Type I Gastric NENs

Type I gastric NENs occur in patients with chronic atrophic gastritis (CAG), with hypergastrinemia caused by an absence of gastric acid. Lesions are usually located in atrophic oxyntic mucosa in the fundus in individuals with CAG with or without pernicious anemia.

Most instances (70%–80%) are seen in patients with CAG and associated hypergastrinemia, whereas pernicious anemia is common (58%). Among 367 individuals with atrophic gastritis, the prevalence was 2.4%, with an annual incidence of 0.4% during endoscopic follow-up. In general, type I NENs occur more frequently in women, and 70% to 80% are diagnosed between the fifth and seventh decades. Because of diminished acid secretion, serum gastrin levels are significantly elevated in patients with atrophic gastritis. On the same basis, a correlation between long-term proton pump inhibitor (PPI)–induced hypergastrinemia and the development of gastric NENs is possible and supported by recent epidemiologic data and case reports. Serum gastrin and chromogranin A (CgA) levels can be elevated in approximately 100% and 95% of patients, respectively.

Lesions, when identified, are usually small (<1 cm), polypoid, and multicentric (≈67% of cases). These tumors are mostly limited to the mucosa or submucosa, do not exhibit angioinvasion, and seem to be benign in behavior. Larger tumors (1–2 cm) may exhibit low-grade malignant behavior, with or without angioinvasion. Tumors in this group may be either single or multiple, exhibit a low rate of lymph node invasion (3%–8%), and are rarely (≈2% of cases) associated with distant metastases. In general, irrespective of their size, lesions are classified as stage I (87%).

At a histologic level, ECL cell lesions have been classified as pseudohyperplasia (cell clustering unassociated with cell proliferation), hyperplasia (diffuse, linear, micronodular, adenomatoid), dysplasia (enlarged, adenomatous or fused micronodules, microinfiltration, nodular growth), and neoplasia (intramucosal or invasive carcinoids). The entire spectrum of ECL cell proliferation, from hyperplasia to dysplasia and neoplasia, has been observed in type I tumors. Hyperplastic and pseudohyperplastic ECL alterations may also occur with some frequency in Helicobacter pylori– related chronic gastritis associated with ulcer disease or dyspepsia. In general, ECL cell type I lesions tend to be well differentiated and are considered by World Health Organization (WHO) classification to be of benign or low-grade malignant potential. They are usually classified (>80%) as G1 ( Table 1 ), with a Ki-67 (if detected) of less than 1%.

At a molecular level, although germ-line mutations in the MEN-1 locus (located on 11q13) could not be identified, a frequent loss of heterozygosity (LOH) within the 11q13–14 region was noted. This finding suggests a potential involvement for the MEN-1 gene and/or a more telomeric tumor suppressor gene in the pathogenesis of type I NENs. More recently, loss of methylation of long interspersed nucleotide element type 1 ( LINE1 ) that has been identified in well-differentiated pancreatic NENs has been examined in gastric NENs. This investigation is based on the premise that LINE1 hypomethylation may serve as a marker of tumor grade and lymph node metastasis. In 11 gastric type I tumors, methylation levels were not different from normal mucosa and no difference was established for higher-grade lesions, suggesting this may not be a potential marker in these lesions. The enzyme α-methylacyl-coenzyme A racemase, which plays a role in the β-oxidation of branched-chain fatty acid and its derivatives and is used to detect prostatic adenocarcinoma, has recently been examined in gastric NENs. None of the grade 1 lesions were immunoreactive, whereas 67% of grade 2 were positive, suggesting that this may be a useful marker for distinguishing between grade 1 (negative) and 2 gastric NENs. The relatively high proportion of grade 1 (≈90%) lesions in type I NENs limits using this enzyme as a marker, but intriguingly, given the role this enzyme plays in metabolism, differential expression could suggest metabolic differences between grade 1 and 2 lesions, suggesting a potential utility as a biomarker.

