Gastric neuroendocrine neoplasms: analysis of a cohort of patients followed at the Brazilian National Cancer Institute

Introduction

Gastric neuroendocrine neoplasms (G-NENs) are rare, representing less than 1% of all gastric malignancies, and approximately 7% of all gastrointestinal (GI) tract neuroendocrine neoplasms (Nikou & Angelopoulos 2012, Gluckman & Metz 2019). The term G-NEN encompasses a heterogeneous group of neoplasm mostly derived from the enterochromaffin-like cells (ECL cells) of the gastric mucosa. Although an increase in incidence has been recently reported, most of this is due to small indolent well-differentiated tumors, diagnosed by the widespread use of upper gastrointestinal endoscopy (Sato et al. 2016).

G-NENs are classified as well-differentiated neuroendocrine tumors (G-NETs) – subdivided as low, intermediate or high grade based on Ki-67 proliferative index and neuroendocrine carcinomas (NECs), which harbors dismal prognosis. A G-NET subtype classification, based on pathogenesis and hormonal-related behavior, is commonly used in addition to the World Health Organization (WHO) pathology classification, as different subtypes show distinct clinical characteristics, pathophysiology, aggressiveness and prognosis (Burkitt & Pritchard 2006, Dias et al. 2017) (Table 1).

Type 1 is the most common and accounts for 70–80% of all G-NETs. It is characterized by high serum gastrin, clinically asymptomatic, usually presents as small and multiple lesions (≤10 mm) in the gastric body or fundus that rarely metastasize due to its nonaggressive behavior and good prognosis. Type 1 is often characterized by slow tumor growth and low risk of metastasis (less than 1%) (Panzuto et al. 2023). They typically occur in the setting of atrophic gastritis, with anti-intrinsic factor or anti-parietal cell autoantibodies, leading to macrocytic anemia, and sometimes posterior cord syndrome (Díez et al. 2013, Sato et al. 2016, Dias et al. 2017, Rinzivillo et al. 2022). Type 2 occurs when hypergastrinemia is associated with gastrinomas, sometimes associated with Zollinger–Ellison syndrome (ZES) and multiple endocrine neoplasia type 1 (MEN-1). Gastrinomas are usually smaller than 20 mm, located in the duodenum or pancreas, and its ECL cell hyperplasia corresponds to 5% of all G-NETs (Jordan et al. 2004, Díez et al. 2013, Panzuto et al. 2023). Type 3 G-NET develops sporadically, independent of gastrin secretion, with single lesions, generally greater than 10 mm, with more aggressive behavior and higher potential to generate metastases. Type 3 represents 15–25% of all G-NETs and has an unfavorable long-term survival (5-year survival rate 70%). Anemia, gastric outlet obstruction or weight loss may be present due to their larger size (Díez et al. 2013, Sato et al. 2016, Dias et al. 2017, Dillon 2020, Panzuto et al. 2023).

Ki-67 proliferation index is utilized to group the G-NETs into well-differentiated tumors, classified into G1 (Ki-67 <3%), G2 (Ki-67 (3–20%) and G3 (Ki-67 >20%), and poorly-differentiated tumors are defined as large- or small-cell neuroendocrine carcinomas with Ki-67 >20% (the majority having Ki-67 above 50%) (Ferrarotto et al. 2013, Sato et al. 2016). Endoscopic ultrasound (EUS) should be considered to evaluate lesions larger than 1 cm to determine depth of invasion and possible regional lymph node involvement. Depth of tumor invasion is related to worse prognosis and survival.

The ideal management of G-NET is still a matter of debate. Treatment of type 1 is centered on endoscopic removal of larger lesions, sampling and follow-up with endoscopic surveillance (Panzuto et al. 2023). However, patients with lesions larger than 10 mm or high Ki-67 labeling should be managed with snare polypectomy or endoscopic mucosal resection (EMR) (Exarchou et al. 2022a , b , Panzuto et al. 2023). Endoscopic full thickness resection (EFTR) with a lymph node sampling should be the treatment choice in patients with type 1 G-NET when EUS shows suspected lymph node metastasis or invasion of the muscularis propria. Antrectomy has been proposed, aiming to remove gastrin production by G cells. Some caveats of this unorthodox procedure are surgical complications and the possibility that some ECL cells have already acquired autonomy and no longer depend on gastrin stimulus for its growth (Dias et al. 2017, Vanoli et al. 2018, Exarchou et al. 2022a , b , Panzuto et al. 2023).

