Staging of digestive endocrine tumours using helical computed tomography and somatostatin receptor scintigraphy

Transcription

1 Original article Annals of Oncology 14: , 2003 DOI: /annonc/mdg160 Staging of digestive endocrine tumours using helical computed tomography and somatostatin receptor scintigraphy F. Panzuto 1, M. Falconi 2, S. Nasoni 1, S. Angeletti 1, A. Moretti 1, M. Bezzi 3, G. Gualdi 3, E. Polettini 3, R. Sciuto 4, A. Festa 4, F. Scopinaro 5, V. D. Corleto 1, C. Bordi 6, P. Pederzoli 2 & G. Delle Fave 1 * Departments of 1 Digestive and Liver Disease, 3 Radiology and 5 Nuclear Medicine, University La Sapienza, Roma; 2 Department of Surgery, University of Verona, Verona; 4 Department of Nuclear Medicine, National Cancer Institute, Roma; 6 Department of Pathology, University of Parma, Parma, Italy Received 22 August 2002; revised 23 October 2002; accepted 21 November 2002 Background: In patients with digestive endocrine tumours, complete pre-operative staging is essential in planning proper management and evaluating treatment efficacy. To date, somatostatin receptor scintigraphy (SRS) is considered the gold standard imaging procedure, and very few data are available concerning the use of helical computed tomography (hct). This study aimed to determine the diagnostic accuracy and the ability to modify the surgical management of hct, alone or combined with SRS. Patients and methods: Sixty patients were staged before surgery by hct, SRS and tumour markers, and included in group 1 if suitable for radical surgery, otherwise in group 2. All patients underwent laparotomy followed by subsequent re-staging. Results: SRS sensitivity was 77%, 48% and 67% for primary, lymph-node and liver lesions, respectively. hct sensitivity was 94%, 69% and 94% for primary, lymph-node and liver lesions, respectively (P = 0.02 versus SRS, for liver lesions). During pre-operative evaluation, hct correctly staged 92% and SRS 75% of patients (P = 0.02). hct provided additional information in 17% of patients. Conclusions: Since hct has been shown to be extremely accurate, providing essential information for the planning of surgical treatment compared with that of SRS, both techniques should be used in the pre-operative work-up of digestive endocrine tumours. Key words: digestive endocrine tumours, helical computed tomography, management, somatostatin receptor scintigraphy, staging Introduction Digestive endocrine tumours (dets) are unusual neoplasms, which manifest either as the result of a syndrome related to uncontrolled hormone secretion [functioning (F) tumours], or due to non-specific symptoms related to the mass effect of a lesion [non-functioning (NF) tumours]. On account of the clinical symptoms associated with F tumours, diagnosis is often possible while they are still small and resectable, and before the development of metastatic disease [1], whereas most NF tumours are diagnosed when metastases are already present [2]. Surgery is widely accepted as first-choice therapy for the treatment of dets [3 5], and can play an important role even in selected cases of advanced disease, when it is possible to achieve improvement or resolution of hormone-associated symptoms, improve the quality of life, prolong survival time and possibly cure a patient [3 7]. In fact, even in patients with multiple liver *Correspondence to: G. Delle Fave, Policlinico Umberto I, Dept. Scienze Cliniche, Università di Roma La Sapienza, Roma, Italy. Tel: ; Fax: ; gianfranco.dellefave@uniroma1.it metastases, resection is now considered a safe procedure, and good candidates are not only symptomatic, but also asymptomatic, patients with resectable primary tumour and liver metastases not responding to non-surgical treatment [4]. Furthermore, early aggressive surgery may improve the efficacy of cytoreduction, enhance the response to medical therapy, and offer a chance to manage the primary tumour [5]. Since its introduction for the identification of endocrine tumours [8], several studies have assessed the role of somatostatin receptor scintigraphy (SRS), which is now considered the gold standard imaging procedure in the staging of these patients. In recent years helical computed tomography (hct) technology has been developed. This technique offers a number of potential advantages over conventional systems: absence of respiratory mis-registration, as the scan acquires data during a single breath-hold; scan reconstruction at shorter intervals than scan collimation, with overlapping slices; and optimal contrast enhancement of parenchyma lesions due to short scanning of arterial/portal/venous phases. The benefits of hct in the detection and characterisation of non-endocrine tumours have been well established [9 11], and due to the technical features it is a promising method to visualise dets, which are often hyper European Society for Medical Oncology

