World j. Surg. 16, 611-619, 1992

World Journal of Surgery O 1992 by the Soci~t6 lnternationale de ChlrursJe

Pancreatic Tumors in Multiple Endocrine Neoplasia Type 1: Clinical Presentation and Surgical Treatment I). Grama, M.D., B. Skogseid, M.D., E. Wilander, M.D., B. Eriksson, M.D., H. M~trtensson, M.D., B. Cedermark, M.D., B. Ahr6n, M.D., A. Kristofferson, M.D., K. Oberg, M.D., J. Rastad, M.D., and 13./~kerstrfm, M.D. Departments of Surgery, Medicine, and Pathology, University Hospital~ Uppsala, Department of Surgery at Helsingborg Hospital, Karolinska Hospital, Stockholm, and University Hospitals, Lund, and Ume~, Sweden Among 33 patients with endocrine pancreatic tumors due to multiple endocrine neoplasia type 1 (MEN-I), 19 (58%) patients had bypergastrinenlia, 7 (21%) patients had hyperinsulinism, and 7 (21%) patients had clinically non-functioning lesions. At least one gross tumor was found in ~oI patients undergoing pancreatic surgery, including those with negative °calization studies prior to operation. The patients also had additional nlacroscopic tumors as well as numerous microadenomas, and the lesions frequently were positive for immunostaining with multiple hormones, ainly pancreatic polypeptide, insulin, glueagon, and somatostatin. aodenal endocrine lesions were found in 4 of 5 investigated patients and stained with gastrin and somatostatin antibodies. Distal, mainly subtotal Pancreatic resection, was performed in 18 patients, eventually combined With caput tumor enucleation or duodenotomy, while a few patients Underwent only tumor enucleation or a Whipple procedure. The longterra outcome of operation was most favorable in patients with hyperinSulinism; only 1 patient had clinical recurrence. Patients with hypergastrirtemia experienced only transitory lowering of serum gastrin values l~ r Pancreatic surgery and 47% of them had or developed metastases. eh tumor spread was seen in 57% of the patients with non-functioning lans. Nine patients died from progressive tumor disease during followP" Consistent with previous studies, we found that surgery is indicated in ~ N - 1 patients with hyperinsulinism even if a lesion is not visualized by i'_~.l°logy. In addition, these indications should be extended to also aelUde patients with only biochemical markers of disease, including elevations of eastrin, as these indicate the presence of ross tumors This strate ~ g " It- g3' should be applied especially in patients with aggressive family tstories to possibly reduce the risk of malignant tumor progression.

~

~

Endocrine pancreatic tumors have been demonstrated by autopsy in 82% of individuals with multiple endocrine neoplasia type 1 (MEN-l) [1]. The pancreatic lesion has frequently been tnOre difficult to diagnose clinically and therefore often considrcedas a late manifestation of MEN-I [2]. More recent studies, °Wever, indicate that early recognition of pancreatic involve~r?~t~emaybe achieved by sensitive biochemical screening [3, 4]. patient with MEN-I, the pancreas commonly has multicentric lesions, particularly numerous microadenomas interStoPr~ented at the International Association of Endocrine Surgeons in R olin, Sweden, August, 199!. 12,:eprint requests: G6ran/~kerstr6m, M.D., Department of Surgery, "versity Hospital, S-751 85 Uppsala, Sweden.

mingled with a few tumors of larger size, all of which characteristically exhibit signs of multiple hormone production [5-7]. Pancreatic surgery would be successful mainly in patients with resectable, gross tumors, provided metastases have not yet occurred. The role of surgery has been debated, however, due to the histologic heterogeneity and multiplicity of the lesions, and due to high rates of recurrence, particularly in patients with gastrin excess [8, 9]. The incidence of malignancy generally seems to be lower than in the corresponding pancreatic tumors of non-familial origin, although pancreatic neoplasms constitute a principal cause of death in the MEN-1 syndrome [5, 10-12]. Recent findings indicate that the malignant behavior of these tumors may vary considerably and that certain kindreds may bear tumors with more malignant potential [13]. The present study evaluated retrospectively the clinical presentation, histopathology, treatment, and long-term follow-up in MEN-1 patients with pancreatic islet cell lesions.

Material and Methods

This study included 33 patients with MEN-1 with associated endocrine pancreatic tumors, 25 of whom were treated from 1973 to 1986 at the University Hospitals of Uppsala, Stockholm, Lund, Malm6 and Ume/i and 8 patients more recently in Uppsala who were included to enable more detailed histological analysis. One patient who initially underwent insulinoma enucleation in 1968 but underwent re-operation during the study period was included. Altogether 16 of the patients were males and 17 patients were females, with age at time of diagnosis or operation of 13-65 years (mean, 44 years). Nineteen (58%) patients had hypergastrinemia with ZollingerEllison's syndrome (ZE), 7 (21%) patients had hypoglycemic symptoms due to insulin or proinsulin excess, while 7 (21%) patients were judged to have clinically non-functioning tumors. The majority of the patients (88%) were diagnosed due to clinical symptoms associated with the pancreatic tumors. Nine patients were detected by screening of MEN-1 families or

612

World J. Surg. Vol. 16, No. 4, July/Aug. 1992

Table 1. Radiological localization of a gross endocrine pancreatic tumor in the MEN-I patients investigated prior to pancreatic surgery. Diagnosis Insulinoma True positive (n) Sensitivity (%) Gastrinoma True positive (n) Sensitivity (%) Non-functioning lesions True positive (n) Sensitivity (%)

