0021-972X/99/$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1999 by The Endocrine Society
Vol. 84, No. 1 Printed in U.S.A.
Screening for Multiple Endocrine Neoplasia Type 1 and Hormonal Production in Apparently Sporadic Neuroendocrine Tumors* ERIC BAUDIN, JEAN-MICHEL BIDART, PHILIPPE ROUGIER, VLADIMIR LAZAR, ´ , JACQUES ROPERS, MICHEL DUCREUX, FRE ´ DE ´ RIC TROALEN, PIERRE RUFFIE JEAN-CHRISTOPHE SABOURIN, ETIENNE COMOY, PHILIPPE LASSER, THIERRY DEBAERE, AND MARTIN SCHLUMBERGER Nuclear Medicine (E.B., M.S.), Clinical Biology (J.M.B., V.L., F.T., E.C.), Medicine (Ph.R., P.R., M.D.), Biostatistics and Epidemiology (J.R.), Pathology (J.-C.S.), Surgery (P.L.), and Interventional Radiology (T.D.), Institut Gustave-Roussy, 94805 Villejuif Cedex, France ABSTRACT Screening was performed in 130 consecutive patients with apparently sporadic neuroendocrine tumors (NET) to assess the prevalence of multiple endocrine neoplasia type 1 (MEN1) and hormonal production. Screening for MEN1 included measurement of serum calcium and PTH [PTH-(1– 84)], gastrin, PRL, and insulin-like growth factor type I (IGF-I) levels. MEN1 genetic testing was performed in patients with two components of the MEN1 syndrome. Screening for hormonal production included measurement of serum neuron-specific enolase (NSE), calcitonin (CT), glycoprotein a-subunit (GPa), hCG b-subunit (free hCGb), and somatostatin levels. Twenty-four-hour urinary free cortisol (UFC) and 5-hydroxyindolacetic acid (5-HIAA) determinations were also performed. Four patients had hyperparathyroidism, none of whom had pituitary or familial disease. Hyperprolactinemia was compatible with a pituitary disease in one patient. No acromegalic feature or any increase in IGF-I was found. Hypergastrinemia, compatible with an associated pancreatic NET, was
found in one patient. Genetic screening of the MEN1 gene was performed in five of the six patients with two components of the MEN1 syndrome. A nonsense mutation (Arg108stop) was identified in the tumor of one patient. Elevated NSE, 5-HIAA, CT, GPa, free hCGb, SMS, and nonsuppressible UFC were found in 47%, 46%, 14%, 19%, 12%, 3%, and 6% of NET patients, respectively. Production of CT, GPa, and free hCGb was highly related to the primary site: all but two of these secretions originated in foregut NET. 5-HIAA secretion was found in 27% of foregut-derived and 85% of midgut-derived NET. In conclusion, MEN1 is a rare event in patients presenting with apparently sporadic NET. It occurred mainly in foregut NET and should be screened for by serum calcium and PTH-(1– 84) measurements. Routine hormonal measurements should depend on the primary site. NSE, 5-HIAA, CT, and aGP should be routinely measured in foregut-derived NET; only serum NSE and 5-HIAA measurements are recommended in midgut-derived NET. (J Clin Endocrinol Metab 84: 69 –75, 1999)
N
eral markers such as NSE and chromogranin A may also be of interest during the work-up of NET (3). Hormonal screening in GEP NET is influenced by the clinical presentation, standard biological results, and knowledge of the primary site (4). Even in NET devoid of clinically apparent symptoms or eutopic hormone secretions, the production of hormones may be demonstrated. Hormonal production in GEP NET can reflect secretion by the primary tumor and its metastases or by an associated NET in the context of the multiple endocrine neoplasia type 1 (MEN1) syndrome. MEN1 syndrome, the gene for which has recently been cloned (5–7), are characterized clinically by the association of parathyroid, pancreatic, and pituitary tumors. The functional characterization of these NET may have multiple consequences; it may help to localize the primary tumor, better define the prognosis, improve the choice of treatment, serve as a marker during the follow-up, and finally, contribute to the understanding of tumor growth patterns. At present, no standard GEP NET biological screening procedure exists because studies measuring a large panel of hormones concomitantly in a large population of GEP NET patients of various origins are lacking. To provide guidelines for GEP NET biological screening, we measured a panel of 11 hormonal and nonhormonal
EUROENDOCRINE tumors (NET) are characterized by typical morphological features and expression of proteins such as the intracytoplasmic enzyme, neuron-specific enolase (NSE), and proteins found in the core of storage vesicles: chromogranin A and synaptophysin (1). This common definition encompasses a wide range of NET, exhibiting major differences in clinical and biological behavior. NET are characterized by their ability to produce multiple endocrine and nonendocrine markers. Considering gastroenteropancreatic (GEP) NET, the spectrum of hormonal secretions in tumors may reflect the normal secretion of the original cell: pancreatic peptides [e.g. insulin, glucagon, somatostatin (SMS), and pancreatic polypeptide] and serotonin are highly conserved markers of pancreatic and digestive tumors, respectively. In addition, a number of other hormonal secretions, generally called ectopic, may be associated with these tumors (2). Finally, serum measurement of gen-
Received June 4, 1998. Revision received September 23, 1998. Accepted October 20, 1998. Address all correspondence and requests for reprints to: Dr. Eric Baudin, Service de Me´decine Nucle´aire, Institut Gustave-Roussy, 39 rue Camille Desmoulins, 94805 Villejuif Cedex, France. E-mail:
[email protected]. * This work was supported in part by Programme Hospitalier de Recherche Clinique (PHRC) IDF 95015.