2 587 vascularised and, therefore, potentially visible during arterial phase scanning. However, little is known about its use in dets, as studies have usually been performed on small series of patients, and therefore few data are available on NF types, and, moreover, most of the tumours are of pancreatic origin [12 14]. Furthermore, to our knowledge, no comparative studies with SRS have been performed until now. As accurate pre-operative staging of dets is essential for planning proper treatment and evaluating its efficacy, due to the lack of prospective comparative studies, the present study was aimed to determine the role of hct, alone or combined with SRS, in the staging of F and NF dets, and to establish whether hct can modify the surgical management of these patients. Patients and methods Patients Sixty consecutive patients with det were enrolled (27 male/33 female; median age 48 years, range years), 12 of whom had already had primary tumour removed and presented only with metastatic lesions (Table 1). Study design Pre-operative staging was performed within 1 month before surgery by hct, SRS and assessment of tumour markers. Two observers evaluated imaging blindly and independently. Positive concordance between imaging findings was considered when both hct and SRS identified tumour lesions in the same locality. Combined result was used when the lesions were detected by at least one technique (hct or SRS). According to imaging findings, patients were classified as group 1 if suitable for radical surgery (complete excision of all tumour lesions), or group 2 if only non-radical surgery could be attempted. All patients underwent laparotomy after staging. Histological examination of surgically resected tumours was carried out in all cases to confirm diagnosis. Post-operative follow-up at 3 months in all patients included hct and tumour markers, and, at 6 months, SRS. Written informed consent was obtained prior to enrolment in the study, which was approved by local ethics committee. Tumour markers Tumour marker assessment included chromogranin A (CgA), as well as specific tumour markers for functioning tumours. Tumour localisation Somatostatin receptor scintigraphy. Planar images were acquired using a pixel matrix 24 h and 48 h after the injection of [ 111 In]pentreotide. Abdominal 360 single-photon emission computed tomography was carried out in all patients at the end of the 24-h planar study; 120 frames of 25 s each were acquired on a pixel matrix. Images were reconstructed in transaxial, coronal and sagittal orthogonal planes. Helical CT. All CT images were obtained with an hct scanner. 1000ml of diluted (2%) barium per os suspension was administered orally 1 h before the scan to provide positive contrast in the entire small bowel. Patients also received 700 ml of water just before scanning to provide negative contrast in the stomach and duodenum. Glucagon (1 mg) was administered i.v. immediately before scanning to maximise distension of the small bowel. Surgery All patients were considered eligible for surgery with the exception of six in whom unresectable tumour lesions were clearly demonstrated by preoperative imaging. However, in these patients, indications for laparotomy were explorative/diagnostic procedure in order to obtain a definitive diagnosis in the two patients (who did not present liver metastases suitable for biopsy) with previous non-diagnostic percutaneous biopsy of a pancreatic mass, and in the remaining four patients with severe bleeding complications requiring surgical treatment. In agreement with Norton [15], the surgical procedure first included standard methods (visualisation and palpation), followed by intra-operative ultrasound, and transillumination where necessary. Surgical exploration also included extensive Kocher manoeuvres and palpation of the pancreas, after incision of the inferior margin. A standard lymphadenectomy was performed in relation to the primary tumour site. Pathology Diagnosis was confirmed by histology and immunohistochemistry as described elsewhere [16]. Reverse transcription polymerase chain reaction (RT-PCR) was performed on total RNA extract from SRS-negative tumour lesions, and somatostatin receptor subtype 2 (sstr2) using specific human primer pairs was amplified. Statistics The two-tailed Fisher s exact test or McNemar s test was used as appropriate to compare percentages. P values <0.05 were considered statistically significant. Table 1. Specific diagnosis and functional state in relation to primary tumour site Overall Pancreas Duodenum Ileum Rectum Unknown (n = 60 a ) (n = 36) (n = 4) (n = 16) (n = 3) (n = 1) Functioning Gastrinoma Gastrinoma-MEN I 3 3 Carcinoids SSoma 1 1 Glucagonoma 1 1 Non-functioning a Primary tumour has been removed prior to study enrolment in 12 out of 60 patients. MEN I, multi endocrine neoplasia type 1; SSoma, somatostatinoma.