Selective angiography

CT

Angiography + CT

5 83

2 66

6 86

7 70

8 73

10 83

3 60

3 60

6 86

Percutaneous ultrasound

pTP~

0

"PTP was considered truly positive upon correct lateralization to the pancreatic head or corpus/tail region. CT: Computed tomography; PTP: Percutaneous portal vein sampling.

patients with parathyroid hyperplasia and only 4 of them appeared to be asymptomatic. The MEN-1 patients constituted 22% of all individuals with endocrine pancreatic tumors undergoing diagnosis or treatment during the study period [14]. The patients were considered to have malignant neoplasms if metastases were present at time of diagnosis or occurred during the follow-up of 1-16 years (mean, 8 years). Data were collected from the patient's records concerning symptoms, hormone values, and efficacy of pre-operative localization procedures. The number and size of the excised pancreatic tumors as well as the result of immunohistochemical stainings were obtained from the surgical records and pathology reports. Detailed histopathological re-examination of the pancreatic (n = 14) and duodenal specimens (n = 4) as well as available metastases was undertaken in 14 patients. The pancreas had been subjected to subtotal (70-80%) resection in 8 of these patients and the entire specimens were sliced at 2--3 mm intervals with careful search for macrotumors and microtumors by palpation and microscopy. This material was fixed in 10% neutral formalin, paraffin-embedded, and serially sectioned at 4 /xm. The sections were stained with hematoxylin-eosin, alkaline Congo Red [15], argentaffin [16], and argyrophil stains [17], and immunohistochemically with antisera against chromogranin A and B (diluted 1 : 1000) [18], insulin (1:2000; Biogenex Lab, San Ramon, California, U.S.A.), glucagon (1:500; Novo Biolabs, Danbury, Connecticut, U.S.A.), somatostatin (1:500; Dako Corp, Carpenteria, California, U.S.A.), pancreatic polypeptide (t:1000; Dako), gastrin (I :5000; Milab, Malmr, Sweden), vasoactive intestinal peptide (VIP) (1:400; Milab), and calcitonin (1 : 1000; Dako). The biotin-avidin peroxidase (ABC) technique (Vector Lab, Burlingame, California, U.S.A.) with primary antisera diluted in TRIS-saline buffer (pH 7.4) before application, and 3,3 diaminobenzidine tetrahydrochloride (0.06%) were used to visualize the immunoreactions. Normal human pancreas, intestine, and thyroid gland specimens constituted positive controls, while negative controls were obtained by omission of the primary antisera and by neutralization tests. Results

Symptoms and Diagnosis Twelve (63%) of the patients with hypergastrinemia displayed histories of peptic ulcer disease, which was recurrent and/or

complicated in the majority. Two patients had dyspepsia without verified ulcer, 4 patients had dyspepsia together with diarrhea, and only 1 patient had diarrhea alone. Five (71%) of the patients with excessive insulin/proinsulin production had typical symptoms of neuroglycopenia and catecholamine excess (one of them when lactating) and 2 patients were obese. General symptoms of fatigue, weight loss, or abdominal pair1 constituted the predominant symptomatology among the patients with non-functioning tumors, one of whom had a palpable abdominal lesion. The general symptoms were also noted at diagnosis of occasional patients with gastrinoma or insulinoma. Two patients with insulin excess and another 2 patients with non-functioning tumors detected by screening experienced no obvious symptoms. Fourteen (74%) of the patients with gastrinoma had elevated basal serum gastrin levels and the same number of patients had BAn/MAn quotients exceeding 0.6. Ten patients had a pathO" logical secretin test. All the insulinoma diagnoses relied on the demonstration of elevated serum insulin/proinsulin values and/or pathological insulin/glucose quotients after fasting. TWO patients also had increased C-peptide levels. All patients under" going more extensive hormonal screening displayed elevated levels of uther'hormones, beside those associated with the classical clinical syndrome. Thus, 5 of the patients with gastri" noma had elevated serum pancreatic polypeptide values, 4 patients exhibited insulin and 1 patient proinsulin excess, while 3 patients had elevated somatostatin and glucagon concentra" tions. The patients, with insulinoma or non-functioning lesions most frequently displayed elevations of serum pancreatic polY~ peptide, but many of them did not undergo careful hormonal screening.

Radiological Localization Pre-operative radiology studies imaged only one, i.e., the largest, tumor in each patient. Angiography was the superior method for visualization of such tumors in the patients witla insulinoma (sensitivity 83%) and computed tomography (CT) localized tumors in 66% of these patients including 1 tumor missed by angiography (Table 1). One patient with insuli n°~a was operated upon after successful tumor localization onlY bY percutaneous transhepatic portal vein sampling (PTP). For the gastrinomas as well as the non-functioning tumors, angiograph~ and CT were equally efficient and the combination was superioJ

I). Grama et al.: Pancreatic Tumors in MEN-I

613

to either method alone. Both measures demonstrated liver metastases in 5 patients. Percutaneous ultrasound showed lower efficacy than both angiography and CT. Regional lymph gland metastases were demonstrated by CT prior to laparatomy in I of the 6 patients exhibiting such dissemination at operation. PTP was performed in 12 patients and provided correct regionalization in 3 of them, but in only 1 patient was a previously Unrecognized tumor visualized. Two patients underwent emergency surgery because of gastrointestinal bleeding and they did not undergo radiological studies prior to operation.