Treatment options include endoscopic and surgical resection and pharmaceutical intervention. Endoscopic resection may be appropriate for intraepithelial tumors less than 2 cm and occasionally tumors less than 1 cm invading into the lamina propria or submucosa. In one study that included 16 type I NENs, complete resection was achieved in all lesions. No procedure-related perforations and only 1 local recurrence occurred during a mean follow-up of 24.4 months. In a separate study, approximately 60% of type I lesions had recurred after endoscopic resection; endoscopic management nevertheless is usually considered safe and effective. Endoscopic resection or local wedge excision may also be repeated unless the lesions are excessive in number or evidence of invasion is present. Large lesions greater than 20 mm that ulcerate or bleed may require more extensive surgical resection, particularly if the patient is young and evidence shows diffuse gastric microcarcinoidosis. In general, surgery may be effective in approximately 80% of tumors. In selected patients at risk for metastatic disease, antrectomy to reduce gastrin levels may be considered.

Given their slow proliferation, type I NENs are not amenable to standard chemotherapeutic approaches. Instead, investigators have focused on disrupting ECL cell growth through either targeting somatostatin receptors or inhibiting the gastrin stimulus. Although somatostatin scintigraphy may be negative for small, localized tumors, receptors are readily evident on both ECL cells and ECL cell tumors, and somatostatin analogs inhibit ECL cell secretion and proliferation. Somatostatin analogs also reduce gastrin levels, and thus gastrin’s proliferative stimuli on ECL cells. Therapeutic use of somatostatin analog therapy is associated with regression of lesions and reductions in circulating gastrin and CgA, the number of visible tumors, and CgA-positive immunoreactive tumor cells. These effects, however, are generally short term (≈1 year), and disease progression and even tumor dedifferentiation have been noted at 5 years after termination of therapy. Treatment of benignly behaving localized type I tumors with somatostatin analogs remains controversial and, if this treatment is initiated, it should not be discontinued because of a potential risk of progression to malignant disease.

The specific gastrin receptor (CCK2) antagonist YF476 is an alternative agent that has been shown to inhibit acid secretion and ECL cell proliferation, and reduce type I NEN lesions size and number. The clinical utility of this approach was recently examined in a prospective study. Eight patients with multiple type I tumors received oral netazepide (YF476) once daily for 12 weeks, with follow-up at 12 weeks in an open-label pilot trial. All patients had a reduction in the number and size of their largest tumor, and plasma CgA was reduced to normal levels at 3 weeks, whereas gastrin levels remained unchanged. This agent seemed to be well tolerated; its utility requires further assessment because CgA levels rebounded after drug withdrawal. A vaccine against gastrin (G17 molecule) was recently tested in 3 patients with CAG/type I, who were followed up endoscopically and clinically for a mean of 36 months. Tumor regression was noted in 2 of the 3 patients who exhibited a significant reduction in plasma CgA levels. These pilot studies that are based on an understanding of the underlying biology of the lesion, suggest that targeting gastrin either at a receptor level or the hormone itself may be a potential therapy for type I lesions.

However, given the low mortality rate associated with these lesions, whether these pharmacologic approaches will be of utility remains unclear. Studies by Rappel and colleagues in 1995 reported an observed 78% and an age-corrected 100% survival rate (Kaplan-Meier) in 88 patients with type I lesions. More recently, 5- and 10-year crude survival rates were estimated at 96.1% and 73.9%, respectively (not different from the general population), for type I tumors. These data reflect an overall benign course for this disease. In general, type I tumors have survival rates that are therefore not different from the general population. The benefit of treatment must therefore be considered in conjunction with the potential morbidity associated with pharmaceutical treatment, repeated endoscopic resection, or surgery. Nevertheless, an active treatment approach is required for the more malignant tumors.

Type II Gastric NENs

These ZES-MEN-1–associated hypergastrinemia-driven lesions are usually multiple, small (<1 cm), and predominantly of the ECL cell type, although some lesions contain heterogeneous cell populations. Serum gastrin levels in type II NENs are usually significantly elevated. A secretin test or measurement of acid secretion levels may suggest the gastrinoma origin of the type II–associated lesion. However, clinical presentations may exhibit wide variations, and include diverse symptomatology indistinguishable from peptic ulcer, gastric polyps, or even carcinoma. The clinicopathologic behavior of type II NENs occupies an intermediate position between that of the aggressive (gastrin-independent) type III sporadic lesions and the more benign type I lesions.

