Predictive and prognostic factors for treatment and survival in 305 patients with advanced gastrointestinal neuroendocrine carcinoma (WHO G3): the NORDIC NEC study.

BACKGROUND: As studies on gastrointestinal neuroendocrine carcinoma (WHO G3) (GI-NEC) are limited, we reviewed clinical data to identify predictive and prognostic markers for advanced GI-NEC patients. PATIENTS AND METHODS: Data from advanced GI-NEC patients diagnosed 2000-2009 were retrospectively registered at 12 Nordic hospitals. RESULTS: The median survival was 11 months in 252 patients given palliative chemotherapy and 1 month in 53 patients receiving best supportive care (BSC) only. The response rate to first-line chemotherapy was 31% and 33% had stable disease. Ki-67<55% was by receiver operating characteristic analysis the best cut-off value concerning correlation to the response rate. Patients with Ki-67<55% had a lower response rate (15% versus 42%, P<0.001), but better survival than patients with Ki-67≥55% (14 versus 10 months, P<0.001). Platinum schedule did not affect the response rate or survival. The most important negative prognostic factors for survival were poor performance status (PS), primary colorectal tumors and elevated platelets or lactate dehydrogenase (LDH) levels. CONCLUSIONS: Advanced GI-NEC patients should be considered for chemotherapy treatment without delay.PS, colorectal primary and elevated platelets and LDH levels were prognostic factors for survival. Patients with Ki-67<55% were less responsive to platinum-based chemotherapy, but had a longer survival. Our data indicate that it may not be correct to consider all GI-NEC as one single disease entity.

Anemia in patients with midgut carcinoid, treated with alpha interferon: effects by erythropoietin treatment on the perceived quality of life.

One important side effect from alpha interferon is depression of bone marrow function and studies have shown that patients with carcinoid tumours treated with alpha interferon suffers from fatigue and impaired physical functions. The aim of this pilot study was to investigate if treatment with erythropoietin (EPO) could have a positive effect on self-rated quality of life (QoL). Eighteen patients with midgut carcinoid treated with alpha interferon were included in the study. There were statistical significant increases in haemoglobin (Hb) levels between baseline and 4 months, between baseline and 8 months as well as between baseline and 2-year follow-up. No EPO related side effects were reported. There were improvements of more than 10 points in self-rated QoL-issues related to anaemia. Even though the analysis did not reveal any statistically significant relation between the observed increase in Hb levels and self-rated QoL, this pilot study has increased the knowledge about benefits, doses and frequency of EPO treatment in patients with midgut carcinoid suffering from interferon related anaemia.

Regulation of insulin and glucagon secretion from rat pancreatic islets in vitro by somatostatin analogues.

Somatostatin is an inhibitor of hormone secretion through specific receptors (sst1-5). The aim of this study was to investigate the putative regulatory role of somatostatin analogues on the secretion of insulin and glucagon by rat pancreatic islets. After 48 h exposure only the non-selective agonists (somatostatin, octreotide and SOM-230) inhibited insulin accumulation. The inhibition of insulin secretion was accompanied by increased islet insulin contents. None of the analogues showed a consistent effect on the glucagon accumulation in the medium after 48 h. Since we observed a difference in the regulatory effect between the non-selective and selective analogues, combinations of selective analogues were studied. Combination of sst2+sst5 agonists inhibited the medium insulin accumulation, while combination of sst1+sst2 analogues caused a decrease in glucagon accumulation. After removal of somatostatin a rebound effect with increased insulin secretion were observed. This effect was reversed after 6 h. For SOM-230 insulin secretion continued to be suppressed even after the analogue was removed and returned to control values after 3 h. As for glucagon secretion there was an initial decline after culture with octreotide, while the other substances failed to induce any changes. In summary, non-selective somatostatin analogues or combinations of receptor selective analogues may cause inhibition of hormone secretion from rat pancreatic islets. For insulin and glucagon, combinations of sst2+sst5 and sst1+sst2, respectively may exert this effects. Thus, our data suggest that more than one sst must be involved to down-regulate islet glucagon and insulin secretion.