Surgel~y Twenty-seven of the patients underwent pancreatic surgery, COnsisting of pancreatic tail resection (8 patients) combined With caput tumor enucleation (2 patients), subtotal (70-80%) resection (10 patients), a Whipple procedure (3 patients), or tumor enucleation alone (6 patients). Resection of liver metastases was performed in 2 patients and lymph gland metastases Were removed in 3 patients. Duodenotomy was performed in 5 Fig. 1. Transected pancreatic specimen from a patient with MEN-I Patients, total gastrectomy in 3 patients, and Billroth II resec- showing 2 larger and 3 smaller (arrows) tumors. tion or vagotomy in one patient each. Two patients had adrenal tumors removed during the pancreatic surgery. Four patients Were subjected to one or two (1 patient) re-operations for of those with hypergastrinemia, and 3 (43%) of the patients with non-functioning lesions. Six of the patients with hypergastrinreCUrrent pancreatic tumor. All patients with insulinoma underwent pancreatic surgery. emia had metastases to regional lymph glands (n = 3) or the In agreement with the pre-operative radiological studies, one liver (n = 3), and 1 patient exhibited such spread in both large tumor of 1--4 cm (mean, 2.5 cm) in diameter was found in locations. Two of the patients with non-functioning lesions had each patient. Three patients had 1 to 4 additional, palpable regional lymph gland and liver metastases, while I patient had tumors with diameters of 0.5-1 cm. Three of the patients with liver metastases only. The largest tumor in the patients with gastrinoma did not undergo surgery, 2 patients because of metastases (0.7-20 cm) did not differ in size from those lacking disseminated liver metastases and 1 patient because no tumor pre-operative or peroperative signs of such dissemination. There was no immediate (within 30 days) postoperative Was visualized by pre-operative radiology. Another patient with liver metastases underwent only open biopsy and yet another, mortality. Pancreatic fistulation/abscesses in 3 patients healed Without metastases, underwent gastric surgery without explo- uneventfully after percutaneous drainage. One patient with the ration of the pancreas. The remaining 14 patients with gastri- ZE syndrome experienced duodeno-jejunal perforation 2 weeks noma had at least one palpable pancreatic tumor of 0.7-8 cm after pancreatic surgery and recovered uneventfully after intes(mean, 2.0 cm) in diameter. Six of these patients had 1 to 6 tinal resection. Another patient subjected to pancreatico-jejunal additional pancreatic tumors, which measured 0.5-1.5 cm in anastomosis after enucleation of a 6 cm caput tumor, together diameter. One to 3 duodenal lesions of 0.5-2 cm were found in with subtotal gastrectomy because of advanced foregut carci4 Patients with gastrin excess, and 1 of these patients also had noid, underwent 2 re-operations for leaking anastomosis. nodules 1-3 mm in size covering virtually the entire inside of the stqaerior and descending portions of the duodenum. The duo- Histopathology denOtomy, however, was negative in 1 patient. All patients with non-functioning lesions exhibited at least 1 Some gross tumors identified by palpation were greyish, firm, Pancreatic tumor measuring 1-20 cm (mean, 6.7 cm) in diame- fibrotic (Fig. 1) and occasionally even calcalous, while others ter. Four of them also had 1-3 additional gross tumors < 1 cm in were reddish-brown and soft. Most of these lesions appeared S~ze. One patient with a 20 cm pancreatic tail tumor, without macroscopically well demarcated by a capsule from the sur~aCroscopic signs of metastases, underwent only open biopsy. rounding pancreas. When the specimens of subtotal pancreatic ApProximately 80%(n = 49) of the palpable gross tumors resections were sliced serially, up to 12 additional tumors of !-3 Were found in the pancreatic corpus or tail and the remaining mm were identified in all the investigated patients. These tumors (n = 13) were located in the caput pancreatis. Palpable lesions mainly appeared as firm white nodules, and all of them tumors in the patients with hypergastrinemia were somewhat had escaped peri-operative palpation, while some were visual~Ore frequently located in caput/corpus pancreatis and the ized by intra-operative ultrasound. Histological examination of the grossly tumor-free pancreatic UOdenum, while tumors of the other patient groups were specimens revealed a considerable number of microadenomas, ~qually common in the head, body, and tail of the pancreas. which varied from slightly larger than a normal islet to a few ~tra-operative ultrasound in 10 patients revealed an additional mm in diameter (Fig. 2A). Pancreatic islets of normal size, "l~l°n-palpable tumors >2 mm in diameter. etastases at time of surgery, or diagnosis in the unoperated cellular morphology, and arrangement (Fig. 2B) were invariably Patients, were found in no patients with insulinoma, in 7 (37%) found close to the microadenomas. The microtumors lacked the

614

World J. Surg. Vol. 16, No. 4, July/Aug. 1992

Fig. 2A. Microadenoma of the pancreas from a patient with MEN-! in chromogranin staining, x 25. B. Normal pancreatic islet of the MEN-I pancreas in chromogranin staining. Note normal cellular organization with peripherally located, intensely stained a-cells and more centrally placed, poorly stained,/3-cells, × 400. C. Macroadenoma of the MEN-I pancreas with fibrotic area and clusters of cells displaying intense pancreatic polypepo tide immunoreactivity, x 25. D. MEN-1 pancreas showing nesidioblastosis, i.e., fibrotic area with exocrine duct proliferation containing clusters of chromogranin reactive endocrine cells, x 180.