69
70
BAUDIN ET AL.
markers in a large single center population of patients with apparently sporadic GEP NET, to assess 1) the prevalence of MEN1 syndrome and 2) the prevalence of hormonal secretion and factors influencing their secretion. Furthermore, genetic testing was performed in patients in whom two components of the MEN1 syndrome were diagnosed. Subjects and Methods Patients All patients with GEP NET referred to our institution from November 1993 to October 1996 were enrolled consecutively. Their medical history and clinical data were recorded. Patients were considered as suspect of MEN1 when at least two components of the MEN1 syndrome were diagnosed. In these patients, blood or tumor tissues were collected to perform MEN1 gene analysis. The histological diagnosis was confirmed by a panel of pathologists (coordinated by J.-C.S.) and, when possible, the degree of differentiation was analyzed according to the classification of Warren and Gould (8). All tumors disclosed NET morphological features, including regular cells, normochromatic nucleus, and eosiniphilic cytoplasm, arranged in ribbons, nests, or sheets separated by a fine fibrovascular stroma. An immunohistochemical study with NSE, chromogranin A, and synaptophysin antibodies (Dako Corp., Glostsup, Denmark) was performed when the morphological structure precluded an unequivocal diagnosis of NET. NET were classified according to their primary site as foregut-derived (head and neck, respiratory tract, pancreas, stomach, and duodenum), midgut-derived (ileum, appendix, and right colon), and hindgut-derived (left colon and rectum) tumors. Patients with mixed tumors and small cell lung cancer were excluded. Patients were considered as having limited disease when only the primary site or lymph node metastases were known and as having extensive disease when distant metastases were known. This study was performed in accordance with local ethical rules.
Samples All blood samples were collected at the same time after overnight fasting without any dietary restrictions. When chemotherapy was prescribed, blood samples were collected before treatment. In patients treated with SMS analogs, samples were collected before the morning injection. Measurements were artificially classified as screening for MEN1 or hormonal production.
Screening for MEN1 Total serum calcium, phosphate, and creatinine levels were measured. Ionized calcium (specific electrode; normal, 1.12–1.30 mm/L) was measured when total serum calcium was found to be elevated after adjustment to albuminemia. Serum intact PTH [PTH-(1– 84); PTH intact, Diagnostic Products Corp., Los Angeles, CA; normal, 10 –70 pg/mL], PRL (ELSA-PROL, CIS-Bio International, Gif-sur-Yvette, France; normal, ,20 ng/mL), insulin-like growth factor I (IGF-I; SM-C-RIA-CT, Medgenix Diagnostics, Fleurus, Belgium; normal, 90 –270 ng/mL), and gastrin (GASK-PR, CIS-Bio International; normal, 25–100 ng/mL) were measured in the first 89 patients. Thereafter, only serum PTH and calcium measurements were performed. Blood and/or tumor DNA were extracted using the Qiagen kit (Qiagen, Chatsworth, CA). Published primer sequences flanking exons 2–10 of the MEN1 gene were used for PCR amplification (5, 6). All PCR products were then directly sequenced on a PE/ABI 377A automated sequencer using the ABI Prism Dye Terminator Cycle Sequencing Ready Reaction kit with AmpliTaq DNA polymerase (Perkin-Elmer, Norwalk, CT).
