Prognostic factors and survival in MEN1 patients with gastrinomas: Results from the DutchMEN study group (DMSG)

Abstract Background and objectives Gastrinomas are the most prevalent functioning neuroendocrine tumors (NET) in multiple endocrine neoplasia type 1 (MEN1). Guidelines suggest medical therapy in most patients, but surgery may be considered in a subgroup. Currently, factors to guide management are necessary. This population‐based cohort study assessed prognostic factors of survival in patients with MEN1‐related gastrinomas. Methods Patients with MEN1 having gastrinomas were identified in the Dutch MEN1 database from 1990 to 2014 based on fasting serum gastrin (FSG) levels and/or pathology. Predictors of overall survival were assessed using Cox regression. Results Sixty‐three patients with gastrinoma (16% of the MEN1 population) were identified. Five‐ and 10‐year overall survival rates were 83% and 65%, respectively. Prognostic factors associated with overall survival were initial FSG levels ≥20x upper limit of normal (ULN) (hazard ratio [HR], 6.2 [95% confidence interval, 1.7‐23.0]), pancreatic NET ≥2 cm (HR 4.5; [1.5‐13.1]), synchronous liver metastases (HR 8.9; [2.1‐36.7]), gastroduodenoscopy suspicious for gastric NETs (HR 12.7; [1.4‐115.6]), and multiple concurrent NETs (HR 5.9; [1.2‐27.7]). Conclusion Life expectancy of patients with MEN1 gastrinoma is reduced. FSG levels and pancreatic NETs ≥2 cm are prognostic factors. FSG levels might guide surveillance intensity, step‐up to additional diagnostics, or provide arguments in selecting patients who might benefit from surgery.

Conclusion: Life expectancy of patients with MEN1 gastrinoma is reduced. FSG levels and pancreatic NETs ≥2 cm are prognostic factors. FSG levels might guide surveillance intensity, step-up to additional diagnostics, or provide arguments in selecting patients who might benefit from surgery. Gastrinomas are the most frequently encountered functioning dpNETs and occur in approximately 30% of patients with MEN1. 3 These tumors produce gastrin which induces gastric acid hypersecretion and subsequently leads to ulcerative peptic disease and gastrointestinal bleeding, known as the Zollinger-Ellison syndrome. 4 MEN1-related gastrinomas are generally located in the duodenal submucosa and are rarely found in the pancreas. 5,6 Duodenal gastrinomas are often small (<1 cm), multiple and accompanied by pancreatic neuroendocrine tumors (pNETs). 6 Approximately 70% to 80% of surgically treated patients have lymph node metastases, and 10% present with synchronous liver metastases. 7,8 Gastric acid hypersecretion-related complications used to be the leading cause of death in patients with MEN1 having gastrinomas before the widespread use of proton pump inhibitors (PPIs). 9,10 Nowadays, compared with the general population, patients with MEN1 have a seriously decreased life expectancy mainly caused by malignant dpNETs. 2 However, the reported prognosis of patients with MEN1 gastrinoma varies widely. 8,[11][12][13] In the French cohort, studied from 1956 to 2005, gastrinomas have been reported as an independent risk factor for death. 9 Actual data on MEN1 gastrinoma survival are scarce and survival rates are difficult to interpret since patients are diagnosed and treated differently among studies. In addition, the understanding that MEN1 gastrinomas mostly originate in the duodenum instead of the formerly assumed pancreatic origin, emphasizes the need for new studies. Besides, data regarding the long-term natural history are important, because guidelines suggest symptomatic management using PPIs in the majority of patients. 14,15 Nevertheless, the only potentially curative oncological treatment remains surgery. Pancreaticoduodenectomy offers the possibility to achieve a biochemical cure for MEN1-related duodenal gastrinomas. 8 However, controversies exist regarding the timing and the extent of surgery, considering the unpredictable tumor course and morbidity associated with extensive surgery. 14,15 Therefore, the necessity of prognostic factors to guide therapy has recently been underscored. 16 Since gastrinomas are hormone producing tumors, we hypothesized that, besides known dpNET-related prognostic factors such as pNET size and liver metastases, gastrin levels might predict survival in this population. 17 Therefore the present study aims to assess prognostic factors and survival in patients with MEN1 having gastrinomas.

| Study design
Patients were selected from the national Dutch MEN1 database from the DutchMEN Study Group (DMSG). 18

| Patient selection
The diagnosis of gastrinomas in patients with MEN1 was challenging, since the reference standard, provocative tests using secretin, is not widely available and routine measurements of basal acid output at gastroduodenoscopy were not routinely performed. Therefore, based on stringent criteria we aimed to identify those patients of whom we were confident of having gastrinomas, also using subsequently elevated fasting serum gastrin (FSG) levels. Gastrinoma diagnosis was based on (a) pathology reports of gastrin immunohistochemistry positive tumors or (b) elevated FSG levels, or (c) gastroduodenoscopy suspicious for gastrinoma, or a combination of these. Serum gastrin reference values were obtained from all UMCs over the study period.
FSG levels were calculated as a factor of the upper limit of normal VAN   Deaths caused by MEN1 manifestations and MEN1-related therapy were considered as MEN1 related. Other causes of death were regarded as non-MEN1 related. 2 MEN1-related NETs at the moment of gastrinoma diagnosis were diagnosed according to pathology reports. 2 If no pathology reports were available, imaging results were used for diagnosis as previously described. 19 Gastric NETs were also diagnosed using gastroduodenoscopies. The size of the largest pNET on conventional imaging was used for further analysis.