A comparison between the efficacy of somatostatin receptor scintigraphy and that of in situ hybridization for somatostatin receptor subtype 2 messenger RNA to predict therapeutic outcome in carcinoid patients.

Predictive tests for treatment with somatostatin analogues have been asked for by clinicians. We have shown previously that somatostatin receptor (sstr) scintigraphy may be used to predict therapeutic outcomes for carcinoid patients receiving somatostatin analogues. However, almost 20% of patients with pathological tracer uptake fail to respond to such treatment. To increase further the reliability and prognostic value of sstr identification, we investigated the presence of mRNA for the subtypes sstr1 and sstr2 by in situ hybridization on tumor specimen from 25 carcinoid patients (22 midgut, 2 foregut, and 1 hindgut), all receiving somatostatin analogue treatment (12 lanreotide, 8 octreotide, and 5 octastatin) and compared this to the therapeutic response evaluated as inhibition of hormone secretion. Expression of sstr2 mRNA could be detected in 15 patients, all responding to somatostatin analogue treatment and showing pathological tracer uptake in tumor lesions at sstr scintigraphy. In the remaining 10 patients, no sstr2 mRNA could be detected, and none of the patients responded to somatostatin analogue treatment. Three of these 10 patients failed to accumulate tracer activity at sstr scintigraphy, whereas 7 had a pathological uptake of [(111)In-DTPA-D-Phe(1)]-octreotide. We conclude that in this group of carcinoid patients, there was complete agreement between the presence of mRNA for sstr2 detected by in situ hybridization and therapeutic outcome. In patients with pathological tracer accumulation without expression of somatostatin sstr2 mRNA, other sstr may be present that can bind the somatostatin analogue but not inhibit hormone secretion.

Determination of somatostatin receptor subtype 2 in carcinoid tumors by immunohistochemical investigation with somatostatin receptor subtype 2 antibodies.

We have shown previously that expression of mRNA for somatostatin receptor subtype 2 (sst2) detected by in situ hybridization correlates to therapeutic outcome in patients with carcinoid tumors treated with somatostatin analogues. However, in situ hybridization is laborious and not practical in clinical routine work. We have, therefore, developed polyclonal antibodies directed against sst2 that may be used for immunohistochemistry on tissue specimens. The staining is specific and is highly correlated to expression of mRNA for sst2 (P < 0.01) as well as to tracer uptake at somatostatin receptor scintigraphy (P < 0.01). There is also a good correlation to the therapeutic response in carcinoid patients treated with somatostatin analogues (P < 0.05). Of 35 patients with carcinoid tumors included in this investigation, 25 stained positive with the antibodies. Twenty-two of these were investigated by somatostatin receptor scintigraphy and showed tracer uptake in metastases. An additional two patients that did not stain with the antibodies showed pathological uptake of the tracer in metastases, which might indicate binding to somatostatin receptor subtype 5. None of the 10 patients without positive immunostaining responded to somatostatin analogue treatment, whereas patients with a positive stain had a biochemical response or remained stable during treatment. Thus, these antibodies may be used to determine the presence of sst2 in carcinoid tumors and to select patients suitable for somatostatin analogue treatment. The method is easily applicable in clinical practice.

Somatostatin receptor scintigraphy in patients with carcinoid tumors: comparison between radioligand uptake and tumor markers.