3 588 Results Surgery At the time of enrolment in the study, the primary tumour had already been removed in 12 patients. In the remaining 48 patients, the primary tumour was localised during the surgical procedure in 47 patients [32 in the pancreas and 15 in the gastrointestinal (GI) tract]. In one gastrinoma patient, the primary tumour was not found at surgery and diagnosis was confirmed by histological findings on liver metastases (Table 2). Sensitivity and specificity Primary tumours. Pre-operative imaging procedures for primary tumours showed a sensitivity of 94% with hct (44 of 47) and 77% with SRS (36 of 47) (P = 0.07). The combination of the two methods (hct or SRS positive) increased the sensitivity to 95% (45 of 47) (P = 0.04 versus SRS alone) (Table 3). The positive concordance was 74% (35 of 47). An example of concordant finding at primary tumour imaging with these two procedures is shown in Figure 1. Specificity was 100% and 92% for hct and SRS, respectively (patients in whom primary tumour had been removed were used as negative controls). All pancreatic localisations were identified by hct and 78% (25 of 32) by SRS (P = 0.07), while 12 of 15 of the GI tract localisations (80%) were detected by hct and 11 of 15 (73%) by SRS. Table 2. Surgical findings Positive at surgery (no. of patients) Overall Functioning Non-functioning Primary tumour Pancreas GI tract <1.5 cm cm cm >5 cm Previously removed Not found 1 1 Metastases None Lymph nodes Liver Single lesion lesions >5 lesions Medium CgA levels, U/l % CI CgA, chromogranin A; CI, confidence interval; GI, gastrointestinal. Table 3. Helical CT and SRS sensitivity (%) hct SRS hct or SRS a Primary tumour (n = 47) 94 b c Pancreas (n = 32) 100 d d GI tract (n = 15) <1.5 cm cm cm >5 cm Lymph node (n = 28) e Liver (n = 33) 94 f f Single lesions >5 lesions a At least one is positive. b P = c P = d P = e P = f P = 0.02 versus SRS alone. GI, gastrointestinal; hct, helical computed tomography; SRS, somatostatin receptor scintigraphy. Taking into consideration the functional state of the tumour, hct and SRS identified 16 of 17 (94%) and 14 of 17 (82%) F dets, respectively, and 28 of 30 (93%) and 22 of 30 (73%) NF dets, respectively. As far as the tumour size is concerned, hct was more sensitive compared with SRS in detecting both the smaller (<3 cm: 90% versus 76%) and the larger lesions (>5 cm: 100% versus 82%). The two techniques showed the same sensitivity to detect primary tumours <1.5 cm (87%) (Table 3). Lymph-node metastases. Sensitivity was 69% (20 of 29) for hct, 48% (14 of 29) for SRS (hct versus SRS, P = 0.18) and 76% (22 of 29) by combining the two procedures (P = 0.09 versus SRS alone) (Table 3). Of the 29 patients, 13 (45%) had positive concordant findings. Specificity was 100% with both techniques. Helical CT visualised lymph-node metastases in 70% (seven of 10) of patients with F and in 68% (13 of 19) of patients with NF dets, whereas SRS was positive in 50% (five of 10) of F and 47% (nine of 19) of NF dets, respectively. Liver metastases. Sensitivity was 94% (34 of 36) with hct and 67% (24 of 36) with SRS (P = 0.02) (Table 3). No benefit was observed by combining the two procedures. In 24 patients (67%) positive concordant findings were found. Specificity was 95% with hct, 91% for SRS and 86% for the two techniques combined. Liver tumour lesions were identified by hct in all patients with F and in 91% (21 of 23) of patients with NF dets, and by SRS in 69% (9 of 13) of patients with F and in 65% (15 of 23) of patients with NF dets. Furthermore, hct was more sensitive both in identifying patients with a single lesion (83% versus 50%), as well as those with multiple lesions (86% versus 71% with two to five metastases, 100% versus 69% with more than five lesions) (Table 3).