| Treatment
Patients were treated medically by PPIs to prevent acid-related complications by somatostatin analogs or surgically to prevent metastatic disease. Treatment regimen was decided by the treating

| Outcome measures
The primary outcomes were 5-and 10-year overall survival (OS).
Possible prognostic factors at gastrinoma diagnosis were analyzed for influence on OS. The date of death or date of last follow-up was used for analysis. Origin of liver metastases is based on the expert panel. Percentage is based on the group of patients without liver metastases at diagnosis (n = 58). VAN

| Patient characteristics
Demographic and clinical characteristics are described in Table 1.
Sixty-three patients with gastrinoma (16%) were identified in the

| Long-term outcomes
Patient outcomes are reported in Table 2

| Survival of patients with MEN1 gastrinoma
OS rates of patients with MEN1 gastrinoma after 5 and 10 years were 83% and 65%, respectively (Table 2; Figures 1A and S2A). Five and 10-year OS rates for patients with MEN1 having FSG measurements not indicative for a gastrinoma were 93% and 87%, respectively ( Figures 1B and S2B). Patient with gastrinomas were older than patients without gastrinomas (51 vs 39 years, P < .001).
Ten-year OS rates were 65% vs 81% for age and gender matched MEN1 patients without gastrinomas (Figures 1C and S2C). Table 3   Abbreviations: HR, hazard ratio; NA, not applicable; NET, neuroendocrine tumor; PPI, proton pump inhibitor; ULN, upper limit of the normal of the reference value. * Conventional imaging: magnetic resonance imaging, computed tomography, endoscopic ultrasonography, or gastroduodenoscopy suspicious for duodenal gastrinoma. In one case the gastroduodenoscopy was suspicious for a gastric gastrinoma.

Prognostic factors for OS are shown in
F I G U R E 2 Overall survival of patients with MEN1 according to initial fasting serum gastrin levels. MEN1, multiple endocrine neoplasia type 1 removed, which is deemed necessary for achieving biochemical cure. 8 In addition, patients with synchronous liver metastases were included in our study. Liver metastases, either NF-pNET or gastrinoma related, are associated with survival. 7,11,13,17,21 Although the OS rate was lower, the age of death (58 years) was slightly higher than in previous series (55-56 years). [10][11][12] Initial FSG levels were associated OS, also after adjusting for age.
More specifically, OS decreased as FSG levels increased. Ito et al 10 described very high FSG levels (>20-fold elevated) more often in deceased patients. We observed a significantly increased HR for death in patients with FSG levels higher than 20x ULN. In line, we observed a HR of 2.66 and a 10-year OS of 66% for patients with FSG levels between 10 and 20x ULN. We believe that the outcomes of this analysis were not statistically significant due to the low number of patients and events in this subgroup. NIH series observed higher FSG levels in patients with an aggressive disease course and in patients with liver metastases, although no survival analysis was conducted. 7,13 The only study focusing on initial FSG levels and survival in patients with MEN1 from the NIH, did not find a significant correlation in MEN1 gastrinoma patients (log rank P = .068). 22

This study was supported by an unrestricted grant from Ipsen
Pharmaceutical. The funding source had no influence on the study question, design, data acquisition, statistical analysis, and interpretation of data.

CONFLICT OF INTERESTS
There is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

AUTHOR CONTRIBUTIONS
DJvB: study design, statistical analysis and interpretation of data, drafting, and final approval of the manuscript.
SN: study design, acquisition of data, interpretation of data, critical revision, and final approval of the manuscript.
CRCP: study design, design of data collection protocol, acquisition of data, critical revision, and final approval of the manuscript.
WWdH: study design, critical revision, and final approval of the manuscript.
ACvdV: study design, critical revision, and final approval of the manuscript.
OMD: study design, critical revision, and final approval of the manuscript.
ANvdH-S: study design, critical revision, and final approval of the manuscript.
MLD: study design, critical revision, and final approval of the manuscript.
PHB: study design, critical revision, and final approval of the manuscript.
BH: study design, critical revision, and final approval of the manuscript.
IHMBR: study design, critical revision, and final approval of the manuscript.
MRV: study design, interpretation of data, critical revision, and final approval of the manuscript and study supervision.
GDV: study design, interpretation of data, critical revision, and final approval of the manuscript and study supervision.

DATA AVAILABILITY
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

DISCLOSURE SUMMARY
The authors have nothing to disclose.