We have performed 100 scintigraphic investigations using [111In-diethylenetriaminepentaacetic acid-D-Phe1]octreotide (111In-octreotide) single photon emission tomography (SPECT) in patients with carcinoid tumors. One or several lesions could be detected in 77 cases, and true negative results were obtained in 11 cases. There were false-negative results in 12 cases compared with results from conventional radiological methods. The ratio between the SPECT signals from the area with the highest uptake and normal lung was used as a tumor:background ratio. An attenuation correction was made in all investigations. We found that lesions in untreated patients had lower tumor:background ratios compared with those in patients treated with somatostatin analogues (medians, 10 and 40, respectively; P < 0.001) or IFN (median, 23; P = 0.03). In untreated patients, there was a correlation between the tumor:background ratio and the levels of urinary 5-hydroxyindoleacetic acid (U-5HIAA) and p-chromogranin A. The data obtained in the present investigation indicate that somatostatin receptor expression might be influenced by the treatment; i.e., a higher tumor:background ratio is found in patients treated with either somatostatin analogues or IFN. Furthermore, it was found that somatostatin receptor expression correlates with the levels of U-5HIAA and p-chromogranin A in untreated patients, and that 111In-octreotide SPECT scintigraphy is more likely positive in patients with elevated U-5HIAA values. This indicates that somatostatin receptor expression and elevated U-5HIAA are more likely present in patients with highly differentiated tumors and, thus, could be of prognostic value.

Expression of somatostatin receptor subtypes 1-5 in pancreatic islets of normoglycaemic and diabetic NOD mice.

OBJECTIVE: Somatostatin acts on five specific receptors (sst1-5) to elicit different biological functions. The non-obese diabetic (NOD) mouse is an experimental model of type 1 diabetes. The aim of this study was to investigate whether the islet expression of sst1-5 is affected during the development of diabetes in NOD mice, with insulitis accompanied by spontaneous hyperglycaemia. METHODS: By immunostaining for sst1-5 the expression and co-expression together with the four major islet hormones in pancreatic islets were investigated in female and male NOD mice at different stages of disease. The NOD related non-diabetic ICR mouse was also examined. RESULTS: The islet cells of diabetic NOD mice showed an increased islet cell expression of sst2-5 compared with normoglycaemic female NOD mice. This correlated to increasing age and extent of insulitis. Major findings from the co-expression investigations were that sst2 was expressed in a majority of beta-cells in the normoglycaemic NOD mice, but absent in the beta-cells in the diabetic NOD mice. A majority of the alpha-cells expressed sst2 and 5 in normoglycaemic and diabetic NOD mice. About 60% of delta-cells showed co-expression of sst4 and 5 in both normoglycaemic and diabetic NOD mice. 60% of pancreatic polypeptide (PP)-cells expressed sst4 in both groups. Insulitis was found to be accompanied by a down-regulation of sst in normoglycaemic animals. CONCLUSIONS: The difference in sst expression in the islets cells of diabetic mice may suggest either a contributing factor in the process leading to diabetes, or a defence response against ongoing beta-cell destruction.

Novel long-acting somatostatin analog with endocrine selectivity: potent suppression of growth hormone but not of insulin.

Somatostatin, also known as somatotropin release-inhibiting factor (SRIF), is a natural cyclic peptide inhibitor of pituitary, pancreatic, and gastrointestinal secretion. Its long-acting analogs are in clinical use for treatment of various endocrine syndromes and gastrointestinal anomalies. These analogs are more potent inhibitors of the endocrine release of GH, glucagon, and insulin than the native SRIF; hence, they do not display considerable physiological selectivity. Our goal was to design effective and physiologically selective SRIF analogs with potential therapeutic value. We employed an integrated approach consisting of screening of backbone cyclic peptide libraries constructed on the basis of molecular modeling of known SRIF agonists and of high throughput receptor binding assays with each of the five cloned human SRIF receptors (hsst1-5). By using this approach, we identified a novel, high affinity, enzymatically stable, and long-acting SRIF analog, PTR-3173, which binds with nanomolar affinity to human SRIF receptors hsst2, hsst4, and hsst5. The hsst5 and the rat sst5 (rsst5) forms have the same nanomolar affinity for this analog. In the human carcinoid-derived cell line BON-1, PTR-3173 inhibits forskolin-stimulated cAMP accumulation as efficiently as the drug octreotide, indicating its agonistic effect in this human cell system. In hormone secretion studies with rats, we found that PTR-3173 is 1000-fold and more than 10,000-fold more potent in inhibiting GH release than glucagon and insulin release, respectively. These results suggest that PTR-3173 is the first highly selective somatostatinergic analog for the in vivo inhibition of GH secretion, with minimal or no effect on glucagon and insulin release, respectively.