Screening for NET hormonal production It included measurements of NSE (CISPACK NSE, CIS-Bio International; normal, ,12.5 mg/L) and calcitonin (CT) levels (ELSA-hCT, CISBio International; normal, ,10 pg/mL). The glycoprotein hormone a-subunit (GPa; normal in men and premenopausal women: ,1 ng/mL; normal in postmenopausal women, ,3 ng/mL) and hCG b-subunit (free
JCE & M • 1999 Vol 84 • No 1
hCGb; normal , 0.1 ng/mL) were measured using specific immunoradiometric assays, as previously described (9). Twenty-four-hour urinary 5-hydroxyindolacetic acid excretion was measured by a high performance liquid chromatography method [5-hydroxyindolacetic acid (5-HIAA) reagent kit, Bio-Rad, Munich, Germany; normal, ,42 mmol/24 h]. NSE, 5-HIAA, CT, and GPa levels were measured in the 130 patients, and free hCGb was measured in the first 89 patients. SMS (SMS RIA kit, Incstar Corp.; normal, ,10 pg/mL) and 24-h urinary free cortisol (UFC; CORT-CT2, CIS-Bio International; normal, ,50 mg/24 h) were also measured in the first 64 patients. Serum PTH-related protein (PTHrp immunoradiometric assay kit, Incstar Corp., Stillwater, MN; normal, ,1.5 pm/L) was measured in two patients with increased ionized calcemia, low PTH-(1– 84) levels, and normal bone scintigraphy. Seven patients with elevated basal CT levels had a pentagastrin stimulation test (0.5 mg/kg peptavlon with CT measurements at 0, 2, and 5 min). Elevated hormonal levels, below twice the upper limit of the normal range, were confirmed by a second measurement. As hypothyroidism and renal insufficiency are known to increase the level of most hormones, all patients had serum TSH (Immulite Third generation TSH, Diagnostic Products Corp., Los Angeles, CA; normal, 0.4 – 4 mIU/mL) and serum creatinine measurements. No case of hypothyroidism was diagnosed during the study, and patients with renal insufficiency were excluded.
Statistical analysis The four main secretions individualized in this study (NSE, 5-HIAA, CT, and aGP), analyzed as elevated or normal, were successively compared to patient (age, sex), tumor (embryological origin, differentiation, disease extent), and treatment characteristics in the whole group of patients, then in foregut-derived NET patients, considering the main influence of embryological origin. Fisher’s exact test was used to compare proportions. When there were more than two means, means were compared using either Wilcoxon or Kruskal-Wallis nonparametric tests.
Results Clinical characteristics of the patients
The clinical characteristics of the 130 consecutive patients (67 males and 63 females; mean age 6 sd, 55 6 12 yr; range, 20 –76 yr) enrolled in this study are reported in Table 1. NET arose in the foregut, midgut, and hindgut in 74, 33, and 3 patients, respectively, and the primary site was unknown in 20 cases. Of the 38 patients with pancreatic NET, 9 had clinical symptoms related to hormonal hypersecretions, including gastrinomas (4), insulinomas (2), somatostatinomas (2), and vasoactive intestinal polypeptide-secreting carcinoma (1). Among the 108 tumors in which differentiation could be investigated, 78 were well differentiated, and 30 were poorly differentiated. At the time of the study, 116 patients had distant metastases, and 14 had a primary tumor without distant metastases, which was isolated in 2 and associated with lymph node involvement in the other 12. Ninety-one (70%) patients had already undergone 1 (39 patients) or multiple (52 patients) treatment modalities: surgery (64 patients), chemotherapy (64 patients), SMS analog therapy (18 patients), interferon (5 patients), and external radiotherapy (5 patients). The mean follow-up since diagnosis was 49 6 50 months (range, 3–306 months). These characteristics were not significantly different between patients with foregut or midgut tumors, except for histological differentiation; poorly differentiated NET were found in 40% and 7% of foregut and midgut tumors, respectively (P 5 0.001). Ninety-one patients were still alive at the end of the study. No familial history of MEN1-related NET was found in any of these patients.
HORMONAL SCREENING IN NEUROENDOCRINE TUMORS
71
TABLE 1. Clinical, histological, and treatment characteristics of the 130 NET patients Mean follow-up since diagnosis (months)
Histological differentiation: not known/well/poorly differentiated (n)
Mean age (yr)
Sex ratio female/male
Foregut Pancreas (38)
53.8
19/19
44.2
Others (36)
55.7
10/26
53.8
5/15/16
30
Midgut (33)
56
19/14
58
3/28/2
33
Hindgut (3) Unknown (20)
60 52
3/0 12/8
114.7 25.5
1/2/0 2/13/5
3 16
Tumor site (n)
MEN1 screening
Two patients (one lung and one pancreatic NET) had undergone surgery for a single parathyroid adenoma before the diagnosis of NET as reported in Table 2. In one of these two patients, who had lung NET, hyperparathyroidism recurred 4 yr later; two parathyroid adenomas and one hyperplastic gland were then found at surgery. Hypercalcemia was diagnosed in 7 other patients. In 2 of these patients, who had pancreatic NET, PTH-(1– 84) levels were in the normal range, which suggested hyperparathyroidism. Thus, 4 of the 130 patients (3%) had hyperparathyroidism, but only 1 patient had multiple parathyroid adenomas. In the other 5 patients with elevated serum calcium, PTH-(1– 84) levels were either low or undetectable. Three had known bone metastases; bone scintigraphy was normal in the other 2 patients, 1 of whom had elevated PTH-rp levels. Eighty-nine patients were