When ZES was first described, the gastrinomas were mainly pancreatic tumors, and patients suffered from refractory peptic ulcer disease, vomiting, diarrhea, and weight loss (Table 13.2). Treatment was based on controlling the acid production (total gastrectomy) and resection of the pancreatic tumor if possible. Nowadays, gastrinomas are often small and up to 60 % are located in the duodenal wall. Symptoms are often those of peptic ulcer disease without diarrhea and weight loss, as PPIs are very effective at controlling the gastric acid secretions. Delay in the diagnosis of ZES has increased as a result of the effective use of PPIs [5, 23–25].

Gastrinomas are rare tumors, occurring with an incidence of 0.5 cases per million population per year [23, 26]. Gastrinomas appear to be more frequently in males, with a male to female ratio of 1.3:1–1.5:1 [4, 24, 27]. The tumor typically presents in late adulthood with a mean age at diagnosis between 41 and 54 years, earlier for patients with MEN1 (33 years) [1, 23, 28, 29]. Historically, patients with gastrinoma were typically diagnosed late in the course of their disease with complications of severe refractory peptic ulcer disease. In the modern era of the use of PPI therapy, a history of peptic ulceration is absent in 29 % of patients at the time of diagnosis [23]. Although many patients with ZES present with more severe manifestations of peptic ulcer disease, the vast majority of ZES patients are indistinguishable from idiopathic peptic ulcer disease [23]. As a result, many patients are initially treated with PPIs early in the course of their disease, leading to an improvement or resolution of their symptoms and a potential development of metastatic disease before a diagnosis is established [1, 5, 23, 24]. Ellison reported an increase in advance stage at presentation when patients were diagnosed between 1986 and 1998. Patients diagnosed up to 1985 had an incidence of metastatic disease of 19 % and a 5-year disease-free survival (DFS) of 29 % compared with an incidence of metastatic disease of 55 % and 5-year DFS of 2 % after 1986 [30].

ZES patients will present primarily with pain (83 %) and diarrhea (71 %) in a background of long-standing history of reflux (GERD) and recurring peptic ulcer disease (PUD) (Table 13.2). Only 11 % of patients have a single symptom. Complicated PUD is decreasing; in 1964, 45 % of patients presented with life-threatening bleeding, whereas in the NIH report from 1999, only 24 % presented with bleeding [1, 3]. Up to 20–25 % of gastrinomas occur in the setting of multiple endocrine neoplasia type 1 (MEN1). The diagnosis of ZES should be considered in any patient with PUD who is H. pylori negative, is refractory to standard treatment, has ulcers in unusual locations, or has MEN1 syndrome [5].

The diagnosis of gastrinoma is based on biochemical testing, demonstrating an elevation of fasting serum gastrin (FSG) level (>150 pg/ml) in the presence of raised gastric acid secretion [28]. In patients receiving acid-suppressing therapy with H2-receptor antagonists or PPIs, FSG is secondarily raised by these medications, and as such, these should be discontinued prior to testing [5, 23]. Gastric secretory studies showing a BAO of more than 15 mEq/h (more than 5 mEq/h in the presence of previous gastric resection) or a pH of less than 2 are the most commonly used criteria to determine acid hypersecretion [28]. FSG can also be elevated in other disease states that must be differentiated from ZES (Table 13.1). Most patients with gastrinoma have highly elevated gastrin levels, with 36 % having levels more than 10 times normal [31]. When the results of FSG and acid secretory studies are borderline or nondiagnostic in patients where a clinical suspicion for ZES exists, provocative testing of FSG after administration of intravenous secretin (2 mg/kg) or calcium is indicated. Because most of these patients are on PPIs, these drugs must be stopped a week before provocative testing. H2-receptor antagonists may be utilized for symptom control, but also must be stopped 48 h before [23]. There is a risk for developing rebound hypersecretion following PPI withdrawal. The day of the secretin test is the ideal time to get a gastric pH and upper endoscopic ultrasound (EUS) to avoid having to withdraw PPIs again at a later date. Following secretin injection, FSG increases from baseline levels within 2–3 min by more than 100 pg/ml in 94 % of patients with gastrinoma and rises by more than 50 % in 84 % of cases [32]. Similarly, provocative testing after calcium infusion results in a rise in gastrin of more than 50 % from baseline levels in 84 % of cases [5]. Combining these two investigations lead to an increased sensitivity for confirming the diagnosis of a gastrinoma. Synthetic secretin has replaced native secretin. An analysis of 293 NIH ZES patients and 537 patients from the literature has established the criteria of a rise ≥120 pg/mL (sensitivity 94 %, specificity 100 %) replacing the old criteria of >200 pg/mL [5, 31, 33].