Somatostatin receptor ligands and their use in the treatment of endocrine disorders.

Somatostatin is a tetradeca peptide hormone produced by many different endocrine cells throughout the body. It is also present in both the central and peripheral nervous system. The peptide has many different moods of action including inhibition of hormone secretion and influence on gastrointestinal motility. Somatostatin was identified in 1973 and about 10 years later the first long-acting analogue, octreotide, became available for use in clinical trials. Somatostatin analogues have been used to treat patients with neuroendocrine tumors, such as carcinoid tumors and endocrine pancreatic tumors, with symptoms due to excessive hormone production. It also has a well-documented effect on hormone levels and symptoms in acromegalic patients, while the use in diabetes mellitus is less well established. Several new analogues have been developed and tested for clinical use and lately non-peptide analogues have been produced. These new somatostatin receptor subtype specific analogues will soon be tested in clinical trials. In this review article the development of new analogues and new preparations of old analogues and their use in the clinic is discussed.

Expression of p68 protein kinase and its prognostic significance during IFN-alpha therapy in patients with carcinoid tumours.

The aim of this study was to evaluate the antiproliferative effects of interferon alpha (IFN-alpha) on neuroendocrine differentiated cell lines and, retrospectively, to assess the prognostic significance of p68 protein kinase (PKR) induction in neuroendocrine gut and pancreatic tumour patients. Archive specimens from 56 patients were studied, 43 before IFN-alpha and 56 during therapy. The tissues were immunostained for p68 protein kinase (PKR) using the monoclonal antibody (MAb) TJ4C4. A significant increase in immunostaining after treatment with IFN-alpha compared with before treatment (3.47 +/- 0.12 versus 2.72 +/- 0.15, P < 0.001) was noted. The p68 score was significantly increased after treatment only in patients with stable disease before = 2.71 +/- 0.19, after = 3.40 +/- 0.14 (P < 0.001) or an objective response before 3.13 +/- 0.22, after = 4.00 +/- 0.24 (P < 0.05) but not in those with progressive disease (before = 2.32 +/- 0.24, after 2.86 +/- 0.26, NS). A low p68 score (< 3.0) during treatment was a predictor of shorter duration of response and overall survival (P = 0.0062 and P < 0.0001, respectively). Furthermore, IFN-alpha showed a significant antiproliferative effect (by [3H]thymidine incorporation) on two carcinoid tumour cell lines in a dose-dependent manner which correlated with a dose-dependent induction of p68 mRNA and protein expression (by Northern and Western blot analysis). We conclude that IFN-alpha can effectively inhibit the in vitro growth of carcinoid tumor cell lines and upregulates the expression of p68 at both mRNA and protein levels in carcinoid tumours. The induction of p68 could be a prognostic indicator of response in patients with carcinoid tumours during IFN-alpha treatment.

Nuclear localization of 111In after intravenous injection of [111In-DTPA-D-Phe1]-octreotide in patients with neuroendocrine tumors.