normal islet cell organization and they were all chromogranin positive but less intensely than the morphologically normal islets. The pattern of peptide immunoreactivity varied between the microscopic and macroscopic lesions within the same patient. Apart from chromogranin, pancreatic polypeptide immunostaining was most frequent in both the smaller and larger lesions, but several also stained positively for insulin, glucagon, and somatostatin. Gastrin immunoreactivity was essentially absent in the microscopic lesions, and present only in two of the gross pancreatic tumors, as well as in the lymph gland metastases of 1 patient with ZE. Only scattered cells within a singular lesion frequently stained with the individual hormone markers and most tumors displayed immunoreactivity for multiPle peptides. Some of the largest lesions were chromogranin positive and argyrophil but negative for all hormone markers. Amyloid was demonstrated in both microscopic and macroscopic tumors

in 3 of 6 specimens. Some tumors displayed a pronounced fibrotic stroma (Fig. 2C). Even the larger neoplasms were histologically well differentiated and their parenchymal organi" zation was trabecular (40%), solid (40%), and more occasionally insular (20%). Only one tumor of 6 cm in diameter exhibited pronounced cellular polymorphism and areas with necrosis and bleeding, which indicated lower degree of differentiation. Six (43%) of the closely examined pancreatic specimens demonstrated nesidioblastos-like features with multiple areas of fibrosis and proliferating small pancreatic ducts (Fig. 2D). These ductules and buds extending from them contained cluS" ters of endocrine cells with chromogranin and commonly als0 pancreatic polypeptide reactivities, while insulin, glucagon, and somatostatin immunostainings were less frequent. The duod¢" nal tumors stained for gastrin and occasionally also somatO" statin, but not for insulin and glucagon. Submucosal micronO&

l). Grama et al.: Pancreatic Tumors in MEN-1

615

during follow-up. The survival of the patients with MEN-I associated gastrin excess approached that of patients with sporadic gastrinoma collected during the study period from the same centers (Fig. 3).

1oo

80

I~ i ] I I

60

Discussion I

1.1 !

"i

40

II.

2O

I 5O

I 1 O0

I 150

I 200

I 250

months

Fig. 3. Survival curves constructed by the Kaplan-Meier life-table alethod of 14 patients with MEN-I gastrinoma (uninterrupted line), and 9 patients with non-familial gastrinoma [14]. Both groups were treated etween 1979-1986 and subjected tO equally long follow-up; the age at diagnosis/operation averaged 44 years and 53 years for the MEN-1 and SPoradic cases, respectively.

~

Ules of chromogranin-positive endocrine cells were identified in the duodenum of 2 patients, and they were numerous in one of them.

Follow_up All the individuals with symptomatic excess of the pancreatic tumor markers developed symptom-free periods after macroScopically radical tumor resection. Four of the patients with Insulin/proinsulin excess were apparently cured during postoperative follow-up of 7 to 14 years. Two other patients subjected to more extensive biochemical investigation developed slight relapse of the pancreatic tumor markers within I to 3 years, but rio recurrent hypoglycemia. Another of these patients developed liver metastases 180 months postoperatively and then Presented a ZE syndrome. Eight (67%) of the 12 ZE patients SUbjected to grossly radical tumor resection experienced subStantial lowering or normalization of basal serum gastrin values POStoperatively. Within I to 4 years postoperatively, however, all Properly followed patients (n = 6) demonstrated positive Seeretin tests and subsequent rises in basal serum gastrin values and/or increments in variable combinations of the other pancreatic tumor markers. Two of the patients displayed rises in serum insulin. Liver metastases developed during follow-up in 2 Patients with the ZE syndrome, and 1 of the patients with ~°n'functioning lesions developed watery diarrhea hypokale~ia achlorhvdria (WDHA) svndrome and radiological evidence liver met~'s~se~ith 5 fluoi'ouracil or Adriamycin resulted in Streptozo s " of 8 atients, while interferon biochemical response m 2 p atients treatment gave response in 3 of 6 patients. Altogether 5 p ~iith hypergastrinemia and another 4 t~atients with non-funconing lesions succumbed from orogressive tumor disease 11 to 192 rno . • p_.. nths (mean, 88 months) after surgery/diagnosis. All these '~uents initially had liver metastases or developed such spread

The present study supports previous investigations indicating that the lesion of the endocrine pancreas in patients with MEN-1 originates from multiple sites [5-7]. Solitary tumors seemed to be confined to the patients in the present series who were subjected to mere tumor excision often without knowledge of the MEN-I diagnosis and without close peri-operative examination of the pancreas. The numerous and scattered pancreatic microadenomas [6] invariably were surrounded by pancreatic islets of normal size and endocrine cell organization, which suggests that islet cell hyperplasia constitutes an inappropriate denomination of this lesion. This suggestion coincides with the hypothesis of tumorogenesis in inheritable neoplasia [19] as even the minute pancreatic tumors of MEN-l demonstrated monoclonality with respect to the loss of alleles on chromosome 11 [20]. In agreement with the generally prevailing indication for pancreatic exploration in MEN-1 [9, 21, 22], all patients subjected to pancreatic surgery had at least one gross tumor even among those with unsuccessful pre-operative localization studies. Thorough pancreatic palpation and intra-operative ultrasound, however, revealed a variable number of additional, small tumors. Substantially more tumors were found upon histological examination of the serially transected pancreatic specimens. These findings suggest continuous occurrence of the MEN-1 associated somatic mutation, although variable growth potential of the individual lesions cannot be excluded. Considering that malignant transformation of these adenomas may result from a complicated chain of chromosomal events [23], surgical intervention at preferentially early stages theoretically may interfere with the process of malignancy as time could be a requisite for the occurrence of the necessary chromosomal rearrangements [13]. The immunohistochemical stainings most frequently revealed pancreatic polypeptide, insulin, glucagon, and more occasionally, somatostatin reactivity in the pancreatic tumors [6]. Gastrin immunoreactivity, however, was limited to few of the gross pancreatic tumors [6], but the transitory decrease in serum gastrin values after tumor removal in many ZE patients suggested that these lesions nevertheless secreted gastrin. Consistent with previous observations [7], we found endocrine cell proliferation within the duodenal submucosa and mainly small duodenal tumors in all but one of our properly explored patients with the ZE syndrome. Some of these lesions stained for gastrin, others for pancreatic polypeptide or somatostatin, but none with insulin or glucagon. In addition half of the pancreatic specimens displayed nesidioblastos-like features consisting of proliferating pancreatic ducts with clusters of endocrine cells [5]. Considering previous hypotheses on the development of endocrine pancreatic tumors [24], these endocrine cells may constitute an origin for the prevalent pancreatic microadenomas in MEN-1. Extensive biochemical investigation of most patients within the present study displayed rise in several hormone markers for

