-NET in the other two. In two cases, lymphnode metastases were suspected on the basis of the resultsof scintigraphy; in one patient, multiple liver metastaticlesions were visible.

In five out of the nine patients, an islet cell tumour of thepancreas was suspected. Based on the clinical state andlaboratory tests (neuroglycopenic symptoms, fasting hypo-glycaemia with inappropriate serum insulin level, positivefasting test), insulinoma was suspected in four of them,although all other imaging tests before [99mTc-EDDA/HYNIC]octreotate SRS had failed to reveal tumours ofthe pancreas. The mean value of CgA in patients withsuspected GEP-NET of the pancreas was 17.5 U/l (range15.1–19.3). In one woman without clinical symptoms of afunctioning pancreatic tumour, ultrasound and CT revealeda pancreatic cyst with a negative history of pancreatitis. Theresult of aspiration biopsy under ultrasound control wasambiguous, suggesting the need for further diagnostic

1546 Eur J Nucl Med Mol Imaging (2007) 34:1545–1555

measures regarding a pancreatic GEP-NET. Because ofpositive SRS, the woman qualified for RGS. No preoper-ative serum glucagon level was obtained.

Surgical treatment with the use of an intra-operativescintillation detector and intra-operative ultrasound wasappropriate for all patients.

None of the patients was treated with cold somatostatinanalogues before SRS. Details of the patients’ data andresults of pre-operative [99mTc-EDDA/HYNIC]octreotateSRS are presented in Table 1.

Patients who qualified for the study gave writteninformed consent prior to inclusion and the study wasapproved by the local ethical committee.

Imaging protocol

Radiopharmaceutical kit

Preparation of [99mTc-EDDA/HYNIC]octreotate has beendescribed previously [20]. Briefly, the content of alyophilised kit developed at the Radioisotope CentrePOLATOM, Poland with [HYNIC]-octreotate (20 μg),tricine (50 mg), mannitol (10 mg) and SnCl2×2H2O(40 μg) was labelled with 1–1.5 GBq 99mTc in 1–2 ml ofa 99Mo/99mTc generator eluate and incubated for 20 min at80°C using EDDA (5 mg) as co-ligand. Overall labellingyield was >90% and no further purification was required.

[99mTc-EDDA/HYNIC]octreotate SRS

[99mTc-EDDA/HYNIC]octreotate (740 MBq) scintigraphywas performed in all patients. Whole-body (WB) scanswere obtained 1, 4 and 24 h after tracer administration, andsingle-photon emission computed tomography (SPECT)was performed 4 h after injection of the tracer. The studieswere performed with a dual-head, large field of view E.CAM gamma camera (Siemens) equipped with parallel-hole, low-energy, high-resolution collimators connectedwith the e.soft and ICON operating systems.

All patients underwent a WB scan in the anterior andposterior projections (256×1,024 matrix, 6 cm/min) andSPECT of the abdomen and/or the chest. The SPECT wasperformed using a 360° orbit (180° for each head), in stepand shoot mode, at 30 s per view. The data acquired werecollected in a 128×128 computer matrix and reconstructedusing iterative reconstruction (OSEM 2D, number ofiterations 4 or 8).

In seven patients, [99mTc-EDDA/HYNIC]octreotate(700 MBq) was additionally injected 24 h before surgery.Only in two patients with positive scans 1 and 4 h after tracerinjection was acquisition of the data not performed after 24 howing to the surgery planned on the following day.

The scans were evaluated by two independent physi-cians. Focal [99mTc-EDDA/HYNIC]octreotate accumula-tion was considered to be specific when scans visualisedthe lesions at both 4 h and 24 h (or, alternatively, 1 h) post

Table 1 Patient data, values of clinical markers of GEP-NET and results of pre-operative [99mTc-EDDA/HYNIC]octreotate SRS in the examinedgroup of patients (n=9)

Patient no. Age(yrs)

Sex CgA (normalrange 2–18 U/l)

5-OHIAAa (normal range5–42 μmol/24 h)

[99mTc-HYNIC/EDDA]octreotate SRS results: target/non-target ratio

Primary tumour Lymph nodemetastases

Livermetastases

WB SPECT WB SPECT WB SPECT

Carcinoidtumour

1 66 M 26 82 2.0 8.5 1.8(1)

7.7 (1) (0) (0)

2 57 M 187 238 2.5 12.5 1.9(2)b

5.4(2)b

2.7(6)b

8.5(6)b

3 53 M 22 58 1.7 2.9 (0) (0) (0) (0)4c 18 M 55 52 1.6 6.4 (0) (0) (0) (0)