Chromogranin-A (CgA), a protein contained within the secretory granules of neuroendocrine cells, is a nonspecific marker for neuroendocrine tumors (NET) and is elevated in 90 % of patients with gastrinoma [34]. Chromogranin-A levels tend to correlate with those of serum gastrin; however, higher levels are not associated with an increased risk of advanced or metastatic disease. However, in a series of 72 patients with serial measurements of gastrin and CgA, these markers did not have sufficient sensitivity to replace serial imaging for the detection of small but important changes in tumor burden [35].

Although sporadic ZES was first described in association with neuroendocrine tumors of the pancreas (pNETs), the primary tumor can be located in the duodenum in an equivalent proportion of cases and may also arise uncommonly from other sites [1, 4, 36]. In 90 % of cases, the primary tumor can be localized within the “gastrinoma or Stabile’s triangle,” a region encompassing the proximal duodenum and head of the pancreas [37]. This triangle is defined by (1) the junction of the cystic and common bile ducts (superiorly), (2) the junction of the second and third parts of the duodenum (inferiorly), and (3) the junction of the neck and body of the pancreas (medially). Duodenal gastrinomas are small tumors arising within the submucosa and are found in the first and second parts of the duodenum in 83–92 % of cases and rarely distal to this [20, 29, 38] (Fig. 13.1). Mean size ranges from 0.6 to 1.3 cm in diameter and 49–80 % of the tumors are less than 1 cm in size, and for this reason, they are rarely seen on imaging studies [4, 17, 38–40]. The pancreas is the site of gastrinoma in 17–55 % of patients with sporadic ZES, and the tumor is located proximally in the head of the pancreas in 33–53 % of cases [1, 4, 36, 41]. Gastrinoma is the second most common functional neuroendocrine tumor of the pancreas, accounting for 36 % of lesions [42, 43]. Pancreatic gastrinomas tend to be larger than their duodenal counterparts with a mean diameter of 2.7–3.2 cm [4, 38, 39, 41, 42]. Sporadic gastrinomas rarely present in multiple sites in 11–16 % of cases, and pancreatic and duodenal gastrinomas may coexist in 9–11 % of cases [4, 27]. In contrast, greater than 75 % of patients with ZES/MEN1 have multiple pancreatic and duodenal neuroendocrine tumors [44, 45].

Ectopic gastrinomas arising outside the gastrinoma triangle are rare accounting for only 6 % of reported ZES cases [46]. Primary ectopic gastrinomas have been reported in the liver, biliary tree, ovary, kidney, heart, stomach, jejunum, and greater or lesser omentum [1, 5, 47, 48]. The occurrence of primary lymph node gastrinomas (primary LN) is controversial. A primary LN gastrinoma is defined by the following conditions: (1) removal of an involved lymph node in the absence of a gastrinoma primarily found elsewhere and (2) normal postoperative FSG and a negative secretin stimulation test [1]. In a large series from NIH, 26 % of patients with ZES undergoing laparotomy that had only involved lymph nodes removed without identification of a primary tumor of the pancreas or duodenum resulted in normalization of the gastrin postoperatively [36]. Whether these nodal lesions represent metastases from a small primary tumor that evaded detection at the time of initial operation or are a true primary remains controversial. In a long-term follow-up from this NIH series, 69 % of the primary LN gastrinomas remained biochemically cured leading these authors to conclude that lymph nodes can be the primary site of gastrinoma in 10 % of patients with ZES [49]. It has been demonstrated that primary LN do not occur in ZES/MEN1 and that the designation as a lymph node primary should be tentative until 10 years after surgery [1, 50].

Metastatic disease affecting regional lymph nodes or the liver can be demonstrated in 52–69 % of patients with gastrinoma on initial imaging studies or at the time of laparotomy [36, 51, 52]. Involvement of regional lymph nodes, particularly peripancreatic and periduodenal nodes, can be demonstrated in 40–75 % of patients with duodenal gastrinoma and 50 % for pancreatic [5, 51, 53]. In contrast to metastatic liver disease, regional lymph node involvement alone has not been associated with a reduction in survival, with an overall 10-year survival approaching 100 % [4, 53, 54]. Hepatic metastases are present in 8–22 % of patients with gastrinoma at the time of initial assessment [51, 53, 54]. Several authors have found that pancreatic gastrinomas are more frequently associated with hepatic metastases than were duodenal gastrinomas (52 % vs. 5 %) [4, 53]. The size of the primary tumor was associated with an increased risk of hepatic metastases, with tumors smaller than 1 cm having a 4 % risk of hepatic metastases, tumors 1–3 cm in size having metastases in 28 % of cases, and 61 % of tumors greater than 3 cm [53, 55]. The development of hepatic metastases significantly worsens survival and is associated with a 10-year survival rate of 26–30 % [53, 54]. Other areas of metastatic disease include peritoneal disease in 6 %, bony metastases in 3–7 %, and lung metastases in 2 % [27, 56].