UNLABELLED: Treatment with tumor-targeting substances is currently being evaluated in clinical trials. For patients with neuroendocrine tumors expressing somatostatin receptors, the 111In-labeled somatostatin analog [diethylenetriaminepentaacetic acid (DTPA)-DPhe1]-octreotide has been used with promising results. To further investigate the clinical effect of the injected conjugate, we analyzed the cellular distribution of 111In by ultrastructural autoradiography. METHODS: Seven patients with somatostatin receptor-expressing midgut carcinoid tumors scheduled for abdominal surgery were investigated by somatostatin receptor scintigraphy. During operation, tumor tissue samples and samples of normal intestine were collected, fixed, and processed for electron microscopy. A thin layer of film emulsion was applied on sections and after the exposure film was developed. The cellular distribution of silver precipitations indicating the presence of isotope was evaluated. RESULTS: Cell surface receptor binding and internalization of [111In-DTPA-D-Phe1]-octreotide in the tumor cells was easily revealed by silver precipitations in the film. Multiple silver grains were seen at the plasma membrane, in the cytoplasmic area among secretory granules and vesicular compartments, and in the perinuclear area. Silver grains were also regularly located in the nucleus. For all patients, the silver precipitation patterns from 111In decay were identical in all examined cells from removed tumors, and in most cells 111In could be seen in the nucleus. The specificity of the silver reaction products is supported by the observation that enterocytes in intestinal tissue specimens from near the tumor did not show any silver grains and no background labeling was seen in the plastic. CONCLUSION: After internalization through the somatostatin receptor system, 111In is translocated to the perinuclear area and into the nucleus. Whether the nuclide is still conjugated to the intact somatostatin analog or to part of it cannot be evaluated in this study. Despite the short irradiation range of 111In, the nuclear localization can explain its clinical effectiveness. The results from this study suggest that [111In-DTPA-D-Phe1]-octreotide may act as a powerful tumor cell-targeting substance.

A comparison between three commercial kits for chromogranin A measurements.

Chromogranin (CgA) has been shown to be an excellent marker for neuroendocrine tumours. There are now three commercial assays on the market. We wanted to compare the usefulness of the different kits in a clinical situation. We have thus measured CgA in 77 patients and compared the results from the different methods. CgA was measured with three different commercial kits according to the recommendations from the manufacturers (CGA-RIA CT; CIS bio international, Gif-sur-Yvette cedex, France, DAKO Chromogranin A ELISA kit; DAKO A/S, Glostrup, Denmark and CgA; EuroDiagnostica, Malmö, Sweden). The sensitivity and specificity differed between the different kits. The CIS kit showed a sensitivity of 67% and a specificity of 96%. The sensitivity and specificity were both 85% for the DAKO kit and 93% and 88% respectively for the EuroDiagnostica assay. We have concluded that CgA is an important tumour marker for all neuroendocrine tumours. However, different analytical properties of the respective kits give different performances, a fact that must be taken into consideration when comparing results from different clinical studies. A recognised international standard for CgA would be a step on the way to harmonisation, but antibody selection and construction of the assays will probably still influence the results.

Transmembrane protein tyrosine phosphatase IA-2 (ICA512) is expressed in human midgut carcinoids but is not detectable in normal enterochromaffin cells.

A potential upregulation of receptor type protein tyrosine phosphatase IA-2 (ICA512) expression was detected by differential display and investigated in midgut carcinoid tumours. Normal intestine tissue and tumour tissue from 13 midgut carcinoid patients were studied by in situ hybridisation using an IA-2 ribonucleotide probe and confocal microscopy using specific IA-2 antibodies. Previously, it had been shown that IA-2 is located in the secretory granules of virtually all neuroendocrine cells. However, we found that IA-2 was not detectable in resting normal enterochromaffin (EC) cells of the small intestine, while high expression of IA-2 mRNA and protein was confirmed in both primary and metastatic carcinoid tissue. This difference in expression was not observed with chromogranin A or serotonin, two secretory granule hormones known to be expressed in EC cells, indicating that IA-2 was seemingly not necessary for the basal production and packaging of these hormones. When comparing patients receiving biotherapy before operation with untreated patients, we found expression of IA-2 to be lower in tumours from patients that had been treated with a combination of alpha-interferon and the somatostatin analogue, octreotide. There was no correlation between IA-2 expression and proliferation rates as measured by immunohistochemistry with antibodies against the Ki 67 antigen. Furthermore, we show that IA-2 is co-localised with serotonin in carcinoid tumours as well as in the pancreatic tumour cell line, BON1, which is interesting as serotonin secretion rate is presumably higher in tumour cells than in resting EC cells. Taken together, these findings may indicate a role for IA-2 in the later stages of the regulated secretory process.