Like type I lesions, type II lesions range from pseudohyperplasia, hyperplasia, dysplasia, and neoplasia. Type II lesions, like type I NENs, consist mainly of ECL cells, and exhibit argyrophil cell hyperplasia/dysplasia throughout the oxyntic mucosa, which, in contrast to CAG/A, is usually rugose. The cytologic characteristics of the lesion are also similar to the type I NENs. In contrast to the latter, and although they are generally considered well differentiated, local infiltration may occur at mucosal and submucosal levels and metastases occur in approximately 12% of lesions. They are nevertheless characterized as WHO type 1 tumors and are considered of benign or low-grade malignant potential.

Studies show that one-third of individuals with MEN-1 develop gastric type II lesions, and LOH has been identified in 75% of ZES-MEN-1 tumor samples and 41% of MEN-1 gastrinomas. The MEN-1 syndrome is an autosomal dominant disorder caused by mutations in the MEN-1 gene. The MEN-1 gene protein product menin is a tumor suppressor involved in transcriptional regulation and genome stability. Germline mutations and/or LOH are associated with disruption of menin function with a consequent loss of proliferative inhibition (perhaps through regulation of retinoblastoma expression, inactivation of cell cycle inhibitors such as P16 ^INK4A , or alterations in AKT/mTOR ) and development of lesions. The dysregulation of menin function during hypergastrinemia is unclear. Animal studies suggest that elevated gastrin is associated with decreased menin expression, whereas cell-based assays show that menin inhibits gastrin expression. Loss of menin may be associated with an elevation in gastrin transcription (secondary to elevated gastrin translation, synthesis, and secretion).

In addition to MEN-1–associated endocrine pathology, type II lesions commonly occur with a variety of other endocrine abnormalities, including hypothyroidism (39%), diabetes (19%), and Addison disease (6%). Although the molecular biology is generally understood to reflect mutations in MEN-1 , recent studies have identified 2 other loci that may be of interest: NF1 and CDNKB . The NF gene product, neurofibromin, is a tumor suppressor and, when nonfunctional, results in activated P21ras signaling and abnormal cell proliferation. An LOH at the NF1 locus has been demonstrated in a gastric NEN derived from a patient with NF1 suggesting that the RAS-MAPK pathway may be associated with the proliferative regulation. Progression through the cell cycle, particularly through G1, is negatively regulated by p27 ^KIP1 or CDKN1B. Recently, a patient with both hyperparathyroidism and a type II lesion was identified with a heterozygous GAGA deletion in the 5′-UTR of CDKN1B ( NM_004064.3 :c.-32_-29del). This germline mutation falls inside the region that is responsible for CDKN1B transcription and is predicted to destroy a secondary stem and loop structure that includes the GAGAGA element responsible for ribosome recruitment. This function suggests that alterations in the transcription/translation rate of CDKN1B mRNA may be the mechanism for tumor susceptibility. These results indicate that the mutational spectrum in gastric lesions is greater than MEN-1 , and investigators could potentially consider other gene targets that are known candidates for molecular-based therapies in other adenocarcinomas (eg, MAPK inhibitors).

Although the metastatic rate (3%–12%) of ZES-MEN-1 (type II) lesions is reported to be higher than for type I lesions, the short-term prognosis is similar. The long-term prognosis of type II disease, however, ultimately reflects the course of the MEN-1 gastrinoma, which itself has a 5-year survival rate of 60% to 75%. Overall, the survival rate of type II lesions is therefore worse than that of type I given the increased propensity for metastasis of the gastric lesion and the additional morbidity conferred by the natural history of MEN-1.

As with type I NENs, treatment options include endoscopic and surgical resection and pharmaceutical intervention. Removal of the source of hypergastrinemia is the critical aim of surgery, and regression of type II lesions may occur after successful gastrinoma excision. Endoscopic resection has also been used for gastric type II NENs, and endoscopic surveillance with extensive sampling of both the lesser and greater curvatures is recommended if the gastrinoma is inoperable. More extensive surgery may be necessary for lesions greater than 20 mm and when an excessive number of lesions is evident or if deeper invasion into the muscular wall is present.

High doses of a PPI is the preferred treatment for patients with symptoms of hyperacidity. The long-acting depot formulations of somatostatin analogs can also be prescribed in the presence of symptoms, and have been used to control hypergastrinemia and ulceration. In addition, these analogs may decrease the gastric endocrine cell mass, but evidence of antiproliferative efficacy is insufficient.