Pancreatictumour

5 25 M 19.3 – 2.9 6.36 62 F 17.9 – 2.8 10.57 31 M – – 2.3 3.58 22 M 17.8 – 1.9 6.89 68 F 15.1 – 2.4 7.2

Number of revealed lymph nodes and liver metastases are presented in parentheses; if there was more than one metastatic focus, the target/non-target ratio is expressed as a mean valueWB whole body, SPECT single-photon emission computed tomographya Only in patients with suspected carcinoidbMean valuec Patient with false positive results of [99m Tc-EDDA/HYNIC]octreotate SRS; Crohn’s disease was finally diagnosed

Eur J Nucl Med Mol Imaging (2007) 34:1545–1555 1547

injection of the tracer. For the evaluation of tumour uptake,the total counts and counts per pixel in focal lesions werecalculated. The background region of interest was definedin the neighbouring normal tissue for each tumour. Target/non-target ratios were assessed for WB (4 h) and SPECT(transverse slices with the highest tumour uptake weretaken) images.

Patients were prepared with laxatives (Fortrans) andwere placed on a liquid diet for 2 days before scintigraphy.

Gamma positioning system

The intra-operative inspection of the abdominal cavitymeasuring the radioactivity of tumours and normal tissueswas performed with a portable, hand-held gamma probedetector [Navigator GPS-cadmium telluride (CdTe) semi-conductor detector; Tyco Healthcare]. The probe was sterileand the connected cord was covered with a sterile sleeve.At surgery, suspected tumours and lymph nodes weremeasured in vivo and ex vivo after excision. For theinspection, two probes with 11- and 14-mm-diameterdetectors were used. The probe was connected to theelectronic controller with visual and audible indication ofthe counting rate (range 0–25,000 counts/s) as the site oftracer accumulation was approached. The point of themaximal count rate was localised by moving the probeslowly horizontally and vertically over the lesion, startingfrom the baseline areas. For RGS, the target/non-target ratiowas expressed as the ratio of the peak of the count rate tothe background radiation.

In all cases, before surgery patients were sent forabdominal ultrasound examination, endoscopic ultrasound(EUS), computed tomography (CT; four-slice CT scanner,collimation at 4×2.5-mm, 3-mm slices) and magneticresonance imaging (MRI) scans as well as gastroscopy andcolonoscopy in cases of carcinoid tumours. Arteriographywas performed in two cases of islet pancreatic tumours.

In all patients, routinely intra-operative ultrasound exam-ination (IOUS) was performed using a Hitachi EUB 6000machine with colour Doppler option. IOUS was performedby operating surgeons and ultrasonographers using a 7.5-MHz side-fired T-shaped linear array probe and a convex-array end-fire probe. Imaging modalities were used to detectthe primary lesion and to assess its location, extent andresectability, as well as to detect potential liver metastases.

At the time of surgical exploration of pancreaticendocrine neoplasms, complete evaluation of the pancreasand peripancreatic regions was performed with digitalpalpation, IOUS and RGS. After resection of the tumour,precise exploration of the abdominal cavity was performed,starting from the liver, encompassing the left pancreaticparenchyma and peripancreatic lymph nodes, and extendingalong the aorta for evidence of pathological radioactive

accumulation. In cases of carcinoid tumours, particularattention was paid to the small bowel and its mesentery.

Results

[99mTc-EDDA/HYNIC]octreotate SRS

[99mTc-EDDA/HYNIC]octreotate SRS revealed focal accu-mulation of the tracer, suggesting a primary GEP-NET, inall nine patients. In three cases with suspicion of a carcinoidtumour, positive results were confirmed by histologicalexamination. In the fourth case with a suspected carcinoidtumour, positive tracer accumulation was related to Crohn’sdisease and inflammatory infiltration. Amongst the patientswith confirmed carcinoid tumours, abdominal lymph nodemetastases were visible in two cases; in one of them, sixliver metastatic lesions were found which had remainedundetected using other imaging examinations.

The mean target/non-target ratios for WB and SPECTscans are listen in Table 1.

Manual probe measurements for carcinoids

In the carcinoid study group, intra-operative gamma probecounting confirmed four somatostatin receptor-positivelesions primarily diagnosed with SRS, including the patientwith Crohn’s disease. Inspection of the abdominal cavityafter laparotomy revealed large mesenteric tumours in twopatients. These tumours, inducing fibroblastic and inflam-mation reactions with adhesions of adjacent bowel loopsand tissues, seemed to be primary lesions. In fact, however,the primary tumours were much smaller, and easily over-looked. In one case, RGS detected the primary lesion as anulcerative infiltration (Fig. 1) and in the other, as a polypoidgrowth without any signs of bowel wall infiltration, whichwas not visible to the surgeon (Fig. 2). Final clinical data ofthe carcinoid subgroup, including location, staging andsurgery treatment, are presented in Table 2. In all cases, thetarget/non-target ratio (RGS) was high for primary tumours,with values ranging from 16.9 to 50.5, as compared withthe lymph node metastases, with values ranging from 1.9 to3.9 (Table 3).