4 589 Figure 1. Patient with Zollinger Ellison syndrome. (A) Helical computed tomography reveals gastrinoma in duodenal lumen. (B) Somatostatin receptor scintigraphy shows a small intense radiotracer uptake in the duodenal region in transverse single-photon emission computed tomography; (C) anterior and (D) posterior views. T, tumour. Physiological radiotracer uptake: K, kidneys; L, liver; S, spleen; B, bowel. Extra-abdominal metastases. Extra-abdominal lesions were identified by SRS in seven patients, localised in the chest in all of them, and also in the bones in five. All of these patients also had diffuse liver metastases. Sstr2 assessment. To further investigate the negative scintigraphic findings, SRS-negative patients were also studied for the presence of sstr2 mrna by RT-PCR; all results were negative. Predictivity of imaging versus surgical findings and subsequent follow-up Of the 60 patients, 21 (35%) were considered without residual disease after surgery, since they had radical resection (all tumour lesions completely removed), and had negative post-operative follow-up (two patients were false positive at SRS due to the presence of liver angiomas confirmed during the surgical procedure). The remaining 39 (65%) patients still presented residual tumour lesions after surgery, as confirmed by hct in all cases, and by SRS in 35 patients (of the four remaining patients, two were negative also at pre-operative SRS). All but three patients (NF tumours) had positive tumour markers. On the basis of pre-surgery staging, of the 21 patients in whom radical surgery was performed, 19 (by hct) and 15 (by SRS) had been ascribed to group 1 (suitable for radical surgery), whereas of the 39 patients in whom radical surgery was not feasible, 36 (by hct) and 30 (by SRS) had been previously ascribed to group 2 (suitable for non-radical surgery) (Table 4). Thus, overall, preoperative staging was correctly performed by hct and SRS in 55 and 45 patients, respectively (92% versus 75%, P = 0.02). hct provided additional information leading to a modification in the surgical management of 10 patients (17%). On the other hand, incorrect pre-operative staging had been made in five (8%) and 15 (25%) patients, by hct and SRS, respectively. Discussion This is one of the first comparative studies to prospectively investigate the accuracy of hct, alone or combined with SRS, in

5 590 Table 4. Pre-operative staging and predictivity hct SRS Residual disease after surgery No (n = 21) Yes (n = 39) n % n % Group 1 (n = 22) Group 2 (n = 38) Group 1 (n = 24) Group 2 (n = 36) Group 1 and group 2 refer to patients suitable for radical and nonradical surgery, respectively. Predictivity is evaluated comparing preoperative imaging findings versus surgical results and post-surgical follow-up. hct, helical computed tomography; SRS, somatostatin receptor scintigraphy. the pre-surgical staging of a series of dets, comprising mainly NF tumours (62%), localised both in the pancreas and in the GI tract. The first finding emerging from this study is the high diagnostic sensitivity of hct in the localisation of both primary (94%) and metastatic (lymph node 69%, liver 94%) tumour lesions. Thus, the diagnostic yield of hct is higher in comparison to conventional dynamic CT, the sensitivity of which in visualising dets has been reported to range between 29% and 56% [17 20]. In addition, this result is further highlighted by the identification by hct of NF tumours, the detection of which has been reported by other authors to be more difficult [20], since F and NF tumours are similarly localised in the present study. Furthermore, hct sensitivity is high irrespective of the primary tumour size, being 90% in detecting lesions <3 cm. The greater ability of hct may be related to the specific technical features: the arterialphase scan, combined with the absence of respiratory misregistration; the optimal fast contrast enhancement; and the double water/barium contrast determining a better visualisation of endocrine lesions arising from the parenchyma or gut wall. Until now, SRS, as already pointed out, has been considered the gold standard imaging procedure to visualise endocrine tumours. Overall, our findings are in agreement with those of other studies, in which an overall SRS