Type III Gastric NENs

These tumors are usually large and evolve in a milieu of seemingly normal plasma gastrin levels within a “normal” gastric mucosa. Overall, type III NENs exhibit a greater similarity to neuroendocrine carcinomas than to NENs per se. In this respect, their biologic behavior is aggressive, and local invasive growth and distant metastases are predictable features of their evolution.

Tumors display a fairly uniform light microscopic appearance with typical carcinoid histopathologic features. Their growth pattern may be trabecular or gyriform, medullary or solid, glandular or rosette-like, or a combination of any of these types. Lesions exhibit a spectrum from well differentiated to poorly differentiated, often demonstrate an aggressive local behavior (at diagnosis, local spread is present in approximately 15% of patients), and have a high incidence (24%–55%) of metastasis.

WHO type 2 tumors include all lesions, including the traditionally annotated type III NENs, and are characterized by being greater than 2 cm in size, have invaded the muscularis propria with or without metastases and are either nonfunctional or secretory tumors. WHO type 3 tumors include the traditionally annotated “atypical” or “sporadic” gastric NENs. This subtype, which is particularly aggressive in its behavior, is associated with rapid local gastric progression, distant metastases, and early death. These tumors may be identified by an increased mitotic count with nuclear polymorphism, hypochromasia, and prominent nucleoli. Whether this tumor is any different from a neuroendocrine carcinoma is unclear.

LOH in the 11q13-14 region is infrequently (<25%) found in type III lesions, suggesting that alterations in the MEN locus do not play a significant role in the etiopathogenesis of this gastrin-autonomous tumor type; the molecular pathogenesis of these lesions is currently unknown.

Type III lesions often display markedly aggressive local behavior, and metastasize. This poor prognosis is directly related to the high metastatic rate, and the overall 5-year survival rate is usually less than 50%. The 5-year survival rate is significantly higher for localized disease (64.3%) and for lesions with regional metastases (29.9%) than for lesions with distant metastases (10%). Although the tendency to metastasize correlates with tumor size, minute tumors have been reported with spread. Factors that predict aggressive behavior include cellular atypia, 2 or more mitoses per 10 high-powered fields, Ki-67 index greater than 2%, angioinvasion, and transmural invasion.

These normogastrinemia-associated tumors occur more often in men older than 50 years. They typically present as gastric adenocarcinomas with loss of appetite and weight, anemia, and evidence of local and advanced metastatic disease. The atypical carcinoid syndrome may also be evident in some type III lesions and is characterized by cutaneous flushing, profound itching, bronchospasm, and lacrimation. This finding usually reflects the sequelae of unregulated ECL cell histamine production (though other agents, such as tachykinins, have been reported) and is often provoked by ingestion of certain kinds of foods, particularly cheeses (tyramine) and wine.

Histamine 1 and 2 receptor blockade may be of benefit in suppressing skin rashes and acid hypersecretion. Whenever possible, surgical resection and lymph node dissection should be performed. If radical resection is not possible, preoperative downstaging with chemotherapy may be beneficial. Intravenous cytotoxic therapy may be used if metastases are evident. In selected patients, hepatic artery perfusion, focal embolization (or chemoembolization), cryoablation, or radiofrequency ablation with or without synchronous chemotherapy may be beneficial, whereas the combination of cisplatin and etoposide should be considered for anaplastic neuroendocrine carcinomas. Peptide receptor radionuclide therapy (PRRT) may be considered as a treatment option, but a positive somatostatin scintigraph is a prerequisite for using this modality. Neither cytotoxics nor PRRT are associated with substantial objective tumor responses (≈20%) but may increase progression-free survival.

Surgically, these lesions should be managed as for gastric adenocarcinoma. Thus, the presence of solitary, large (>1 cm), or invasive tumors mandates an attempt at surgical cure. In patients whose general condition is consistent with an acceptable operative risk, complete or partial gastrectomy (en bloc resection) with local lymph node resection is appropriate. Lesions greater than 2 cm associated with local invasion require subtotal gastrectomy or extended local resection.

These lesions often display markedly aggressive local behavior. The overall 5-year survival rate is usually less than 50%, and a poor prognosis is related to metastases. Lesions with regional (29.9%) and distant metastases (10.0%) have significantly lower survivals.