616

the pancreatic tumors. This finding coincides with comprehensive screening studies of MEN-I families demonstrating elevation of serum pancreatic polypeptide, insulin, proinsulin, and glucagon as well as chromogranin in the affected individuals [13]. A rise in serum gastrin generally seems to occur decades later than the early markers of the MEN-I pancreatic involvement [13]. Furthermore, the present results substantiate that a rise in serum gastrin may indicate the presence of gross tumor within the MEN-I pancreas, provided achylia is excluded. Although a considerable proportion of microadenomas in MEN-1 demonstrate insulin immunoreactivity, hypoglycemia seemed to be restricted to patients harbouring macroscopic (>0.5 cm) pancreatic lesions [6]. Emphasizing the therapeutic difficulties of hypoglycemia, it is generally agreed that MEN-I patients with insulin excess should undergo surgery irrespective of the outcome of the pre-operative localization studies [22, 25]. The present series also indicated that these patients exhibit favorable postoperative courses, since clinical recurrence and malignant transformation were rare and since even the presence of metastases was compatible with longer-term survival. Patients with MEN-I infrequently suffer from a WDHA or VIPoma syndrome [22]. This was only seen in one of our patients, who primarily presented a non-functioning tumor. The severity of this syndrome should indicate attempts to surgically remove the primary pancreatic tumor and metastases, if present. Glucagon immunoreactivity was frequently demonstrated within the pancreatic microadenomas [6], but a glucagonoma syndrome has only rarely been described in MEN-I patients [6, 13]. Consistent with other studies, the ZE syndrome was the most common manifestation and comprised about 60% of the present MEN-I patients. Treatment of these individuals is controversial and ranges from a general reluctance towards surgery to undertaking operation provided larger lesions are radiologically visualized or lateralized by PTP [8, 9, 21, 22, 26]. The restricted precision of pre-operative localization techniques, however, seems to hamper their utility in the selection of operative cases. Although the postoperative lowering of serum gastrin levels was invariably transitory in our patients, considerable symptomatic improvement may be obtained by surgery [26]. However, the peptic ulcer disease may with few exceptions be efficiently controlled also by acid inhibitors, particularly Omeprazole, Hz-receptor blockers, Octreotide, and parathyroidectomy if the patient is hypercalcemic [9]. The rise in gastrin, however, appears to serve as an important marker for the presence of gross tumors, and surgery could thus participate in preventing malignant transformation and dissemination in these patients. As certain MEN-1 kindreds may exhibit pronounced malignant potential of the pancreatic lesions [13], this knowledge rather than the profile of peripheral tumor markers, could constitute a principal indication for pancreatic surgery at earlier stages of the disease. Utilizing this generally liberal attitude towards operation, all patients explored so far have had macroscopic and generally multicentric pancreatic tumors. In kindreds with more benign pancreatic disease, however, radiological demonstration of gross pancreatic tumors may still be utilized to justify operation. MEN-I patients undergoing pancreatic surgery should be subjected to pre-operative CT and selective angiography to attempt localization of gross tumors, but also to substantiate

World J. Surg. Vol. 16, No. 4, July/Aug. 1992

absence of disseminated liver metastases, which generally contradicts surgical intervention. At laparotomy the pancreatic head and uncinate process, as well as the body and tail, are mobilized from the retroperitoneal tissue [9]. The pancreatic neck is freed from the mesenteric vessels and portal vein, whereafter bidigital palpation and scanning by ultrasonography may be performed from the anterior and posterior surfaces of the entire gland [27-30]. The duodenum is longitudinally incised and the mucosa palpated after inversion of the intestine. The distribution of gross tumors within the present study supported the view [22, 25] that the MEN-1 patients preferentially should be subjected to a distal 70% to 80% pancreatic resection, together with enucleation of caput tumors, and clearance of regional lymph nodes as well as excision of accessible liver metastases. A Whipple procedure is only rarely necessary, since even large caput lesions usually can be enucleated and it is in fact the small lesions that may be more difficult to excise. Repeated pancreatic resection or enucleation of recurrent tumors could result in near-total or even total pancreatectomY [31], especially in individuals belonging to kindreds with more highly malignant pancreatic lesions. Metastases of the pancreatic tumors were present at time of diagnosis/operation or developed, generally after considerable delay (5 to 15 years), in 14 of our patients, indicating malig" nancy rates of 14% for the patients with hyperinsulinism, 47~ for those with hypergastrinemia, and 57% in the patients with non-functioning lesions. Four patients with hypergastrinemia and another 4 patients with non-functioning tumors died with liver metastases and progressive tumor disease. Only moderately higher rates of malignancy are reported for patients with corresponding types of sporadic endocrine pancreatic tumors [10, 14]. The longer time of survival which characterized the MEN-I patients compared to sporadic patients with the ZI~ syndrome was accompanied by considerably lower age at the time of diagnosis and, when longer term follow-up was extended to comparable ages, the rates of survival tended to become similar in the two patient groups. This seems to substantiate that development of malignancy in MEN-I endo" crine pancreatic tumors is time-dependent and may lend further support to early surgery in these patients. R~sum~