sensitivity ranging from 60% to 100% in the detection of dets has been reported [8, 17, 19 25]. However, compared with our data, a higher SRS sensitivity has been reported both for lymph node (73 100%) and liver metastases (83 92%) [19, 24, 26]. This discrepancy could be explained by various factors. First, the lack of sstr2, the presence of which on the tumour cell surface correlates mainly with positive scintigraphic findings [27]. In the present study, data on disease extension support this hypothesis, since it seems unlikely that 31% of patients with more than five hepatic lesions, as well as 18% of patients with a primary tumour >5 cm, were completely missed by SRS, and the lack of octreotide-preferring somatostatin receptors on these lesions was confirmed by RT- PCR. Secondly, as reported elsewhere [28], SRS is less effective in revealing closely adjacent metastases, sometimes missing some of them, and reducing its ability to precisely define the disease stage. Moreover, other factors need to be taken into account to explain the relatively low SRS sensitivity, such as a modification in local tumour blood supply, unlabelled endogenous somatostatin resulting in competition with, or downregulation of, the somatostatin receptors, and expression of receptor subtypes pattern with less affinity for the radioligand [27, 29]. Furthermore, the different characteristics of the patients enrolled, in the present study, could also play a role, since our series included a relatively large number of patients with NF tumours (62%), in contrast to most other studies, which were carried out on patients with F tumours such as Zollinger Ellison syndrome and carcinoids [19 25]. Helical CT played a prominent role in pre-operative staging of dets. In fact, this technique showed a greater ability to predict the surgical outcome of patients with respect to SRS (92% versus 75%, P = 0.02), providing additional information useful in modifying the surgical management of 10 patients (17%). This role would appear to depend upon its ability to detect liver metastases (P = 0.02 versus SRS), the presence of which is predictive of poor prognosis, and to identify the site of primary and metastatic lesions and their relationship with the surrounding structures (such as vessels), which greatly affect the surgical strategy. Analysing the combined imaging findings, a negligible improvement is observed compared to hct alone, even if the combined primary tumour sensitivity became statistically significant (P = 0.04) compared with SRS alone. In conclusion, this study shows that hct is the most reliable method for accurate pre-operative staging of dets, on account of its ability to detect the primary tumour, whether localised in the pancreas or in the GI tract, and metastatic lesions. However, SRS is a total-body scan, and is needed to identify somatostatin receptors on tumour cells, the presence of which offers information critical for the medical treatment of these patients, allowing the use of either unlabelled or radiolabelled somatostatin analogues. On the basis of these considerations, since complete preoperative staging and subsequent management should include identification and characterisation of primary and metastatic lesions, as well as demonstration of the presence of somatostatin receptors on the surface of tumour cells, both hct and SRS should be performed in all patients with dets prior to starting any treatment. Acknowledgements This study was supported by interuniversity co-financed programme (1999) from the Italian Ministry for University and Technological Research (MURST), by a grant from Ministero Salute ICS 060.2/ RF00-57 (2000) Roma, Italy, and by Fondazione Italiana per le Malattie Digestive (FIMAD). References 1. Lloyd RV. Endocrine pancreas. In Lloyd RV (ed): Endocrine Pathology. New York, NY: Springer Verlag 1990;