Type I Gastric NENs

Type I gastric NENs occur in patients with chronic atrophic gastritis (CAG), with hypergastrinemia caused by an absence of gastric acid. Lesions are usually located in atrophic oxyntic mucosa in the fundus in individuals with CAG with or without pernicious anemia.

Most instances (70%–80%) are seen in patients with CAG and associated hypergastrinemia, whereas pernicious anemia is common (58%). Among 367 individuals with atrophic gastritis, the prevalence was 2.4%, with an annual incidence of 0.4% during endoscopic follow-up. In general, type I NENs occur more frequently in women, and 70% to 80% are diagnosed between the fifth and seventh decades. Because of diminished acid secretion, serum gastrin levels are significantly elevated in patients with atrophic gastritis. On the same basis, a correlation between long-term proton pump inhibitor (PPI)–induced hypergastrinemia and the development of gastric NENs is possible and supported by recent epidemiologic data and case reports. Serum gastrin and chromogranin A (CgA) levels can be elevated in approximately 100% and 95% of patients, respectively.

Lesions, when identified, are usually small (<1 cm), polypoid, and multicentric (≈67% of cases). These tumors are mostly limited to the mucosa or submucosa, do not exhibit angioinvasion, and seem to be benign in behavior. Larger tumors (1–2 cm) may exhibit low-grade malignant behavior, with or without angioinvasion. Tumors in this group may be either single or multiple, exhibit a low rate of lymph node invasion (3%–8%), and are rarely (≈2% of cases) associated with distant metastases. In general, irrespective of their size, lesions are classified as stage I (87%).

At a histologic level, ECL cell lesions have been classified as pseudohyperplasia (cell clustering unassociated with cell proliferation), hyperplasia (diffuse, linear, micronodular, adenomatoid), dysplasia (enlarged, adenomatous or fused micronodules, microinfiltration, nodular growth), and neoplasia (intramucosal or invasive carcinoids). The entire spectrum of ECL cell proliferation, from hyperplasia to dysplasia and neoplasia, has been observed in type I tumors. Hyperplastic and pseudohyperplastic ECL alterations may also occur with some frequency in Helicobacter pylori– related chronic gastritis associated with ulcer disease or dyspepsia. In general, ECL cell type I lesions tend to be well differentiated and are considered by World Health Organization (WHO) classification to be of benign or low-grade malignant potential. They are usually classified (>80%) as G1 ( Table 1 ), with a Ki-67 (if detected) of less than 1%.

At a molecular level, although germ-line mutations in the MEN-1 locus (located on 11q13) could not be identified, a frequent loss of heterozygosity (LOH) within the 11q13–14 region was noted. This finding suggests a potential involvement for the MEN-1 gene and/or a more telomeric tumor suppressor gene in the pathogenesis of type I NENs. More recently, loss of methylation of long interspersed nucleotide element type 1 ( LINE1 ) that has been identified in well-differentiated pancreatic NENs has been examined in gastric NENs. This investigation is based on the premise that LINE1 hypomethylation may serve as a marker of tumor grade and lymph node metastasis. In 11 gastric type I tumors, methylation levels were not different from normal mucosa and no difference was established for higher-grade lesions, suggesting this may not be a potential marker in these lesions. The enzyme α-methylacyl-coenzyme A racemase, which plays a role in the β-oxidation of branched-chain fatty acid and its derivatives and is used to detect prostatic adenocarcinoma, has recently been examined in gastric NENs. None of the grade 1 lesions were immunoreactive, whereas 67% of grade 2 were positive, suggesting that this may be a useful marker for distinguishing between grade 1 (negative) and 2 gastric NENs. The relatively high proportion of grade 1 (≈90%) lesions in type I NENs limits using this enzyme as a marker, but intriguingly, given the role this enzyme plays in metabolism, differential expression could suggest metabolic differences between grade 1 and 2 lesions, suggesting a potential utility as a biomarker.