Parmi 33 patients ayant une tumeur pancr6atique endocrine due b. une n6oplasie endocrine multiple de type 1 (MEN-l), 19 (58°~) avaient une hypergastrin6mie, 7 (21%) un hyperinsulinisme et 7 (21%) une 16sion cliniquement muette. On a mis en 6vidence a0 minimum une grosse tumeur chez tousles patients, y compris chez ceux dont les examens pr6op6ratoires de d6pistage ttlmoral 6taient n6gatifs. Les patients 6taient 6galement porteurS de tumeurs macroscopiques et de nombreux microad6nomeS" Les 16sions montraient souvent un immunomarquage positif pour de multiples hormones, principalement le polypeptide pancr6atique, l'insuline, le glucagon et la somatostatine. Des 16sions endocrines duod6nales furent retrouv6es chez 4 des 5 patients explor6s; elles montraient un immunomarquage ave¢ les anticorps angigastrine et anti-somatostatine. Une r6sectio0 pancr6atique distale, le plus souvent subtotale, a 6t6 r6alis6¢ chez 18 patients. Elle 6tait 6ventuellement compl6t6e par uoe 6nucl6ation tmorale de la t6te ou par une duod6notomie, peu de

1). Grama et al.: Pancreatic Tumors in MEN-I

Patients ont b6n6fici6 d ' u n e simple 6nucl6ation ou d'une intervention de Whipple. L'6volution postop6ratoire h long terme a 6t6 plus favorable en cas d'insulinome puisque seul un patient a eu une r6cidive clinique. Les patients atteints de gastrinome n'ont pr6sent6 que transitoirement une diminution des taux S6riques de gastrine apr6s la chirurgie pancr6atique. Quarante sept pour cent de ces patients avaient ou ont d6velopp6 des rn6tastases contre 57% des patients porteurs de 16sions sans traduction clinique. N e u f patients sont d6c6d6s en raison de l'extension tumorale au cours du suivi. Conform6ment ~ des SUggestions ant6rieures, la chirurgie semble indiqu6e chez les Patients atteints de M E N - I avec hyperinsulinisme m6me si la radiologie ne visualise pas de 16sion. Mais cette indication peut 6tre 61argie aux patients dont seuls les param~tres biologiques Sont en faveur d ' u n e grosse tumeur (dont l'hypergastrin6mie). Cette strat6gie pourrait convenir particuli~rement aux patients ayant des ant6c6dents familiaux importants; elle permettrait Peut-~tre de r6duire le risque d'extension tumorale. Resumen

Entre 33 individuos con tumores pancre~iticos endocrinos como COmponente del sfndrome de neoplasia endocrina mtiltiple tipo 1 (NEM-1), 19 pacientes (58%) tenfan hipergastrinemia, 7 (21%) hiperinsulinismo y 7 (21%) lesiones clinicas " n o funcionantes". En la totalidad de los pacientes sometidos a cirugfa pancre~itica rue hallado por lo menos un tumor, incluso en aquellos con exarnenes de localizaci6n negativos anteriores a ta operaci6n. EStos pacientes tambi6n albergaban tumores macrosc6picos, as[ como numerosos microadenomas; con frecuencia las leSiones demostraron inmunocoloraci6n con diferentes hormohas, principalmente polip6ptido, insulina, glucag6n y somatostatina. Se encontraron lesiones endocrinas duodenales en 4 de cada 5 pacientes investigados, las cuales colorearon con gastrina y anticuerpos a la somatostatina. Se practic6 resecci6n Pancrefitica distal (principalmente resecci6n subtotal) en 18 Pacientes, eventualmente combinada con enucleaci6n del turaor (cuando 6ste se hallaba ubicado en la cabeza del pancreas) o duodenectomia; solamente unos pocos pacientes fueron Sornetidos a simple enucleaci6n del tumor o al procedimiento de Whipple. El resultado a largo plazo fue m~is favorable en los Pacientes con hiperinsulinismo, puesto que s61o uno present6 recurrencia clfnica. Los pacientes con hipergastrinemia exhibieron apenas una disminuci6n transitoria de los valores de gastrina s6rica luego de la cirugia pancre~itica. Cuarenta y siete Pot ciento del conjunto tuvo o desarroll6 met,'lstasis, en tanto que la extensi6n local del tumor se present6 en 57% de los casos Con lesiones no funcionantes. Nueve pacientes murieron por Progresi6n de la neoplasia en el curso del seguimiento. En aeuerdo con sugerencias previas, se considero quo la cirugfa est~i indicada en pacientes con NEM-I e hiperinsulinismo, at~n en los casos en que no se visualiza radiol6gicamente la lesi6n, Iaero que la indicaci6n puede ser ampliada para incluir tambi6n Pacientes con s61o marcadores bioquimicos, tales como niveles elevados de gastrina, indicativos de la presencia de tumores ~acrosc6picos. Esta estrategia debe ser aplicada principalente en aquellos pacientes con historia familiar agresiva, con t~ CUal tal vez se reduce el riesgo de progesi6n maligna del urnor.