[111In-DTPA-D-Phe1]octreotide scintigraphy in patients with carcinoid tumours: the predictive value for somatostatin analogue treatment.

This study was performed to evaluate whether the presence or absence of somatostatin receptors in malignant carcinoid tumours detected by [111In-DTPA-D-Phe1]octreotide scintigraphy can be used to predict response to somatostatin analogue treatment. Thirty patients were investigated, 28 with midgut carcinoid tumours and two with foregut carcinoid tumours. Twenty-seven patients showed pathological uptake in tumour lesions at scintigraphy; of these, 22 responded to somatostatin analogue treatment using octreotide, somatuline or octastatin, while five patients failed to respond. None of the three patients displaying negative scintigraphic investigations responded to treatment with somatostatin analogues. These results show a good correlation between the somatostatin receptor status and the patients' ability to respond to somatostatin analogue treatment (p = 0.014). We conclude that somatostatin receptor scintigraphy using [111In-DTPA-D-Phe1]octreotide can be used to select patients with malignant carcinoid tumours suitable for somatostatin analogue treatment and exclude those that will not benefit from such medication.

Somatostatin receptor scintigraphy during treatment with lanreotide in patients with neuroendocrine tumors.

To investigate possible changes in somatostatin receptor expression during treatment with high dose lanreotide, eight patients with neuroendocrine tumors were investigated by [(111)In-DTPA-D-Phe1]-octreotide scintigraphy before and during treatment. The spleen-to-background ratio decreased in all patients, whereas tumor-to-background ratio revealed a heterogeneous pattern with an average increase of 50% (-79% to +1,087%). This finding indicates that lanreotide treatment may influence the binding of radioactively labeled somatostatin to the spleen, while changes in the binding to functioning somatostatin receptors in tumor cells are more complex and not clearly related to treatment.

Expression of the five different somatostatin receptor subtypes in endocrine cells of the pancreas.

Knowledge concerning tissue-specific expression of the five somatostatin receptor subtypes is of great importance in understanding their physiological function. We developed rabbit polyclonal antibodies specific for each human somatostatin receptor subtype and report our results concerning the expression in normal endocrine pancreatic cells. The antibodies were produced by immunizing rabbits with fragments specific for the five cloned somatostatin receptor subtypes. Colocalization of these somatostatin receptors with the four major islet hormones--insulin, glucagon, somatostatin, and pancreatic polypeptide--was studied in normal human endocrine pancreatic cells, using double-immunofluorescence staining. High expression of somatostatin receptor subtypes 1, 3, and 4 was found in all endocrine pancreatic cells. Somatostatin receptor subtype 2 was frequently expressed in alpha and beta cells, whereas expression was low in pancreatic polypeptide cells and intermediate in delta cells. Somatostatin receptor subtype 5 was expressed in most beta and delta cells but almost absent in alpha and pancreatic polypeptide cells. There is a variability in the normal expression of somatostatin receptor subtypes among the different human endocrine pancreatic cells. Knowledge of this expression and the physiological function mediated by these receptors will be valuable in the future when considering treatment of endocrine disorders.

Induction of MxA mRNA in patients with neuroendocrine tumors after interferon treatment. Lack of correlation with antitumor response.