Treatment options include endoscopic and surgical resection and pharmaceutical intervention. Endoscopic resection may be appropriate for intraepithelial tumors less than 2 cm and occasionally tumors less than 1 cm invading into the lamina propria or submucosa. In one study that included 16 type I NENs, complete resection was achieved in all lesions. No procedure-related perforations and only 1 local recurrence occurred during a mean follow-up of 24.4 months. In a separate study, approximately 60% of type I lesions had recurred after endoscopic resection; endoscopic management nevertheless is usually considered safe and effective. Endoscopic resection or local wedge excision may also be repeated unless the lesions are excessive in number or evidence of invasion is present. Large lesions greater than 20 mm that ulcerate or bleed may require more extensive surgical resection, particularly if the patient is young and evidence shows diffuse gastric microcarcinoidosis. In general, surgery may be effective in approximately 80% of tumors. In selected patients at risk for metastatic disease, antrectomy to reduce gastrin levels may be considered.

Given their slow proliferation, type I NENs are not amenable to standard chemotherapeutic approaches. Instead, investigators have focused on disrupting ECL cell growth through either targeting somatostatin receptors or inhibiting the gastrin stimulus. Although somatostatin scintigraphy may be negative for small, localized tumors, receptors are readily evident on both ECL cells and ECL cell tumors, and somatostatin analogs inhibit ECL cell secretion and proliferation. Somatostatin analogs also reduce gastrin levels, and thus gastrin’s proliferative stimuli on ECL cells. Therapeutic use of somatostatin analog therapy is associated with regression of lesions and reductions in circulating gastrin and CgA, the number of visible tumors, and CgA-positive immunoreactive tumor cells. These effects, however, are generally short term (≈1 year), and disease progression and even tumor dedifferentiation have been noted at 5 years after termination of therapy. Treatment of benignly behaving localized type I tumors with somatostatin analogs remains controversial and, if this treatment is initiated, it should not be discontinued because of a potential risk of progression to malignant disease.

The specific gastrin receptor (CCK2) antagonist YF476 is an alternative agent that has been shown to inhibit acid secretion and ECL cell proliferation, and reduce type I NEN lesions size and number. The clinical utility of this approach was recently examined in a prospective study. Eight patients with multiple type I tumors received oral netazepide (YF476) once daily for 12 weeks, with follow-up at 12 weeks in an open-label pilot trial. All patients had a reduction in the number and size of their largest tumor, and plasma CgA was reduced to normal levels at 3 weeks, whereas gastrin levels remained unchanged. This agent seemed to be well tolerated; its utility requires further assessment because CgA levels rebounded after drug withdrawal. A vaccine against gastrin (G17 molecule) was recently tested in 3 patients with CAG/type I, who were followed up endoscopically and clinically for a mean of 36 months. Tumor regression was noted in 2 of the 3 patients who exhibited a significant reduction in plasma CgA levels. These pilot studies that are based on an understanding of the underlying biology of the lesion, suggest that targeting gastrin either at a receptor level or the hormone itself may be a potential therapy for type I lesions.

However, given the low mortality rate associated with these lesions, whether these pharmacologic approaches will be of utility remains unclear. Studies by Rappel and colleagues in 1995 reported an observed 78% and an age-corrected 100% survival rate (Kaplan-Meier) in 88 patients with type I lesions. More recently, 5- and 10-year crude survival rates were estimated at 96.1% and 73.9%, respectively (not different from the general population), for type I tumors. These data reflect an overall benign course for this disease. In general, type I tumors have survival rates that are therefore not different from the general population. The benefit of treatment must therefore be considered in conjunction with the potential morbidity associated with pharmaceutical treatment, repeated endoscopic resection, or surgery. Nevertheless, an active treatment approach is required for the more malignant tumors.

Type II Gastric NENs

These ZES-MEN-1–associated hypergastrinemia-driven lesions are usually multiple, small (<1 cm), and predominantly of the ECL cell type, although some lesions contain heterogeneous cell populations. Serum gastrin levels in type II NENs are usually significantly elevated. A secretin test or measurement of acid secretion levels may suggest the gastrinoma origin of the type II–associated lesion. However, clinical presentations may exhibit wide variations, and include diverse symptomatology indistinguishable from peptic ulcer, gastric polyps, or even carcinoma. The clinicopathologic behavior of type II NENs occupies an intermediate position between that of the aggressive (gastrin-independent) type III sporadic lesions and the more benign type I lesions.