screened for PRL, gastrin, and IGF-I. Hyperprolactinemia was found in six patients, was slightly elevated in five (,36 ng/mL), and was considered drug induced. In the other patient, who had an ileum NET, PRL attained 498 ng/mL, and magnetic resonance imaging suggested cystic prolactinoma. An elevated gastrin level was found in two patients (147 and 407 ng/mL, respectively). One of these two patients had a pancreatic tumor, whereas the other had a lung NET associated with a pancreatic mass. Gastric fibroscopy was normal as was the octreoscan in both subjects. In the first patient, with a previous history of chronic autoimmune thyroiditis, hypergastrinemia was attributed to chronic atrophic gastritis, and in the other patient, an associated gastrinoma was hypothesized because the gastrin level rose to 2800 ng/L before death. No clinical features indicative of acromegaly or elevated IGF-I levels were found. Therefore, two associated NET were diagnosed in six patients who were then considered as suspect for MEN1. In five of them (see Table 2), blood and/or tumor tissues were available for MEN1 genetic screening, which did not reveal any mutation in four patients. In one patient, with lung NET and involvement of multiple parathyroid glands, a nonsense (Arg108stop) mutation was detected in the tumor sample. Screening for NET hormonal production
The results of the biological markers used for screening for NET are presented in Table 3 and Fig. 1. Increased levels of
11/20/7
Metastatic patients (n)
34
Previous therapies (n)
Surgery (18), radiotherapy (1), chemotherapy (23), interferon (2), somatostatin analog (6) Surgery (14), radiotherapy (1), chemotherapy (17), interferon (1), somatostatin analog (3) Surgery (23), radiotherapy (1), chemotherapy (15), interferon (1), somatostatin analog (8) Surgery (2), chemotherapy (1) Surgery (7), radiotherapy (2), chemotherapy (8), interferon (1), somatostatin analog (1)
NSE were found in 61 of the 130 patients (47%); in 32 of 74 (43%), 14 of 33 (42%), 2 of 3 (66%), and 13 of 20 (65%) cases of foregut, midgut, hindgut, and NET of an unknown primary site, respectively. NSE levels were above 20 mg/L in 42 patients and above 100 mg/L only in patients with foregutderived NET. However, mean NSE levels in foregut and midgut NET did not differ. Elevated urinary excretion of 5-HIAA was found in 60 of the 130 patients (46%) and exceeded 100 mmol/24 h in 50 patients. Elevated urinary excretion of 5-HIAA was found in 20 of 74 (27%), 28 of 33 (85%), 1 of 3, and 11 of 20 (55%) cases of foregut, midgut, hindgut NET, and NET of unknown primary site, respectively. Mean 5-HIAA values found in foregut and midgut NET were not different. The CT level was elevated in 18 of the 130 patients (14%) and was above 20 pg/mL in 13. Fifteen of them had a foregutderived NET, including 3 head and neck, 7 lung and 5 pancreatic tumors, in only 1 ileal tumor, and 2 NET of unknown primary site. Thyroid ultrasonography performed in all patients demonstrated thyroid nodules more than 1 cm in diameter in 4 of them and enlarged neck lymph nodes in 2. A pentagastrin test, performed in 7 patients, showed no significant response in 5, including the 2 patients with enlarged neck lymph nodes, and found a 2-fold increase in the CT level in 2 patients with lung NET whose thyroid gland was normal at ultrasonography. GPa levels were elevated in 25 of the 130 patients (19%) and were above 3 ng/mL in 23. Elevated GPa levels were found in 21 patients with foregut-derived primary NET, including 1 mediastinal, 10 lung and 10 pancreatic tumors, in only 1 patient with a rectal NET, and in 3 patients with a primary NET of an unknown origin. None of the patients with an elevated GPa level had pituitary disease, as shown by biological screening and, in 11 of them, by CT scan. Free hCGb levels were elevated in 11 of 89 patients (12%). Elevated free hCGb levels were found in 10 patients with foregut-derived NET, including 1 mediastinal, 3 lung, and 6 pancreatic NET and in 1 patient with NET of an unknown primary site. Five patients had elevated levels of both free hCGb and GPa (4 pancreatic and 1 NET of unknown primary site), but none had elevated hCG levels. Elevated SMS levels were found in 2 of 64 patients (3%) without a known pancreatic tumor. Both patients had lung
Not done 1995: gastrin, 407 ng/L; 1996: gastrin, 2800 ng/mL
Ileum
Lung
60/M
66/F
No
No mutation (lymphocytes) 1996: PRL, 498 ng/mL
Pancreatic gastrinoma 66/M
No
No mutation (lymphocytes) 1996: Cai, 1.34 mmol/L; PTH, 30 ng/L
Pancreas nonfunctioning 56/F
No
No mutation (tumor) 1994: Cai, 1.48 mmol/L; PTH, 35 ng/L
Pancreas nonfunctioning 54/F
No
1997: Cai, 1.1 mmol/L; PTH, 51 ng/L
No mutation (tumor 1 lymphocytes)
Arg108stop mutation, exon 2 (tumor)
2 parathyroid adenomas, 1 hyperplasic gland, 1 necrotic gland, no other MEN1 lesion Single adenoma, 2 other normal parathyroid glands seen, no other MEN1 lesion Hyperparathyroidism, nonsurgical management, no other MEN1 lesion Hyperparathyroidism, nonsurgical management, no other MEN1 lesion Prolactinoma, external radiotherapy, no other MEN1 lesion Associated pancreatic tumor: gastrinoma, no other MEN1 lesion 1994: Ca total, 2.21 mmol/L; PTH, 13 ng/L
Parathyroid surgery 1985, relapse 1989, parathyroid surgery 1992 Parathyroid surgery 1992 Lung 63/F
Presenting lesion Age (yr)/sex
TABLE 2. Screening for MEN1 syndrome
MEN history
MEN biological screening or last measurements
MEN1 gene screening
BAUDIN ET AL.