Germ-line mutation of the MEN1 gene, a tumor suppressor gene located on the short arm of chromosome 11 (11q31), is responsible for the MEN1 syndrome [6]. The MEN1 gene codes for the nuclear protein menin which is involved in cell growth regulation. Twenty-five percent of gastrinomas are associated with the MEN1 syndrome [1, 7]. Mutations of the MEN1 gene are inherited in an autosomal dominant pattern, and between 36–42 % of patients with this mutation will develop gastrinoma [44]. Mutation of the MEN1 gene has also been implicated in the pathogenesis of sporadic gastrinoma. Mutations of this gene have been found between 17 and 58 % of gastrinomas in patients who do not possess germ-line MEN1 mutations or a family history of MEN1-related tumors [57, 58]. It has been proposed that MEN1 gene mutation is an early event in the pathogenesis of both sporadic and familial gastrinomas, with mutations occurring equally in duodenal and pancreatic gastrinomas [7]. The biological behavior of gastrinomas displaying MEN1 gene mutation is similar to those where this mutation is not found, with no increased risk of metastatic disease or of recurrence after surgical resection [57, 58].

A number of other genetic mechanisms are involved in the pathogenesis of gastrinoma in addition to mutations affecting the MEN1 gene. Chromosomal deletions causing loss of heterozygosity of chromosomes 1q and 22q have also been identified in gastrinomas [59, 60]. Inactivation or deletion of the p16 tumor suppressor gene at chromosome 9p21, coding for the p16 protein involved in the regulation of cell growth and cell cycle inhibition, has also been demonstrated in 52 % of gastrinomas [61]. Gene expression using microarrays has identified numerous additional gene alterations in pancreatic NETs; however, there is no clear concordance between the studies, and none have established as clearly important for pathogenesis [5, 7, 55, 62].

Cross-sectional imaging with computed tomography (CT) and magnetic resonance (MRI), somatostatin receptor scintigraphy (SRS), and endoscopic ultrasound (EUS) have been utilized for the localization and staging of patients with ZES and to detect recurrent disease. NETs are hypervascular and therefore demonstrate a greater degree of enhancement than the normal pancreas [1]. Cross-sectional imaging with CT and MRI does not possess the resolution to detect duodenal primary tumors small but can detect pancreatic gastrinomas with a sensitivity of 54–71 % and metastatic regional lymph nodes [51, 63, 64] (Fig. 13.2). Somatostatin receptors are expressed by 85–100 % of gastrinomas, and for this reason, SRS using radiolabeled octreotide can be used to detect primary gastrinomas and nodal or distant metastatic lesions [63] (Fig. 13.3). SRS has a sensitivity of 11–30 % for the identification of duodenal lesions, 25–77 % for pancreatic lesions, 72–82 % for nodal metastases, and 67–100 % for hepatic metastases from gastrinoma [1, 63, 65]. In a study of 122 patients with ZES by Termanini et al., SRS was positive for disease in 61 % of patients and was the only imaging modality that localized tumor in 12 % of cases, changing the management of 47 % of cases [66]. In another study, SRS was able to detect primary tumors and metastatic lesions with a sensitivity superior to that of CT, MRI, and angiography combined; however, it was unable to identify 33 % of the primary tumors subsequently found at laparotomy [63]. These results correlated with the tumor size and location as SRS detected only 30 % of gastrinoma <1 cm, 64 % of those between 1.1 and 2 cm, and 96 % of those >2 cm and missed small duodenal lesions.

EUS has been shown to be useful in the detection of pancreatic neuroendocrine tumors including gastrinomas, identifying even small lesions with a sensitivity of 82 % and a specificity of 95 % [67]. This technique also has the advantage of defining the relationship of primary tumors to pancreatic ductal and vascular structures, and needle biopsy of pancreatic lesions can also be performed. EUS was able to detect up to 91 % of pNETs missed by CT [67]. EUS unfortunately is less sensitive in the detection of duodenal gastrinomas due to their small size.

Invasive testing involving arteriography or selective venous sampling has been utilized to assist in the localization of gastrinoma in patients where CT, MRI, and SRS do not localize the tumor. Selective angiography with secretin injection (SASI) involves cannulation of the right or left hepatic veins and sampling of gastrin levels from this location after intra-arterial injection of secretin or more commonly calcium into the gastroduodenal, hepatic, splenic, left gastric, and superior mesenteric arteries [68]. Secretin stimulates gastrin release from the tumor present in the vascular distribution of the arterial injection, producing an increase in gastrin levels as measured in the hepatic veins. Utilizing the same principles, intravenous calcium injection has also proven to be effective in localizing the gastrinoma [69]. The proportion of patients with gastrinoma where the tumor can be regionalized by SASI varies from 71 to 89 % [1].