The main advantage of RGS in comparison to SRS ishigh sensitivity in the detection of metastatic lymph nodes(Table 4). In two cases the use of a hand-held gammaprobe extended the surgical procedure, resulting in thesuccessful excision of additional metastatic lymph nodes. Inthe first case RGS revealed three metastatic lymph nodescompared with two shown by SRS. In the second, threemetastases detected with RGS were close to the primarytumour infiltration, probably resulting in negative SRS. Thesmallest lymph node detected with RGS was 6 mm.

1548 Eur J Nucl Med Mol Imaging (2007) 34:1545–1555

Amongst all the cases, only one metastatic lymph node wasundetected according to histopathology. The target/non-target ratio (RGS) for metastases was substantially lowerthan that for the primary tumours owing to their locationclose to the mesenteric root and retroperitoneal area(Table 3). RGS also confirmed liver metastases that werenot visible on other preoperative diagnostic examinationswith the exception of SRS. It is important to realise thatliver examination with a hand-held gamma probe is alsolimited insofar as it can yield false positive results: in oneof our patients a liver tumour of 5 mm turned out to be acavernous haemangioma.

This subgroup of patients is being followed up in ourpoliclinic. In two patients, after resection of the primary tumourand all detected lymph node metastases the control SRS did notreveal a recurrence of the disease. One patient with livermetastases was referred to [90Y-DOTA]-TATE therapy.

Manual probe measurements for islet pancreatic tumours

Islet cell tumours were suspected in five patients with afocal accumulation of [99mTc-EDDA/HYNIC]octreotate in

the pancreatic parenchyma together with specific clinicalsymptoms. All patients were referred to surgery with theuse of RGS on the basis of a positive SRS in spite ofnegative results using other imaging examinations (Fig. 3).The final histopathology revealed five islet cell tumours:four insulinomas and one glucagonoma. In one insulinoma,tumour capsule infiltration was found. Intra-operativeinspection together with IOUS and RGS did not revealany lymph node or liver metastases. The clinical data on thestage, location of the tumour and surgical proceduresperformed are listed in Table 5.

In all cases RGS resulted in successful localisation of thepancreatic tumour. The smallest lesion identified with thehand-held probe (not palpated by the surgeon) was atumour of 8 mm diameter. These results were consistentwith SRS as well as IOUS examinations and the probe wasmore sensitive than palpation.

The target/non-target count ratio for insulinomas in SRS(SPECT) was within the range of 3.5–10.5 (Table 1). ForRGS, the target/non-target ratio was much lower, reaching1.8 for insulinomas and 2.5 for glucagonoma, owing to theretroperitoneal location of the pancreas and the high

Fig. 1 A 66-year-old man with midgut carcinoid. a Pre-operative[99mTc-EDDA/HYNIC]octreotate SRS (SPECT) showed two foci ofpathological tracer uptake within the peritoneal cavity (arrows). b, cMesenteric metastatic lymph node infiltrating surrounding tissues,

confirmed with RGS (arrow). d Ex vivo hand-held gamma proberadioactivity counting of the removed lymph node. e Small primarytumour detected with RGS

Eur J Nucl Med Mol Imaging (2007) 34:1545–1555 1549

background activity (Table 6). The target/non-target countratios for WB (1.9–2.9; Table 1) and for RGS werecomparable. The maximal count rates did not differdepending on tumour size or pancreatic location.

All patients operated on are free of clinical symptomsand remain in the follow-up programme.

Discussion

GEP-NET of the small intestine, with their relatively highmalignant potential, and small pancreatic GEP-NET causethe most difficulties in pre-operative localisation. The useof a hand-held gamma probe helps to localise these tumours

F ig . 2 a [ 9 9mTc -EDDA/HYNIC]octreotate SRS of a 57-year-old man with disseminatedcarcinoid and an ileal primaryfocus detected with SRS fol-lowed by RGS (thick arrowsindicate primary focus andmetastatic lesion to neighbour-ing lymph node; thin arrowsshow liver metastases). b SRSof the same patient after resec-tion of the primary focus, lymphnodes and one of the livermetastases. c Ex vivo radioac-tivity counting of the ileal pri-mary carcinoid tumour. d Smallprimary tumour in the form of apolypoid growth

Table 2 Final location of the primary tumours, staging and extent of the operation in the subgroup of patients with histologically confirmedmidgut carcinoid tumours (n=3)

Patient no. Tumour location Tumoursize (mm)

Surgical procedure Infiltrationbeyond primarysite

Lymphnode status

Distantmetastases

WHOclassification

Carcinoidtumour

1 Jejunum 12 Partial resection of smallbowel

+ + Livera 2

2 Ileum (14 cmfrom caecum)

20 Partial resection andright hemicolectomy

+ + Liver 2

3 Ileum (5 cmfrom caecum)

25 Partial resection andright hemicolectomy

+ + – 2

a Patient after resection of liver carcinoid metastasis in 2000; currently no liver metastases were detected by [99m Tc-EDDA/HYNIC]octreotateSRS, RGS or IOUS

1550 Eur J Nucl Med Mol Imaging (2007) 34:1545–1555

with greater precision and sensitivity, especially in caseswith positive SRS and negative results of other imagingmethods. SRS followed by RGS can identify primary GEP-NET and their metastases which have not been diagnosedwith other procedures, and the density and distribution ofcell membrane-bound receptors in a lesion are moredeterminant than the tumour size [19, 21].