We have investigated the possibility of using induction of MxA mRNA in patients with neuroendocrine tumors undergoing interferon-alpha(IFN-alpha) treatment as a predictive test for antitumor effect. A total of 122 patients with various types of neuroendocrine tumors were included in the study. Blood samples were drawn 12 hours after administration of either recombinant or natural IFN-alpha for analysis of induction of MxA mRNA in peripheral blood leukocyte (PBLs). Total RNA was isolated and slot-blot hybridization was performed using a MxA cDNA fragment as a probe. All patients displayed induction of MxA mRNA. Out of 13 untreated patients, 4 had MxA mRNA as well as 2 out of 11 patients treated with somatostatin analogue. All IFN-alpha treated patients showed induction of MxA mRNA and there was no difference between patients demonstrating partial remission, stable disease or progressive disease. Moreover, there was no difference between recombinant or natural leukocyte IFN-alpha in the ability to induce MxA mRNA. Six patients developed neutralizing IFN-alpha antibodies with 1 patient presenting a tire of 3200 NU/ml. Development of neutralizing antibodies did not abrogate the induction of MxA mRNA, but in 3 patients with high antibody titres the antitumor effect was lost. We therefore conclude that IFN-alpha is able to induce MxA mRNA in PBLs from patients with neuroendocrine tumors but we could not find any correlation with the therapeutic outcome. Furthermore, development of neutralizing IFN-alpha antibodies, although abrogating the antitumor effect, might not block the antiviral activity.

Levels of angiogenic peptides in sera from patients with carcinoid tumours during alpha-interferon treatment.

BACKGROUND: Alpha-interferon, a known inhibitor of angiogenesis and cell proliferation, is used in the standard treatment of patients with carcinoid tumors. We studied the levels of two angiogenic peptides (bFGF and VEGF) in sera from patients with carcinoid tumours before and during treatment with alpha-interferon. The aim was to investigate if the antitumoral effect of alpha-interferon in these patients could be at least in part explained by a reduction in the measured angiogenetic peptides. PATIENTS AND METHODS: Sera from 29 patients with carcinoid tumours were collected before and during alpha-interferon treatment and analyzed using commercially available ELISA-kits. RESULTS: Interferon alpha treatment did not cause reduction of bFGF and VEGF levels in serum from patients with carcinoid tumours. In fact there was no correlation between changes in bFGF or VEGF levels and treatment effect. CONCLUSION: The action of alpha-interferon does not seem to be mediated by bFGF or VEGF in patients with carcinoid tumours. If alpha-interferon has an anti-angiogenic effect in this patient group, it is probably mediated by angiogenic peptides other than bFGF and VEGF.

Treatment of malignant endocrine pancreatic tumors with a combination of alpha-interferon and somatostatin analogs.

Somatostatin analogs and alpha-interferon induce good responses as single drugs in the treatment of endocrine pancreatic tumors. We examined the efficacy and tolerability of the combination of alpha-interferon and somatostatin analogs in 16 patients with metastatic endocrine pancreatic tumors. All patients except one had received prior treatment and were in a progressive state. Doses of alpha-interferon and somatostatin analogs were individually titrated. The alpha-interferon doses varied between 9 and 25 million units per week and were combined with 100-1500 microg of octreotide or 6000 microg of lanreotide daily. Radiological response was seen in 3 of 16 (19%) patients (median duration 23 mo). Biochemical response was seen in 10 of 16 (62.5%) patients (median duration 22 mo). All three patients previously progressing on both alpha-interferon and somatostatin analog as single drugs achieved a stabilization of the disease when treated with the combination (median duration 10 mo). Seven of eight (88%) patients previously progressing on alpha-interferon treatment benefited from the combination with biochemical partial response or stabilization. All six patients previously progressing during somatostatin analog treatment achieved biochemical partial response or stabilization. More than 80% of patients who progressed during previous treatment with either drug benefited from the combined treatment, which also was well tolerated. Thus, a combination of alpha-interferon and somatostatin analogs may be considered for patients previously progressing on treatment with alpha-interferon or somatostatin analogs. However, in this study, the value of sequential treatment has not been evaluated.