Like type I lesions, type II lesions range from pseudohyperplasia, hyperplasia, dysplasia, and neoplasia. Type II lesions, like type I NENs, consist mainly of ECL cells, and exhibit argyrophil cell hyperplasia/dysplasia throughout the oxyntic mucosa, which, in contrast to CAG/A, is usually rugose. The cytologic characteristics of the lesion are also similar to the type I NENs. In contrast to the latter, and although they are generally considered well differentiated, local infiltration may occur at mucosal and submucosal levels and metastases occur in approximately 12% of lesions. They are nevertheless characterized as WHO type 1 tumors and are considered of benign or low-grade malignant potential.

Studies show that one-third of individuals with MEN-1 develop gastric type II lesions, and LOH has been identified in 75% of ZES-MEN-1 tumor samples and 41% of MEN-1 gastrinomas. The MEN-1 syndrome is an autosomal dominant disorder caused by mutations in the MEN-1 gene. The MEN-1 gene protein product menin is a tumor suppressor involved in transcriptional regulation and genome stability. Germline mutations and/or LOH are associated with disruption of menin function with a consequent loss of proliferative inhibition (perhaps through regulation of retinoblastoma expression, inactivation of cell cycle inhibitors such as P16 ^INK4A , or alterations in AKT/mTOR ) and development of lesions. The dysregulation of menin function during hypergastrinemia is unclear. Animal studies suggest that elevated gastrin is associated with decreased menin expression, whereas cell-based assays show that menin inhibits gastrin expression. Loss of menin may be associated with an elevation in gastrin transcription (secondary to elevated gastrin translation, synthesis, and secretion).

In addition to MEN-1–associated endocrine pathology, type II lesions commonly occur with a variety of other endocrine abnormalities, including hypothyroidism (39%), diabetes (19%), and Addison disease (6%). Although the molecular biology is generally understood to reflect mutations in MEN-1 , recent studies have identified 2 other loci that may be of interest: NF1 and CDNKB . The NF gene product, neurofibromin, is a tumor suppressor and, when nonfunctional, results in activated P21ras signaling and abnormal cell proliferation. An LOH at the NF1 locus has been demonstrated in a gastric NEN derived from a patient with NF1 suggesting that the RAS-MAPK pathway may be associated with the proliferative regulation. Progression through the cell cycle, particularly through G1, is negatively regulated by p27 ^KIP1 or CDKN1B. Recently, a patient with both hyperparathyroidism and a type II lesion was identified with a heterozygous GAGA deletion in the 5′-UTR of CDKN1B ( NM_004064.3 :c.-32_-29del). This germline mutation falls inside the region that is responsible for CDKN1B transcription and is predicted to destroy a secondary stem and loop structure that includes the GAGAGA element responsible for ribosome recruitment. This function suggests that alterations in the transcription/translation rate of CDKN1B mRNA may be the mechanism for tumor susceptibility. These results indicate that the mutational spectrum in gastric lesions is greater than MEN-1 , and investigators could potentially consider other gene targets that are known candidates for molecular-based therapies in other adenocarcinomas (eg, MAPK inhibitors).

Although the metastatic rate (3%–12%) of ZES-MEN-1 (type II) lesions is reported to be higher than for type I lesions, the short-term prognosis is similar. The long-term prognosis of type II disease, however, ultimately reflects the course of the MEN-1 gastrinoma, which itself has a 5-year survival rate of 60% to 75%. Overall, the survival rate of type II lesions is therefore worse than that of type I given the increased propensity for metastasis of the gastric lesion and the additional morbidity conferred by the natural history of MEN-1.

As with type I NENs, treatment options include endoscopic and surgical resection and pharmaceutical intervention. Removal of the source of hypergastrinemia is the critical aim of surgery, and regression of type II lesions may occur after successful gastrinoma excision. Endoscopic resection has also been used for gastric type II NENs, and endoscopic surveillance with extensive sampling of both the lesser and greater curvatures is recommended if the gastrinoma is inoperable. More extensive surgery may be necessary for lesions greater than 20 mm and when an excessive number of lesions is evident or if deeper invasion into the muscular wall is present.

High doses of a PPI is the preferred treatment for patients with symptoms of hyperacidity. The long-acting depot formulations of somatostatin analogs can also be prescribed in the presence of symptoms, and have been used to control hypergastrinemia and ulceration. In addition, these analogs may decrease the gastric endocrine cell mass, but evidence of antiproliferative efficacy is insufficient.