Diagnosis
72
JCE & M • 1999 Vol 84 • No 1
tumors, and SMS levels were above 50 pg/mL. Four other patients, 1 lung, 2 pancreatic, and 1 mediastinal NET, respectively, had a slight increase in the SMS level (,36 pg/ mL). Three of these patients had diabetes, and 1 had hypoglycemia related to an insulinoma. These 4 patients were not considered as having tumor-induced SMS secretion. Elevated 24-h excretion of UFC was found in 4 of 64 patients (6%; range, 58 –337 mg/24 h), including 1 patient with clinical Cushing’s syndrome. None of these patients received corticosteroid treatment. These 4 patients had foregutderived NET, including 1 head and neck, 1 lung, and 2 pancreatic tumors. In 2 of these patients with basal 24-h excretion of UFC below 200 mg/24 h, a standard dexamethasone suppression test (2 mg/day during 2 days) showed incomplete suppression of UFC, and a CRH test was abnormal in only 1 patient, but immunohistochemistry with ACTH antibodies was positive on tumor tissue in the other patient. Multiple hormonal secretions and relationships with clinical and tumoral characteristics
Among the 130 patients screened, at least 1 of these markers was elevated in 108 patients (83%), only 1 was elevated in 58 patients (45%), and at least 2 were elevated in 50 patients (38%). In 22 (17%) patients all markers were normal. In the subgroup of 64 patients in whom all markers were measured, isolated or combined secretions were, respectively, found in 30 (47%) and 25 (39%) patients, and all measurements were normal in only 9 subjects (14%). At the end of hormonal screening, 20 of 74 (27%) and 2 of 33 (6%) of foregut and midgut NET were negative for all markers screened. No significant statistical relationship was found between NSE and the clinical parameters analyzed. 5-HIAA secretion was correlated with the embryological origin (P , 0.0001) and histological differentiation (P 5 0.01); 5-HIAA secretion was found in 1) 85% of midgut-derived NET compared to 27% of foregut-derived NET, and 2) 56% of well differentiated compared to 30% of poorly differentiated NET. However, when foregut-derived NET were considered alone, 5-HIAA secretion was found in a similar proportion of patients with either well or poorly differentiated NET. CT secretion was correlated with sex (P 5 0.0006) and embryological origin (P 5 0.04). CT secretion was found in: 1) 23% of male compared to 3% of female patients, and 2) all CT secretions but one arose in foregut-derived NET. GPa secretion was correlated with sex (P 5 0.008) and embryological origin (P 5 0.0004). GPa secretion was found in 1) 28% of male compared to 9% of female patients, and 2) all GPa secretions but one arose in foregut-derived NET. Furthermore, all patients with GPa secretions had metastatic disease. When taking into account only foregut-derived NET, GPa secretion and disease extent were significantly correlated (P 5 0.05). A trend toward a positive statistical significance between a previous therapy, including SMS analog therapy, and positive hormonal secretions, particularly 5-HIAA, was found. Indeed, patients with hormonal secretions may be diagnosed and hence treated earlier.
HORMONAL SCREENING IN NEUROENDOCRINE TUMORS
73
TABLE 3. Frequency of hormonal secretions according to NET primary site
NSE 5-HIAA GPa CT hCGb UFC SMS PTH-rp
Foregut (mean/median/range)
Pancreas (mean/median/range)
Midgut (mean/median/range)
15/36(42%) (35.8/26/13–99) 16/36(44.4%) (582/268/64 –1871) 11/36(30.5%) (153/23/2.9 –757) 10/36(28%) (1521/145/11–7600) 4/24(17%) (0.57/0.45/0.21–1.2) 2/23(9%) (58 –78) 2/23(9%) (52–92) 1/35(3%) (2.1)
17/38(45%) (68/40/20 –166) 4/38(10%) (299/104/84 –904) 10/38(26%) (78/12/1.3–527) 5/38(13%) (5125/151/11–24400) 6/29(21%) (1.2/1.2/0.13– 4.3) 2/20(10%) (297–337) 0/20
14/33(42%) (31.4/20/14 – 81) 28/33(85%) (809.2/427/53– 4068) 0/33
0/38
Hindgut
Total (mean/median/range)
61/130(47%) (44/29/13–166). 60/130(46%) (615/347/53– 4068) 25/130(19%) (105/20/1.3–757) 18/130(14%) (2504/166/11–24400) 11/89(12%) (1.08/0.5/0.13– 4.3). 4/64(6%) (192/187/58 –337) 2/64(3%) (52–92) 1/130(,1%) (2.1)
0/13
0/1
13/20(65%) (38/23/15– 87) 11/20(55%) (329/196/72– 830) 3/20(15%) (46/35/5–100) 2/20(10%) (25–1526) 1/10(10%) (2.2) 0/7
0/13
0/1
0/7
0/33
0/3
0/20
1/33(3%) (182) 0/24
2/3 (16 –50) 1/3 (111) 1/3 (13) 0/3
Unknown (mean/median/range)
0/2
Mean, median, and range are given when at least three abnormal values were observed; individual values are given elsewhere.