6 Lloyd RV. Neuroendocrine cell and neoplasms of the gastrointestinal tract. In Lloyd RV (ed): Endocrine Pathology. New York, NY: Springer Verlag 1990; Nave H, Mossinger E, Feist H et al. Surgery as primary treatment in patients with liver metastases from carcinoid tumors: a retrospective, unicentric study over 13 years. Surgery 2001; 129: Jaeck D, Oussoultzoglou E, Bachellier P et al. Hepatic metastases of gastroenteropancreatic neuroendocrine tumors: safe hepatic surgery. World J Surg 2001; 25: Chung MH, Pisegna J, Spirt M et al. Hepatic cytoreduction followed by a novel long-acting somatostatin analog: a paradigm for intractable neuroendocrine tumors metastatic to the liver. Surgery 2001; 130: Nagorney DM, Florencia GQ. Cytoreductive hepatic surgery for metastatic gastrointestinal neuroendocrine tumors. In Mignon M, Jensen RT (eds): Endocrine Tumors of the Pancreas: Recent Advances in Research and Management. Front Gastrointestinal Res. Basel, Switzerland: Karger 1995; 23: Eriksson B, Oberg K. Neuroendocrine tumours of the pancreas. Br J Surg 2000; 87: Krenning EP, Bakker WH, Breeman WAP et al. Localisation of endocrine-related tumours with radioiodinated analogue of somatostatin. Lancet 1989; 1: Mertz HR, Sechopoulos P, Delbeke D et al. EUS, PET, and CT scanning for evaluation of pancreatic adenocarcinoma. Gastrointest Endosc 2000; 52: Davies J, Chalmers AG, Sue-Ling HM et al. Spiral computed tomography and operative staging of gastric carcinoma: a comparison with histopathological staging. Gut 1997; 41: Valls C, Andía E, Sánchez A et al. Hepatic metastases from colorectal cancer: preoperative detection and assessment of resectability with helical CT. Radiology 2001; 218: Van Hoe L, Gryspeerdt S, Marchal G et al. Helical CT for the preoperative localization of islet cell tumors of the pancreas: value of arterial and parenchymal phase images. Am J Roentgenol 1995; 165: Ichikawa T, Peterson MS, Federle MP et al. Islet cell tumor of the pancreas: biphasic CT versus MR imaging in tumor detection. Radiology 2000; 216: Procacci C, Carbognin G, Accordini S et al. Nonfunctioning endocrine tumors of the pancreas: possibilities of spiral CT characterization. Eur Radiol 2001; 11: Norton JA. Intra-operative procedures to localize endocrine tumours of the pancreas and duodenum. Ital J Gastroenterol Hepatol 1999; 31 (Suppl 2): S195 S Solcia E, Klöppel G, Sobin LH in collaboration with nine pathologists from four countries. Histological Typing of Endocrine Tumors, 2nd edition. World Health Organization. Berlin: Springer Verlag 2000; Corleto VD, Scopinaro F, Angeletti S et al. Somatostatin receptor localization of pancreatic endocrine tumors. World J Surg 1996; 20: Jensen RT, Gardner JD. Gastrinoma. In Go VLW, Dimagno EP, Gardner JD et al. (eds): The Pancreas: Biology, Pathobiology, and Disease, 2nd edition. New York, NY: Raven Press 1993; Gibril F, Reynolds JC, Doppman JL et al. Somatostatin receptor scintigraphy: its sensitivity compared with that of other imaging methods in detecting primary and metastatic gastrinomas. A prospective study. Ann Intern Med 1996; 125: Zimmer T, Ziegler K, Bader M et al. Localisation of neuroendocrine tumor of the upper gastrointestinal tract. Gut 1994; 35: Modlin IM, Tang LH. Approaches to the diagnosis of gut neuroendocrine tumors: the last word (today). Gastroenterology 1997; 112: Jensen RT, Gibril F. Somatostatin receptor scintigraphy in gastrinomas. Ital J Gastroenterol Hepatol 1999; 31 (Suppl 2): S179 S Lebtahi R, Cadiot G, Sarda L et al. Clinical impact of somatostatin receptor scintigraphy in the management of patients with neuroendocrine gastroenteropancreatic tumors. J Nucl Med 1997; 38: Cadiot G, Lebtahi R, Sarda L et al. Preoperative detection of duodenal gastrinomas and peripancreatic lymph nodes by somatostatin receptor scintigraphy. Gastroenterology 1996; 111: Cadiot G, Bonnaud G, Lebtahi R et al. Usefulness of somatostatin receptor scintigraphy in the management of patients with Zollinger Ellison syndrome. Groupe de Recherche et d Etude du Syndrome de Zollinger Ellison (GRESZE). Gut 1997; 41: Scherubl H, Bader M, Fett U et al. Somatostatin receptor imaging of neuroendocrine gastroenteropancreatic tumors. Gastroenterology 1993; 105: John M, Meyerhof W, Richter D et al. Positive somatostatin receptor scintigraphy correlates with the presence of somatostatin receptor subtype 2. Gut 1996; 38; Termanini B, Gibril F, Reynolds JC et al. Value of somatostatin receptor scintigraphy: a prospective study in gastrinoma and its effect on clinical management. Gastroenterology 1997; 112: Balon HR, Goldsmith SJ, Siegel BA et al. Procedure guideline for somatostatin receptor scintigraphy with (111)In-Pentreotide. J Nucl Med 2001; 42;