Whole body positron emission tomography (PET) using 18-Fluorodeoxyglucose (FDG) is a powerful modality for oncological imaging, yet unfortunately does not accumulate in most well-differentiated NET including gastrinomas. 11C-5-hydroxytryptophan has been shown to be superior to CT and SRS in the detection of neuroendocrine primary tumors or their metastases [70]. PET using 68-gallium-labeled DOTA-TOC, a somatostatin analogue with a greater affinity for somatostatin receptors than octreotide, has been associated with a sensitivity of 97 % for the detection of NETs and metastases [71, 72]. In one study of 52 NET patients, 68-gallium DOTA-TOC proved to be superior to CT and/or MRI for detection and staging. This modality impacted the therapeutic decision making in almost all patients [72].

In patients with negative imaging, an experienced surgeon becomes an invaluable localizing tool. In a recent report, Norton et al. were able to find the gastrinoma in 57/58 patients with negative imaging. These tumors were small (mean 0.8 cm) and found in the duodenum 67 %, pancreas 17 %, and primary lymph nodes (10 %) [40]. Immediate biochemical cure was established in 35/58 (60 %), with long-term biochemical cure (mean 9.4 years) in 27 (46 %) patients.

With the introduction of effective medical therapy for the control of gastric acid hypersecretion in patients with ZES, the role of surgery in the management of patients with gastrinoma has shifted from the control of gastric acid secretion to the requirement for oncological resection of the primary tumor to prevent disease progression and metastatic spread [30, 43, 73]. Surgical resection is the only modality capable of achieving cure in patients with ZES. Given the malignant potential of the primary tumor, surgery also has an important role in preventing the progression to metastatic disease. In their study comparing the long-term follow-up of 160 ZES patients undergoing surgical resection of gastrinoma with 35 ZES patients managed nonoperatively, Norton et al. found that significantly fewer patients undergoing surgical resection developed hepatic metastases during long-term follow-up (5 % vs. 29 %) [36]. More importantly this group demonstrated that surgery can increase survival. The 15-year disease-specific survival was 98 % in the surgery group versus 74 % in the nonoperative group. In those patients who did not undergo surgery, 90 % of deaths were secondary to progressive metastatic disease. Similarly at Ohio State University, Ellison et al. found that patients undergoing R0 resections had a significantly greater long-term survival than those who had incomplete resection of disease or who did not undergo surgery (10-year survival 85 % vs. 30 %) [4].

Surgical resection of the primary tumor and involved regional lymph nodes is curative in patients with gastrinomas that have not metastasized to distant sites. Inspection and palpation of the structures in the gastrinoma triangle is performed after kocherization of the duodenum and can successfully detect 61–71 % of primary gastrinomas [20, 38, 39]. Intraoperative ultrasound (IOS) has been shown to improve the ability to localize primary tumors of the pancreas and identify metastatic disease within the liver, but is not more sensitive than palpation in the detection of duodenal primaries [17, 39]. Small primary tumors within the duodenum are best identified by inspection and palpation of the mucosal surface via a longitudinal duodenotomy of the second part of the duodenum and inversion of the mucosa [19, 20]. After introducing routine duodenotomy to the operative exploration of patients with ZES, Norton et al. demonstrated a significant improvement in the ability to localize the primary tumors from 76 to 98 % of cases [17]. Routine duodenotomy also had significantly higher rates of long-term biochemical cure postoperatively occurring in 52 % of patients undergoing duodenotomy versus 26 % of patients who did not.

Options for resection of the primary gastrinoma are dependent on the location and size of the tumor (Table 13.3). The majority of duodenal lesions are small and can be resected locally, either removing from the submucosa or with an elliptical resection of the full thickness of the duodenal wall (Fig. 13.1). Lesions of the pancreas can be enucleated or formally resected depending on their size and location within the gland. Lesions within the head of the pancreas smaller than 2 cm can be enucleated in most cases. Pancreaticoduodenectomy is required for large or locally invasive tumors of the head of the pancreas, whereas a distal pancreatectomy is the procedure of choice for the rare gastrinoma located in the body or tail of the pancreas [51, 74] (Fig. 13.4a, b). Spread to regional lymph nodes can be demonstrated in 36–67 % of patients with gastrinoma, and therefore resection of regional lymph node groups should occur at the time of resection of the primary tumor [43, 52, 54]. A recent retrospective study in 41 ZES patients found higher rates of biochemical cure and a prolonged disease-specific survival in those ZES patients undergoing systemic lymphadenectomy compared to those who did not [75]. The authors concluded that in sporadic gastrinoma, systemic lymphadenectomy during the initial surgery may reduce the risk of persistent disease and improve survival.