Type III Gastric NENs

These tumors are usually large and evolve in a milieu of seemingly normal plasma gastrin levels within a “normal” gastric mucosa. Overall, type III NENs exhibit a greater similarity to neuroendocrine carcinomas than to NENs per se. In this respect, their biologic behavior is aggressive, and local invasive growth and distant metastases are predictable features of their evolution.

Tumors display a fairly uniform light microscopic appearance with typical carcinoid histopathologic features. Their growth pattern may be trabecular or gyriform, medullary or solid, glandular or rosette-like, or a combination of any of these types. Lesions exhibit a spectrum from well differentiated to poorly differentiated, often demonstrate an aggressive local behavior (at diagnosis, local spread is present in approximately 15% of patients), and have a high incidence (24%–55%) of metastasis.

WHO type 2 tumors include all lesions, including the traditionally annotated type III NENs, and are characterized by being greater than 2 cm in size, have invaded the muscularis propria with or without metastases and are either nonfunctional or secretory tumors. WHO type 3 tumors include the traditionally annotated “atypical” or “sporadic” gastric NENs. This subtype, which is particularly aggressive in its behavior, is associated with rapid local gastric progression, distant metastases, and early death. These tumors may be identified by an increased mitotic count with nuclear polymorphism, hypochromasia, and prominent nucleoli. Whether this tumor is any different from a neuroendocrine carcinoma is unclear.

LOH in the 11q13-14 region is infrequently (<25%) found in type III lesions, suggesting that alterations in the MEN locus do not play a significant role in the etiopathogenesis of this gastrin-autonomous tumor type; the molecular pathogenesis of these lesions is currently unknown.

Type III lesions often display markedly aggressive local behavior, and metastasize. This poor prognosis is directly related to the high metastatic rate, and the overall 5-year survival rate is usually less than 50%. The 5-year survival rate is significantly higher for localized disease (64.3%) and for lesions with regional metastases (29.9%) than for lesions with distant metastases (10%). Although the tendency to metastasize correlates with tumor size, minute tumors have been reported with spread. Factors that predict aggressive behavior include cellular atypia, 2 or more mitoses per 10 high-powered fields, Ki-67 index greater than 2%, angioinvasion, and transmural invasion.

These normogastrinemia-associated tumors occur more often in men older than 50 years. They typically present as gastric adenocarcinomas with loss of appetite and weight, anemia, and evidence of local and advanced metastatic disease. The atypical carcinoid syndrome may also be evident in some type III lesions and is characterized by cutaneous flushing, profound itching, bronchospasm, and lacrimation. This finding usually reflects the sequelae of unregulated ECL cell histamine production (though other agents, such as tachykinins, have been reported) and is often provoked by ingestion of certain kinds of foods, particularly cheeses (tyramine) and wine.

Histamine 1 and 2 receptor blockade may be of benefit in suppressing skin rashes and acid hypersecretion. Whenever possible, surgical resection and lymph node dissection should be performed. If radical resection is not possible, preoperative downstaging with chemotherapy may be beneficial. Intravenous cytotoxic therapy may be used if metastases are evident. In selected patients, hepatic artery perfusion, focal embolization (or chemoembolization), cryoablation, or radiofrequency ablation with or without synchronous chemotherapy may be beneficial, whereas the combination of cisplatin and etoposide should be considered for anaplastic neuroendocrine carcinomas. Peptide receptor radionuclide therapy (PRRT) may be considered as a treatment option, but a positive somatostatin scintigraph is a prerequisite for using this modality. Neither cytotoxics nor PRRT are associated with substantial objective tumor responses (≈20%) but may increase progression-free survival.

Surgically, these lesions should be managed as for gastric adenocarcinoma. Thus, the presence of solitary, large (>1 cm), or invasive tumors mandates an attempt at surgical cure. In patients whose general condition is consistent with an acceptable operative risk, complete or partial gastrectomy (en bloc resection) with local lymph node resection is appropriate. Lesions greater than 2 cm associated with local invasion require subtotal gastrectomy or extended local resection.

These lesions often display markedly aggressive local behavior. The overall 5-year survival rate is usually less than 50%, and a poor prognosis is related to metastases. Lesions with regional (29.9%) and distant metastases (10.0%) have significantly lower survivals.