~

617 Acknowledgments

Supported by the Swedish Medical Research Council (Study group for endocrine abdominal tumors) and the Swedish Cancer Society. References

1. Ballard, H.S., Frame, B., Hartsock, R.J.: Familial multiple endocrine adenoma-peptic ulcer complex. Medicine (Baltimore) 43:481, 1964 2. Marx, S.J., Spiegel, A.M., Brown, E.M., Aurbach, G.D.: Family studies in patients with primary parathyroid hyperplasia. Am. J. Med. 62:698, 1977 3. Friesen, S.R., Kimmel, J.R., Tomita, T.: Pancreatic polypeptide as screening marker for pancreatic polypeptide apudomas in multiple endocrinopathies. Am. J. Surg. 139:61, 1980 4. Skogseid, B., (3berg, K.; Benson, L., Lindgren, P.G., L6relius, L.-E., Lundqvist, G., Wide, L., Wilander, E.: A standardized meal stimulation test of the endocrine pancreas for early detection of pancreatic endocrine neoplasia type 1 syndrome: Five years' experience. J. Clin. Endocrinol. Metab. 64:1233, 1987 5. Thompson, N.W., Lloyd, R.V., Nishiyama, R.H., Vinik, A.I., Strodel, W.E., Allo, M.D., Eckhauser, F.E., Talpos, G., Mervak, T.: MEN 1 pancreas: A histological and immunohistochemical study. World J. Surg. 8:561, 1984 6. Kl6ppel, G., Willemer, S., Stamm, B., H~icki, W.H., Heitz, P.U.: Pancreatic lesions and hormonal profile of pancreatic tumors in multiple endocrine neoplasia type 1: An immunocytochemical study in nine patients. Cancer 57:1824, 1986 7. Pipelleers-Marichal, M., Somers, G., Willems, G., Foulis, A., lmrie, C., Bishop, A.E., Polak, J.M., H~icki, W.H., Stamm, B., Heitz, P.U., Path, F.R.C., Kl6ppel, G.: Gastrinomas in the duodenum of patients with multiple endocrine neoplasia type 1 and the Zollinger-Ellison syndrome. N. Engl. J. Med. 322:723, 1990 8. van Heerden, J.A., Smith, S.L., Miller, L.J.: Management of the Zollinger-Ellison syndrome in patients with multiple endocrine neoplasia type 1. Surgery 100:971, 1986 9. Fraker, D.L., Norton, J.A.: The role of surgery in the management of islet cell tumors. Gastroenterol. Clin. North Am. •8:805, 1989 10. Friesen, S.R,: Treatment of the Zollinger-Ellison syndrome: A 25-year assessment. Am. J. Surg. 142:331, 1982 11. Gogel, H.K., Buchman, M.T., Cadieux, D., McCarthy, D.M.: Gastric secretion and hormonal interactions in multiple endocrine neoplasia type 1. Arch. Intern. Med. 145:855, 1985 12. Eriksson, B.: Recent advances in the diagnosis and management of endocrine pancreatic tumors (Dissertation, Comprehensive summaries). Acta Univ. Upsalla 160:7, 1988 13. Skogseid, B., Eriksson, B., Lundqvist, G., L6relius, L.-E., Rastad, J., Wide, L., Wilander, E., Akerstr6m, G., Oberg, K.: Multiple endocrine neoplasia type 1: A ten-year prospective screening study in four kindreds. J. Clin. Endocrinol. Metab. 73:281, 1991 14. Grama, D., Eriksson, B., MS.rtensson, H., Cedermark, B., Ahr6n, B., Kristoffersson, A., Rastad, J., Oberg, K., Akerstr6m, G.: Clinical characteristics, treatment and survival of patients with pancreatic tumors causing hormonal syndromes. World J. Surg. 16:632, 1992 I5. Puchtler, H., Sweat, F., Levine, M.: On the binding of Congo red by amyloid. J. Histochem. Cytochem. •0:355, 1982 16. Singh, 1.: A modification of the Masson-Hamperi method for staining of argentaffin cells. Anat. Anz. 115:81, 1964 17. Grimelius, L., Wilander, E.: Silver stains in the study of endocrine cell of the gut and pancreas. Invest. Cell Pathol. 3:3, 1980 18. Eriksson, B., Arnberg, H., Oberg, K., Hellman, V., Lundquist, G., Wernstedt, C., Wilander, E.: A polyclonal antiserum against chromogranin A and B: A new sensitive marker for nem-oendocrine tumours. Acta Endocrinol. (Copenh.) 122:145, 1990 19. Knudson, A.G.: Mutation and cancer: Statistical study of retinoblastoma. Proc. Natl. Acad. Sci. U.S.A. 68:820, 1971 20. Larsson, C., Skogseid, B., Oberg, K., Nakamura, Y., Norden-

618

21. 22.

23.

24. 25.

World J. Surg. Vol. 16, No. 4, July/Aug. 1992

skrld, M.: Multiple endocrine neoplasia gene maps to chromosome 11 and is lost in insulinoma. Nature 332:85, 1988 Ellison, E.C., Carey, L.C., Sparks, J., O'Dorisio, T.M., Mekhjian, H.S., Fromkes, J.J., Caldwell, J.H., Thomas, F.B.: Early surgical treatment of gastrinomas. Am. J. Med. 82:!7 (Suppl. 5B), 1987 Sheppard, B.C., Norton, J.A., Doppman, J.L., Maton, P.N., Gardner, J.D., Jensen, R.T.: Management of islet cell tumors in patients with multiple endocrine neoplasia: A prospective study. Surgery 106:1108, 1989 Vogelstein, B., Fearon, E.R., Hamilton, S.R., Kern, S.E., Preisinger, A.C., Leppert, M., Nakamura, Y., White, R., Smith, A.M.M., Bos, J.: Genetic alteration during colorectal tumor development. N. Engl. J. Med. 319:525, 1988 Rutter, W.A.: The development of the endocrine and exocrine pancreas. In The Pancreas, P.J. Fitzgerald, A.B., editors, Baltimore, Williams and Wilkins, 1980, pp. 30-38 Rasbach, D.A., van Heerden, J.A., Telander, R.L., Grant, C.S., Carney, J.A.: Surgical management of hyperinsulinism in the multiple endocrine neoplasia, type 1 syndrome. Arch. Surg. 120: 584, 1985

Invited Commentary J.A. Norton, M.D., and J.R. Lange, M.D. Surgical Metabolism Section, Surgery Branch, National Cancer Institute, Bethesda, Maryland, U.S.A.