Somatostatin receptor subtype 2 (sstr2) is the mostfrequently expressed subtype in midgut primary GEP-NET and their metastases [22]. According to Bertherat etal., sstr2 is also present in insulinomas together withsubtype 5 (sstr5) in 70% of cases and is thus detectableby SRS [23]. On the other hand, a relatively low sensitivityof SRS in the detection of insulinomas was reported byKrenning et al. owing to the low density of ssrt2 expressedon these tumours [24].

It is known that tumours with higher malignancy have alower density of somatostatin receptors. The frequency andexpression pattern of each subtype vary considerably notonly in different tumour types but also in each patient [25].

It is said that successful use of SRS in the diagnosis ofGEP-NET is mainly determined by the overexpression ofsstr2 and, to a lesser degree, sstr5 on the tumour.

Among 99mTc-labelled somatostatin analogues, [99mTc-EDDA/HYNIC]TOC is mainly used in the imagingdiagnosis of GEP-NET. Gabriel et al. reported [99mTc-EDDA/HYNIC]TOC SRS to have an overall sensitivity of80%, a specificity of 94,4% and an accuracy of 82.9% inpatients suffering from GEP-NET [18]. Nevertheless, thereare only a few reports on the use of 99mTc-labelledsomatostatin analogues in RGS. One of them, by Fettichet al., describes successful intra-operative detection of twogastrinomas and two insulinomas after injection of [99mTc-EDDA/HYNIC]TOC [26].

The new somatostatin analogue HYNIC-octreotate usedin our study, with terminal threonine replacing threoninol inthe amino acid sequence of [HYNIC]octreotide, showshigher receptor affinity and a higher internalisation rate at alower liver background, yielding intense tumour uptake[27, 28]. In vitro studies using cell lines transfected with

Table 3 Intra-operative RGS counts of the primary carcinoids and lymph node metastases: in vivo and ex vivo measurements (n=3)

Patient no. Primary tumour (counts/s)a Lymph node metastases (counts/s)b Target/non-target ratio (in vivo)c

In vivo Ex vivo In vivo Ex vivo Primary focus Lymph node metastases

Carcinoid tumour 1 140 130 240 (1) 113 (1) 28.1 3.92 130 107 100 (3) 84 (3) 16.9 2.33 150 110 90 (3) 50 (3) 50.5 1.9

a Data expressed as maximal count rateb Data expressed as mean value of maximal count rates for all positive lymph nodes; number of positive lymph nodes in parenthesesc Data expressed as the ratio of the count rates of a target tissue to the background radiation (for lymph node metastases as a mean value for allpositive lymph nodes)

Table 4 The results of pre-operative [99mTc-EDDA/HYNIC]octreotate SRS, RGS findings and histopathology results in the examined group ofpatients (n=9)

Patient no. Detection of pathological lesions

Primary tumour Lymph node metastases Liver metastases

SRS RGS Pathology SRS RGS Pathology SRS RGS Pathology

Carcinoid tumour 1 + + + 1 1 1 0 0 02 + + + 2 3 3 6 Confirmeda Confirmeda

3 + + + 0 3 4 0 0b 0b

4c + + – 0 0 0 0 0 0Pancreatic tumour 5 + + + 0 0 0 0 0 0

6 + + + 0 0 0 0 0 07 + + + 0 0 0 0 0 08 + + + 0 0 0 0 0 09 + + + 0 0 0 0 0 0

a Liver carcinoid metastasis confirmed by histopathology in resected lesionb False positive result: histopathology revealed cavernous haemangiomac Patient with false positive results of [99m Tc-EDDA/HYNIC]octreotate SRS; Crohn’s disease was finally diagnosed

Eur J Nucl Med Mol Imaging (2007) 34:1545–1555 1551

somatostatin receptors revealed that Tyr3-octreotate showseight- to tenfold higher affinity for ssrt2 than TOC [19].