FIG. 1. Percentage of abnormal hormonal secretions according to the embryological origin of the NET primary site. Foregut derived tumors include pancreatic NET.
Discussion
Some preliminary remarks are mandatory concerning the population studied. 1) Most patients had advanced disease, and recommendations could not be directly applied to patients with limited disease; 2) their lesions did not cover the entire spectrum of GEP-NET: few patients had thymic NET or clinically functional pancreatic NET in whom the frequency of MEN1 ranges up to 25%. 3) Most patients had received previous treatments, but no significant difference was observed in the 4 main secretions analyzed regarding previous therapy. 4) Both well and poorly differentiated NET were studied. 5) Finally, the results of the 4 main secretions were analyzed rather than the secretory activity of NET, because even after measurements of 11 biological markers this definition may be equivocal. Apart from patients with gastrinoma, prospective data on the incidence of MEN1 in patients presenting with apparently sporadic GEP NET are lacking. Previous retrospective studies suggest a 4 –7% prevalence of GEP NET, mainly of foregut origin (10 –12). After the biological
screening, six patients were found to have associated NET compatible with a MEN1 syndrome, including three patients with atypical NET association, as two cardinal tumors of the syndrome were not present. The responsible MEN1 gene has recently been isolated, and its germline mutations, scattered along the gene sequence, have been identified in most familial MEN1 (5–7). Direct sequencing of the MEN1 gene was performed in five of six patients with two components of MEN1 syndrome and did not show any mutation in four patients. Only one tumor nonsense mutation, at codon 108, was found in a patient with a history of multiple parathyroid tumors, highly suggestive of a MEN1 syndrome. As no familial NET history was known in this patient, it may correspond to a sporadic MEN1 syndrome. This observation suggests a low prevalence of MEN1 syndrome in patients presenting with apparently sporadic GEP NET (,1%). Serum calcium and PTH-(1– 84) measurements appear to be sufficient for MEN1 screening, and they should be performed especially in patients with foregut NET. Even if the genetic screening
74
JCE & M • 1999 Vol 84 • No 1
BAUDIN ET AL.
was not performed in the entire population, the high penetrance of MEN1 by age 50 yr (12), the characteristics of our population, and the results of the biological screening suggest that the prevalence of MEN1 in the other patients is extremely low. Five markers were increased in more than 10% of patients. NSE and 5-HIAA levels were elevated in almost half the patients, and CT, GPa, and free hCGb were elevated in 12–19% of the patients. Only 17% of the patients had no increase in any marker level, and a single hormone secretion was found in the majority of patients. These results were only slightly modified when all the markers studied were taken into account, because many tumors gave rise to multiple secretions. The biological behavior of NET was highly dependent on the primary tumor site. All peptidic and glycoproteic hormone secretions originated in foregut-derived NET, except in two patients. 5-HIAA secretion was more frequently observed in midgut-derived NET. These biological results confirmed the relevance of using the classification based on embryological origin, as first suggested by Williams and Sandler (13), to define the biological behavior of NET. These results are consistent with immunohistochemical findings that revealed that normal foregut neuroendocrine cells and NET often exhibit multihormonal production with many peptide-producing cells (14). In contrast, normal midgut neuroendocrine cells and NET contain a limited variety of endocrine cells that mainly produce serotonin and tachykinins (15). Despite the diversity of secretion typical of foregutderived NET, a greater percentage (27%) of these patients, compared to midgut-derived NET patients (6%), remained negative for hormonal screening. Our results may help both to define which hormones should be screened when the primary is known and to search for the primary site when a metastasis is the first known event. NSE, 5-HIAA, CT, and GPa ought to be routinely measured in foregut-derived NET patients, and only NSE and 5-HIAA should be measured in those with midgut-derived NET. Interestingly, no relationship was found between histological