There is considerable controversy about the management of the pancreatic islet cell tumors in patients with multiple endocrine neoplasia type 1 (MEN-I). Grama and coworkers document some of the difficulties in the management of the pancreatic neuroendocrine tumors in these patients. As has been reported by others, they found that the Zollinger-Ellison (ZE) syndrome or gastrinoma is the most common functional islet cell tumor in the M E N - I patient population and that islet cell tumors are commonly multifocal in these patients. Patients with ZE were not cured of the hypergastrin.emia despite the fact that duodenal tumors were removed in 4 patients and that the tumors stained positive for gastrin by immunoperoxidase. In contrast, the MEN-1 patients with insulinoma nearly always had an apparent cure following excision of a dominant pancreatic islet cell tumor. Grama and associates found that the malignant potential of the pancreatic islet cell tumors in M E N - I patients was similar to that in sporadically occurring islet cell tumors. This finding differs from that previously reported by Malagelada and colleagues from the Mayo clinic [1]. Grama reported metastases at the time of diagnosis in 37% of patients with Zollinger-Ellison syndrome and MEN-1 as well as in 43% of patients with non-functional islet cell tumors and MEN-1. Patients who died during the follow-up period each died of the malignant spread of tumor to the liver. This large, retrospective study provides valuable insight into the management of these patients and may be useful in planning prospective strategies to better treat these islet cell tumors. We now consider some areas in the management of MEN-1 patients with islet cell neoplasms that were either n o t r e p o r t e d in the above work or that are especially controversial and will require additional studies and protocols to resolve. Grama and associates did not give information regarding family history or

26. Thompson, N.W., Bondeson, A.-G., Bondeson, L., Vinik, A.: The surgical treatment of gastrinoma in MEN I syndrome patients. Surgery 106:1081, 1986 27. Gigot, J.F., Gianello, P., Dardanne, A.N., Pringot, J., Detry, R., Otte, J.B., Kestens, P.J.: lntraoperative ultrasonography in endocrine pancreatic surgery: Preliminary results in 6 cases of insulinoma. J. Belge Radiol. 69:57, 1986 28. Grant, C.S., van Heerden, J.A., Charboneau, J.W., James, E.M., Reading, C.C.: Insulinoma: The value of intraoperative ultrasonography. Arch. Surg. 123:843, 1988 29. Norton, J.A., Cromack, D.T., Shawker, T.H., Doppman, J.L., Comi, R., Gorden, P., Maton, P.N., Gardner, J.D., Jensen, R.T.: Intraoperative ultrasonographic localization of islet cell tumors: A prospective comparison to palpation. Ann. Surg. 207:160, 1988 30. Skogseid, B., Grama, D., Ahlstrrm, H., Andersson, T., Eriksson, B., Lindgren, P.-G., Rastad, J., Wilander, E., /~kerstr6m, G., Oberg, K.: Pre- and intra-operative localization of pancreatic tumors in multiple endocrine neoplasia type 1. (Submitted) 31. Tisell, L.E., Ahlman, H., Jansson, S.: Total pancreatectomy in the MEN I syndrome. Br. J. Surg. 75:154, 1988

associated endocrinopathies in their patient population. In the management of patients with islet cell tumors, exact determi" nation of the presence or absence of MEN-1 is important, as it may alter management. Currently the diagnosis of MEN-1 is made on the basis of a positive family history and/or the occurrence of associated endocrinopathies. The clinical evidence for M E N - I is circumstantial at best. A specific probe for the gene responsible for MEN-1 is needed. Since the MEN-I gene is known to map to chromosome 11 [2, 3], a specific gene marker will no doubt soon be available. An individual patient could then have specific and sensitive documentation of MEN-I and management decisions altered accordingly. Patients with Z E and M E N - I commonly have primary hyperparathyroidism pre-existing at the time of diagnosis of the ZI~ [4]. Surgical correction of the primary hyperparathyroidisra may dramatically lower serum levels of gastrin and attenuate the gastric acid hypersecretion associated with ZE. Such patients may then be able to reduce the dosage of anti-ulcer medications. Approximately 20% of such patients may no longer have any biochemical evidence of ZE [4]. In contrast, gastrinoma excision in M E N - I patients does not result i0 significant or long-lasting reduction in serum levels of gastrin nor does it improve medical management [1, 5]. Therefore, correction of associated hyperparathyroidism appears to con" trol the symptoms of Z E better than does surgery directed at gastrinoma excision. Many factors contribute to the controversy over what con" stitutes appropriate surgical management of the islet cell tumors in these patients. First, the tumors are multifocal within the pancreas [5, 6] and can be extrapancreatic and malignant [I, 7J. Second, the malignant potential of islet cell tumors in MEN "1 patients has been reported to differ from kindred to kindred, b0t to be rather homogeneous among the members of a given kindred [8]. Third, long-term results differ with histologY; patients with insulinoma or VIPoma have been found to have a better outlook than those with gastrinoma [1, 7]. With these complexities in mind, we suggest the following strategies for the management of islet cell tumors in patients with MEN-1. F o r patients with M E N - I and evidence of insulinonaa,