In our study, [99mTc-EDDA/HYNIC]octreotate SRS fol-lowed by RGS localised all primary midgut carcinoids andidentified more lymph node metastases than SRS. Neverthe-

less, histopathology revealed more lymph node metastases inone patient. A lymph node undetected with RGS was 6 mmin diameter and was found in proximity to the primarytumour, which made it difficult to differentiate them intra-operatively. Compared with SRS using 111In-OctreoScan,

Table 5 Final location of the primary tumours, staging and extent of the operation in the subgroup of patients with histologically confirmed isletcell pancreatic tumours (n=5)

Patient no. Diagnostic tests Tumourlocation

Pathology Tumoursize (mm)

Surgicalprocedure

Malignancy WHOclassification

Pre-operative IOUS

SRS Other

Pancreatictumours

5 + − Positive Tail Insulinoma 25 Leftpancreatectomya

2.5/10 HPFcapsuleinfiltration

2

6 + − Positive Tail Insulinoma 18 Leftpancreatectomy

– 1

7 + − Positive Tail Insulinoma 20 Leftpancreatectomy

– 1

8 + − Positive Head/body

Insulinoma 8 Enucleation oftumour

– 1

9 +b −b Positive Head/body

Cysticglucagonoma

40 Local excision – 1

a Left pancreatectomy with spleen preservationb Pre-operatively diagnosed as pancreatic cyst

Fig. 3 A 24-year-old man with clinical signs and symptoms ofinsulinoma. a,b Results of [99mTc-EDDA/HYNIC]octreotate SRSbefore (a) and after (b) resection of the tumour of the pancreatic tail

(arrow). c Intra-operative location of the pancreatic tumour. d Ex vivoradioactivity counting of the insulinoma: cross-section of the tumour

1552 Eur J Nucl Med Mol Imaging (2007) 34:1545–1555

RGS has been reported to have a better sensitivity indetecting primary foci and their metastases in different GEP-NET [14, 16, 29]. According to Adams et al., RGS seems tobe the most sensitive diagnostic tool for detecting micro-scopic and occult endocrine tumours [29]. In a group of 12patients, RGS revealed 70 lesions, whereas 52 werevisualised preoperatively and only 31 were palpable duringsurgery. The smallest lesion identified by the probe reportedby Adams et al. was 6 mm, which is concordant with ourresults. Although in the cited report SRS was performed with111In-OctreoScan, which has lower ssrt2 affinity than [99mTc-EDDA/HYNIC]octreotate, the results of the two studies arevery similar, showing intra-operative detection of occultGEP tumours to be highly feasible.

The hand-held gamma probe enabled us to locate lymphnode metastases, including those invisible during operation,as well as remnant neoplasm infiltrations in the tumour bed,leading to radical surgical treatment. It had a crucial effecton the final outcome of GEP-NET patients. In our study,detection of previously unknown lymph node metastaseswith the use of the hand-held gamma probe markedlyaffected decision-making with regard to the extent of thesurgical procedure.

Removal of a primary tumour together with all disseminat-ed lymph nodes, even in the presence of multiple livermetastases, significantly delayed progression of the diseasewhen appropriately treated [7]. Additionally, it has beendocumented that metastases to the mesenteric lymph nodesmay be located near to the bowel wall. As they induceintense fibroblastic and inflammatory reactions as well ascontraction and adherence of adjacent bowel loops andtissues, mesenteric metastases may appear larger than theprimary lesion. In such cases a small primary lesion can beoverlooked by the surgeon. RGS allows localisation of thesesmall tumours. The limitation of the method is the incidenceof false positive results due to tracer accumulation in theactivated lymphocytes within inflammatory infiltrates whichexpress somatostatin receptors [21]. We had such a falsepositive result of [99mTc-EDDA/HYNIC]octreotate SRS in a

patient with unrecognised Crohn’s disease. In that case,carcinoid was additionally suspected on the basis of typicalclinical symptoms such as persistent diarrhoea and a highCgA level, probably related to the treatment with protonpump inhibitors.

The goals of surgical therapy for pancreatic endocrineneoplasms include safe resection of the tumour mass andmaximal preservation of pancreatic parenchyma, as well asprovision of relief from symptoms related to hormoneexcess. Insulinomas are mostly small tumours and approx-imately 20–60% of them cannot be localised preoperative-ly; however, more than 90% can be localised by palpationalone or with the aid of IOUS [30–32]. In 1985 Norton etal. first described localisation of insulinoma (8 mm) withthe use of IOUS. The tumour was not surgically palpableand had not been detected by selective arteriography [33].The development of EUS considerably improved the rate ofpre-operative detection of small endocrine pancreatictumours (gastrinomas and insulinomas) and lesions assmall as 5 mm. In our study, [99mTc-EDDA/HYNIC]octreotate SRS followed by RGS detected all insulinomas.EUS examinations were negative, probably owing tolocalisation of the tumours in the pancreatic tail known asthe “blind spot” at the splenic hilum.

The smallest lesion identified by the hand-held probe wasnot palpable by the surgeon. Similarly, numerous reportshave shown that RGS is more sensitive than intra-operativepalpation [15, 16, 29]. [99mTc-EDDA/HYNIC]octreotateSRS followed by RGS seems to be a sensitive method forthe pre- and intra-operative detection of insulinomas. Apositive SRS result, with accompanying clinical symptomsand biochemical tests suggesting pancreatic NET, shouldserve as the indication for the operation, even with negativeresults of other imaging methods.