differentiation and the four main secretions studied. Nevertheless, an impact of histological differentiation on other hormonal secretions should not be ruled out. The practical interest of serum NSE measurement has been shown in patients with small cell lung cancer (16), but remains debatable in patients with other NET. According to previous results (3, 17), NSE is a poorly specific and sensitive marker of NET, as it is elevated in only 47% of the patients. Chromogranin A has been put forward as a more sensitive and specific general NET marker (3). The sensitivity of these general markers is higher than that of other secretions in foregut-derived NET. Biogenic amines are well known secretions characterizing NET, and 5-HIAA is the most sensitive marker of the carcinoid syndrome (18). Peptides such as neurotensin, neuropeptide K, neurokinin A, and substance P are produced by GEP NET mainly of midgut origin (18 –20), but both their sensitivity and specificity are debatable, and assays are not routinely available. Elevated 5-HIAA was found in 46% of our overall population, and was predominantly produced by midgut-derived tumors in the same range as that previously reported (18 –21). CT, GPa, and free hCGb secretions have previously been
demonstrated in GEP NET (3, 21–26). In our study, these silent markers were found to be secreted in 12–19% of patients. We previously reported that CT secretion is a specific marker not only of MTC, but also of other NET (26). A 14% incidence of CT secretion was found and was associated with foregut-derived NET in all cases but one. The increase in pentagastrin-stimulated CT, used to differentiate GEP NET from medullary thyroid carcinoma, did not exceed 2-fold the basal level. Isolated secretion of GPa should also be regarded as specific of NET. Indeed, less than 3% of patients with non-NET malignant tumors exhibited GPa secretion exceeding 3 ng/mL (24). In the present study, GPa secretion was found in 19% of NET patients with levels exceeding 3 ng/mL in 92% of these cases, which is definitively higher than those observed in pituitary adenomas and other tumors (24, 27). Furthermore, GPa secretions were exclusively observed in foregut-derived NET in contrast to previous studies reporting GPa secretion in midgut-derived NET (21, 25). Our results, however, are consistent with a previous immunochemistry study that detected GPa in 21–55% of foregut-derived NET, but not in any ileal NET (28). All patients with GPa secretions had metastatic disease. The interest of GPa as a prognostic marker of malignancy should be evaluated in prospective studies. GPa secretion was associated with free hCGb in 20% of cases. Interestingly, none of these patients produced dimeric hCG. In contrast with GPa secretion, an increase in free hCGb levels has been reported to be associated with various neoplastic diseases and hence should not be considered as specific for NET (24). Recent results suggest that free hCGb may be a marker of aggressive neoplastic disease (29). Finally, increases in UFC, SMS, or PTH-related peptide levels were rare events, and no paraneoplastic acromegaly was observed. Secretions of ACTH, GHRH, and PTH-related peptide have been associated with various histological types of tumor (30). Screening for these markers should depend on clinical presentation. Our study shows that MEN1 is a rare event in patients with apparently sporadic NET. Ionized calcium and PTH-(1– 84) measurements are sufficient diagnostic tools when searching for MEN1 and should be routinely performed in patients with foregut NET. Routine hormonal screening is dependent on the primary site. We suggest that NSE, 5-HIAA, CT, and aGP should be screened for in foregut-derived NET, and NSE and 5-HIAA should be screened for in patients with midgut-derived NET. The prognostic value of these secretions as well as their interest in the follow-up of NET await assessment in future studies. Acknowledgments We are indebted to Ingrid Kuchenthal and Catherine Martin for secretarial assistance, to Christine Machavoine and the nurses of the Nuclear Medicine Department for technical assistance, and to Lorna Saint-Ange for editing. We also thank the Groupe d’Etude des Ne´oplasies Endocriniennes Multiples for useful discussions during the preparation of the manuscript.
References 1. Bishop AE, Polak JM. 1993 Modern morphological and other investigative methods. In: Polak JM, ed. Diagnostic histopathology of neuroendocrine tumours. London: Churchill Livingston; 1–14.