SRS is very rarely used for the detection of glucagonomaowing to its size and the advanced stage at the time ofdiagnosis, when the chance of cure is limited. Therefore ourlast patient, with a diagnosis of cystic glucagonoma, is veryinteresting. Glucagonoma is a rare, usually solid and highly

Table 6 Intra-operative RGS counting of the islet cell tumours: in vivo and ex vivo measurements (n=5)

Patient no. Countinga (counts/s) Target/non-target ratiob

Primary focus Pancreatic parenchyma

In vivo Ex vivo In vivo Ex vivo

Pancreatic tumour 5 760 60 420 10 1.86 680 80 350 14 1.97 840 110 400 18 2.18 560 70 300 7 1.89 730 100 290 9 2.5

a Data expressed as maximal count ratesb Data expressed as the ratio of a target tissue to the background radiation count rates

Eur J Nucl Med Mol Imaging (2007) 34:1545–1555 1553

vascular tumour of pancreatic alpha cells without a tendencytowards cystic degenerative changes. Cystic glucagonoma is aunique variant (only a few cases have been described in theliterature) of the disease and should be mainly considered inthe differential diagnosis of cystic pancreatic neoplasms [34].In spite of suspicion of a non-functional pancreatic NET,RGS and IOUS inspection of the tumour bed, surroundingtissues and liver were performed, with no liver or lymphnode metastases being found. [99mTc-EDDA/HYNIC]octreo-tate SRS 1 year after surgery was negative. The patientremains in the follow-up programme.

The gamma probe, providing real-time response andpositioning flexibility, has become the new tool for theinspection and palpation traditionally performed by thesurgeon. Flexibility in positioning of the probe is essentialfor the examination. The major limitation of intra-operativegamma probe detection is high background activity from theliver, spleen and kidneys. It is extremely important to avoiddirecting the probe towards such physiological traceraccumulations. During exploration of the pancreas, theprobe should be placed towards the posterior surfaces ofthe pancreas, directing it up towards the abdominal wall. Inthe case of carcinoid tumours, the examination is easier.Moving the bowel loop beyond the peritoneal cavity, theexamination can be performed while avoiding abdominalbackground activity. Intra-operative gamma probe scanningrequires careful, slow searching of suspected areas, which istime consuming owing to subtle activity differences occur-ring due to background type. This has also been emphasisedby other authors [14].

According to the literature, successful RGS tumourdetection requires tumour to non-tumour tissue count ratiosof at least 1.5–2.0; however, a ratio above 2.4 makesdetection more reliable. In clinical practice a several-foldhigher ratio can be observed [16, 29]. In a group of patientswith midgut carcinoid tumours, the target/non-target countratio reached 50 for primary lesions and averaged 2.7 forlymph node metastases. The target/non-target count ratiowas substantially lower for islet cell pancreatic tumours.There was no correlation between the in vivo and ex vivomeasurements because the in vivo count rate varied greatlywith the type and location of the target tissue.

Use of a CsI(Tl) gamma probe in surgery was firstdescribed by Harris et al. in 1956 [35]. Currently, there isgrowing interest in the use of positron emission tomogra-phy imaging with, for example, [68Ga-DOTA]TOC forbetter visualisation of GEP-NET, although it is still limitedmainly to clinical research [36]. These rather new tracersseem to be particularly useful in detecting small tumours ortumours bearing only a low density of somatostatinreceptors [36, 37]. In the future, the development of morespecific ligands and other radioisotopes should furtherincrease the efficiency of RGS in patients suffering from

GEP-NET. Currently, not only have new radiopharmaceu-ticals been developed to improve the identification ofspecific tissue during surgery, but in addition the scintilla-tion probes have been upgraded. Widespread introductionof the imaging gamma detectors in the future shouldfacilitate the excision of target tissue surrounded by non-uniform background radioactivity [38].

Conclusion

The results of this study show that [99mTc-EDDA/HYNIC]octreotate SRS followed by RGS is a promising techniqueto improve both the rate of detection of GEP-NET and thetreatment efficacy, especially in patients with occultendocrine tumours. RGS is one of the most sensitivelocalising methods for the detection of unknown metastaticlymph nodes in the abdominal cavity. The imagingproperties of [99mTc-EDDA/HYNIC]octreotate, with goodaffinity to sstr2 and the possibility of a 1-day imagingprotocol before surgery, offer opportunities for morewidespread use of SRS followed by RGS in oncology.

References

1. Benjegärd SA, Forssell-Aronsson E, Wängberg B, Skänberg J,Nilsson O, Ahlman H. Intraoperative tumour detection using 111InDTPA-D-Phe-octreotide and a scintillation detector. Eur J NuclMed 2001;28:1456–62.

2. Kitagawa Y, Kitajima M. Diagnostic validity of radio-guidedsentinel node mapping for gastric cancer: a review of currentstatus and future direction. Surg Technol Int 2006;15:32–6.

3. Caudle AS, Brier SE, Calvo BF, Kim HJ, Meyers MO, Ollila DW.Experienced radio-guided surgery teams can successfully performminimally invasive radio-guided parathyroidectomy without intra-operative parathyroid hormone assays. Am Surg 2006;72:785–90.

4. Bozkurt MF, Ugur O, Hamaloglu E, Sayek I, Gulec SA.Optimization of the gamma probe-guided parathyroidectomy.Am Surg 2003;69:720–5.

5. Waddington WA, Kettle AG, Heddle RM, Coakley AJ. Intra-operative localization of recurrent medullary carcinoma of thethyroid using indium-111 pentetreotide and a nuclear surgicalprobe. Eur J Nucl Med 1994;21:363–4.

6. Waddington WA, Davidson BR, Todd-Pokropek A, Boulos PB,Short MD. Evaluation of a technique for the intraoperativedetection of a radiolabelled monoclonal antibody against colorec-tal cancer. Eur J Nucl Med 1991;18:964–72.

7. Ramage JK, Davies AH, Ardill J, Bax N, Caplin M, Grossman A,et al. Guidelines for the management of gastroenteropancreaticneuroendocrine (including carcinoid) tumours. Gut 2005;54:1–16.

8. Warner RR. Endocrine tumours other than carcinoid: a review ofcurrent clinical advances. Gastroenterology 2005;128:1668–84.

9. Gabriel M, Decristoforo C, Donnemiller E, Ulmer H, WatfahRychlinski C, Mather SJ, et al. An intrapatient comparisonof 99mTc-EDDA/HYNIC-TOC with 111In-DTPA-octreotide fordiagnosis of somatostatin receptor-expressing tumors. J Nucl Med2003;44:708–16.

1554 Eur J Nucl Med Mol Imaging (2007) 34:1545–1555

10. Chiti A, Briganti V, Fanti S, Monetti N, Masi R, Bombardieri E.Results and potential of somatostatin receptor imaging in gastro-enteropancreatic tract tumours. Q J Nucl Med 2000;44:42–9.

11. Ellison EC, Schirmer WJ, Olsen JO, Pozderac RV, Hinkle G, HillT, et al. Localization of neuroendocrine tumors using somatostatinreceptor imaging with indium-111-pentetreotide (OctreoScan).Cancer Control 1997;4:35–9.

12. Orlefors H, Sundin A, Garske U, Juhlin C, Oberg K, Skogseid B, etal. Whole-body 11C-5-hydroxytryptophan positron emission tomog-raphy as a universal imaging technique for neuroendocrine tumors:comparison with somatostatin receptor scintigraphy and computedtomography. J Clin Endocrinol Metab 2005;90:3392–400.

13. Adams S, Baum RP. Intraoperative use of gamma-detecting probesto localize neuroendocrine tumors. Q J Nucl Med 2000;44:59–67.

14. Banzo J, Vidal-Sicat S, Prats E, Galofre G, Razola P, Mane S, etal. In-111 DTPA octreotide scintigraphy and intraoperativegamma probe detection in the diagnosis and treatment of residuallymph node metastases of a rectal carcinoid tumor. Clin Nucl Med2005;30:308–11.

15. Öhrvall U, Westlin JE, Kjellberg F, Nilsson S, Juhlin C, Rastad J,et al. A gamma detector probe with ex vivo detection of carcinoidtumors superior to intraoperative palpation. Cancer 1997;80:2495–500.

16. Öhrvall U, Westlin JE, Nilsson S, Juhlin C, Rastad J, LundqvistH, et al. Intraoperative gamma detection reveals abdominalendocrine tumors more efficiently than somatostatin receptorscintigraphy. Cancer 1997;80:2490–4.

17. Decristoforo C, Mather J, Cholewinski W, Donnemiller E,Riccabona G, Moncayo R. 99mTc EDDA/HYNIC-TOC: a new99mTc-labelled radiopharmaceutical for imaging somatostatinreceptor-positive tumours: first clinical results and intra-patientcomparison with 111In-labelled octreotide derivatives. Eur J NuclMed 2000;27:1318–25.

18. Gabriel M, Muehllechner P, Decristoforo C, von Guggenberg E,Kendler D, Prommegger R, et al. 99mTc-EDDA/HYNIC-Tyr(3)-octreotide for staging and follow-up of patients with neuroendo-crine gastro-entero-pancreatic tumours. Q J Nucl Med MolImaging 2005;49:237–44.

19. Reubi JC, Schar JC, Waser B, Wenger S, Heppeler A, Schmitt JS,et al. Affinity profiles for human somatostatin receptor subtypesSST1–SST5 for somatostatin radiotracers selected for scintigraph-ic and radiotherapeutic use. Eur J Nucl Med 2000;27:273–82.