HORMONAL SCREENING IN NEUROENDOCRINE TUMORS 2. DeLellis RA, Tischler AS, Wolfe HJ. 1984 Multidirectional differentiation in neuroendocrine neoplasms. J Histochem Cytochem. 32:899 –904. 3. Nobels FR, Kwekkeboom DJ, Coopmans W, et al. 1997 Chromogranin A as serum marker for neuroendocrine neoplasia: comparison with neuron-specific enolase and the a-subunit of glycoprotein hormones. J Clin Endocrinol Metab. 82:2622–2628. 4. Dayal Y, Wolfe HJ. 1984 Regulatory substances in clinically nonfunctioning gastrointestinal carcinoids: evolution and tumor pathology of the neuroendocrine system. In: Falkmer S, Hakanson R, Sundler F, eds. Evolution and tumor pathology of the neuroendocrine system. Amsterdam: Elsevier; 497–517. 5. Chandrasekharappa SC, Guru SC, Manickam P, et al. 1997 Positional cloning of the gene for multiple endocrine neoplasia-type 1. Science. 276:404 – 407. 6. Agarwal SK, Kester MB, Debelenko LV, et al. 1997 Germline mutations of the MEN1 gene in familial multiple endocrine neoplasia type 1 and related states. Hum Mol Genet. 6:1169 –1175. 7. The European Consortium on MEN1. 1997 Identification of the multiple endocrine neoplasia type 1 (MEN1) gene. Hum Mol Genet. 6:1177–1183. 8. Warren WH, Gould VE, Faber LP, Kittle CF, Memoli VA. 1985 Neuroendocrine neoplasms of the bronchopulmonary tract. J Thorac Cardiovasc Surg. 89:819 – 825. 9. Ozturk M, Bellet D, Manil L, Hennen G, Frydman R, Wands J. 1987 Physiological studies of human chorionic gonadotropin (hCG), hCGa, and hCGb as measured by specific monoclonal immunoradiometric assay. Endocrinology. 120:549 –558. 10. Duh QY, Hybarger CP, Geist R, et al. 1987 Carcinoids associated with multiple endocrine neoplasia syndromes. Am J Surg. 154:142–148. 11. Calender A, Giraud S, Cougard P, et al. 1995 Multiple endocrine neoplasia type 1 in France: clinical and genetic studies. J Intern Med. 238:263–268. 12. Trump D, Farren B, Wooding C, et al. 1996 Clinical studies of multiple neoplasia type 1 (MEN1). Q J Med. 89:653– 669. 13. Williams ED, Sandler M. 1963 The classification of carcinoid tumours. Lancet. 1:238 –239. 14. Sheppard MN. 1993 Neuroendocrine tumours of the lung. In: Polak JM, ed. Diagnostic histopathology of neuroendocrine tumours. London: Churchill Livingston; 151–165. 15. Solcia E, Rindi G, Sessa F, et al. 1993 Endocrine tumours of the gastrointestinal tract. In: Polak JM, ed. Diagnostic histopathology of neuroendocrine tumours. London: Churchill Livingston; 123–149. 16. Johnson PW, Joel SP, Love S, et al. 1993 Tumour markers for prediction of survival and monitoring of remission in small cell lung cancer. Br J Cancer. 67:760 –766.
75
17. Cunningham RT, Johnston CF, Irvine GB, Buchanan KD. 1992 Serum neurone-specific enolase levels in patients with neuroendocrine and carcinoid tumours. Clin Chim Acta. 212:123–131. 18. Feldman J, O’Dorisio TM. 1986 Role of neuropeptides and serotonin in the diagnosis of carcinoid tumors. Am J Med. 81:41– 46. 19. Feldman J. 1986 Urinary serotonin in the diagnosis of carcinoid tumors. Clin Chem. 32:840 – 844. 20. Norheim I, Theodorsson-Norheim E, Brodin E, Oberg K. 1986 Tachykinins in carcinoid tumors: their use as a tumor marker and possible roˆle in the carcinoid flush. J Clin Endocrinol Metab. 63:605– 612. 21. Norheim I, Oberg K, Theodorsson-Norheim E, et al. 1987 Malignant carcinoid tumors. An analysis of 103 patients with regard to tumor localisation, hormone production, and survival. Ann Surg. 206:115–125. ¨ berg K, Skogseid B. 1989 Neuroendocrine pancreatic tumors. 22. Eriksson B, O Clinical findings in a prospective study of 84 patients. Acta Oncol. 28:373–377. 23. Prinz RA, Bermes EW Jr, Kimmel JR, Marangos PJ. 1983 Serum markers for pancreatic islet cell and intestinal carcinoid tumors: a comparison of neuronspecific enolase b-human chorionic gonadotropin and pancratic polypeptide. Surgery. 94:1019 –1023. 24. Marcillac I, Troalen F, Bidart JM, et al. 1992 Free human chorionic gonadotropin b subunit in gonadal and nongonadal neoplasms. Cancer Res. 52:3901–3907. 25. Grossmann M, Trautmann ME, Poertl S, et al. 1994 a-Subunit and human chorionic gonadotropin-b immunoreactivity in patients with malignant endocrine gastroenteropancreatic tumours. Eur J Clin Invest. 24:131–136. 26. Ghillani PP, Motte P, Troalen F, et al. 1989 Identification and measurements of calcitonin precursors in serum of patients with malignant diseases. Cancer Res. 49:6845– 6851. 27. Oppenheim DS, Kana AR, Sangha JS, Klibanski A. 1990 Prevalence of a-subunit hyperecretion in patients with pituitary tumors: clinically nonfunctionning and somatotroph adenomas. J Clin Endocrinol Metab. 70:859 – 864. 28. Heitz PU, von Herbay G, Kloppel G, et al. 1987 The expression of subunits of human chorionic gonadotropin (hCG) by nontrophoblastic, nonendocrine, and endocrine tumors. Am J Clin Pathol. 88:467– 472. 29. Lazar V, Diez SG, Laurent A, et al. 1995 Expression of human chorionic gonadotropin b subunit genes in superficial and invasive bladder carcinomas. Cancer Res. 55:3735–3738. 30. Odell WD. 1997 Endocrine/metabolic syndromes of cancer. Semin Oncol. 24:299 –317.