20. Hubalewska-Dydejczyk A, Fross-Baron K, Mikolajczak R,Maecke HR, Huszno B, Pach D, et al. 99Tc-EDDA/HYNIC-octreotate scintigraphy, an efficient method for the detection andstaging of carcinoid tumours: results of 3 years’ experience. Eur JNucl Med Mol Imaging 2006;33:1123–33.

21. Warner RR, O’dorisio TM. Radiolabeled peptides in diagnosis andtumor imaging: clinical overview. Semin Nucl Med 2002;32:79–83.

22. Guillermet-Guibert J, Lahlou H, Pyronnet S, Bousquet C, SusiniC. Endocrine tumours of the gastrointestinal tract. Somatostatinreceptors as tools for diagnosis and therapy: molecular aspects.Best Pract Res Clin Gastroenterol 2005;19:535–51.

23. Bertherat J, Tenenbaum F, Perlemoine K, Videau C, Alberini JL,Richard B, et al. Somatostatin receptors 2 and 5 are the majorsomatostatin receptors in insulinomas: an in vivo and in vitrostudy. J Clin Endocrinol Metab 2003;88:5353–60.

24. Krenning EP, Kwekkeboom DJ, Bakker WA, Breeman WA, KooijPP, Oei HY, et al. Somatostatin receptor scintigraphy with [111In-DTPA-D-Phe]octreotide and [123I-tyr]-octreotide: the Rotterdamexperience with more than 1000 patients. Eur J Nucl Med1993;20:716–31.

25. Patel YC. Somatostatin and its receptor family. Front Neuro-endocrinol 1999;20:157–98.

26. Fettich J, Repse S, Snoj M, Zitko-Krhin M, Markovic M. 99mTc-EDDA/HYNIC-TOC is a suitable radiopharmaceutical for radio-guided surgery of neuroendocrine tumours. Abstract book ofInternational Symposium on Trends in Radiopharmaceuticals(ISTR) Vienna, Austria 2005; p 25–26 (IAEA-CN-130/015)

27. Behe M, Maecke HR. New somatostatin analogues labelled withtechnetium-99m [abstract]. Eur J Nucl Med 1995;22:791.

28. Storch D, Béhé M, Walter MA, Chen J, Powell P, Mikolajczak R,et al. Evaluation of [99mTc/EDDA/HYNIC]octreotide derivativescompared with [111In-DOTA0,Tyr3, Thr8]octreotide and [111In-DTPA0]octreotide: Does tumor or pancreas uptake correlate withthe rate of internalisation? J Nucl Med 2005;46:1561–9.

29. Adams S, Baum RP, Hertel A, Wenisch HJ, Staib-Sebler E,Herrmann G, et al. Intraoperative gamma probe detection ofneuroendocrine tumors. J Nucl Med 1998;39:1155–60.

30. Huai JC, Zhang W, Niu HO, Su ZX, McNamara JJ, Machi J.Localization and surgical treatment of pancreatic insulinomasguided by intraoperative ultrasound. Am J Surg 1998;175:18–21.

31. Hashimoto LA, Walsh RM. Preoperative localization of insulino-mas is not necessary. J Am Coll Surg 1999;189:368–73.

32. Oberg K, Eriksson B. Endocrine tumours of the pancreas. BestPract Res Clin Gastroenterol 2005;19:753–81.

33. Norton JA, Sigel B, Baker AR, Ettinghausen SE, Shawker TH,Krudy AG, et al. Localization of an occult insulinoma byintraoperative ultrasonography. Surgery 1985;97:381–4.

34. Brown K, Kristopaitis T, Yong S, Chejfec G, Pickleman J. Cysticglucagonoma: a rare variant of an uncommon neuroendocrinepancreas tumor. J Gastrointest Surg 1998;2:533–6.

35. Harris C, Bieglow R, Francis J, Kelley G, Bell P. A CsI(Tl)-crystalsurgical scintillation probe. Nucleonics 1956;14:102–8.

36. Kowalski J, Henze M, Schuhmacher J, Macke HR, Hofmann M,Haberkorn U. Evaluation of positron emission tomographyimaging using 68Ga-DOTAD Phe1-Tyr3)-octreotide in comparisonto 111In-DTPAOC SPECT. First results in patients with neuroen-docrine tumors. Mol Imaging Biol 2003;5:42–8.

37. Raylman RR. Performance of a dual, solid-state intraoperativeprobe system with 18F, 99mTc, and 111In. J Nucl Med 2001;42:352–60.

38. Pitre S, Menard L, Ricard M, Solal M, Garbay JR, Charon Y. Ahand-held imaging probe for radio-guided surgery: physicalperformance and preliminary clinical experience. Eur J NuclMed Mol Imaging 2003;30:339–43.

Eur J Nucl Med Mol Imaging (2007) 34:1545–1555 1555