As G-NET type 2 can be associated with ZES and MEN-1, management involves localizing and treating the gastrinoma, which can be multiple. In these cases, if resection is not feasible, removal of the most prominent gastrinoma followed by surveillance is a viable option. For symptom management, somatostatin analogs should be considered (Norton 2005, Epelboym & Mazeh 2014). It is also important to screen the patients with MEN-1 for associated tumors on the pituitary and parathyroid glands. The sporadic gastrinoma are more often solitary, larger than 20 mm and located in the pancreas. When there is no evidence of metastatic spread, surgical removal is indicated (Dias et al. 2017, Gluckman & Metz 2019). Before the development of an effective acid suppression therapy, the greater morbidity and mortality of ZES were related to complications of peptic ulcer disease, such as upper GI bleeding. The development of proton pump inhibitors resulted in a significant reduction in these events and prevented need for gastrectomy (Quatrini et al. 2005).

Type 3 G-NETs are candidates for radical resection, with total or subtotal gastrectomy and lymphadenectomy, due to the often advanced stage at the time of diagnosis and metastatic potential, especially with grade 3 tumor (Ki-67>20%) or diameter above 20 mm (Exarchou et al. 2022a , b ). However, recent studies have demonstrated less invasive treatment options in type 3 G-NETs. Endoscopic mucosal resection appears to be more appropriate in those patients with tumor diameter less than or equal to 10 mm, classified as G1. It may also be an option in larger tumors, with a diameter smaller than 15 mm and Ki-67 <10%, when the risk of a surgical resection is high. In the remaining G1 patients, who do not meet these criteria, depth of invasion, tumor size and presence of lymph vascular invasion should be considered. EFTR could be safely indicated in these patients, measuring less than 20 mm, limited to the submucosa and without evidence of lymph vascular invasion (Min et al. 2018, Hirasawa et al. 2021, Panzuto et al. 2023). Less invasive treatment for type 3 G2 G-NET is still under debate. EFTR with lymph node sampling may be considered in those measuring between 10 and 20 mm after an individualized selection (Hirasawa et al. 2021). One study with 50 patients showed that type 3 G-NET smaller than 20 mm, invading up to submucosal layer, and no evidence of lymph vascular invasion could be successfully managed endoscopically (Kwon et al. 2013).

When surgical or endoscopic resection is not recommended due to several factors, such as advanced age, comorbidities, tumor location or disease recurrence requiring multiple endoscopic resections, the use of somatostatin analogs (SSA) is an option. It is associated with high response rate, 25–100%, but disease recurrence is frequent with therapy discontinuation (Strosberg & Kvols 2010, Rossi et al. 2020, Panzuto et al. 2023). Netazepide is a drug not available in Brazil that works through gastrin receptor inhibition and can be utilized in place of SSA. A complete response rate of 30% was observed on a phase II trial with 16 patients (Boyce et al. 2017). Similarly to SSA, drug discontinuation is associated with frequent relapses, so its clinical benefit should be better evaluated before recommendation for clinical practice.

In metastatic disease, treatment modalities should also be individualized and carried out through a multidisciplinary team approach. For patients with metastatic and potentially resectable disease, surgery should be considered, as it may improve control of symptoms and delay tumor growth. However, even with complete resection, most patients may experience disease recurrence (Riihimäki et al. 2016, Gluckman & Metz 2019).

A large proportion of patients with neuroendocrine tumors of the GI tract have liver metastasis. Therefore, in patients with unresectable and predominant hepatic disease, locoregional therapies should be considered, such as surgical cytoreduction, ablative therapy with embolization or radiofrequency ablation (Riihimäki et al. 2016, Machairas et al. 2021). Patients with disseminated and unresectable metastatic disease benefit from SSA as initial therapy for controlling symptoms and tumor growth. It is indicated for G-NET G1 or G2 with low Ki-67 (<10%) and high expression of somatostatin receptors in nuclear imaging. Patients with elevated Ki-67 may benefit from SSA in case of slow tumor growth or small disease volume (Strosberg & Kvols 2010). Peptide receptor radiation therapy (PRRT) using a radiolabeled somatostatin analog (lutetium 177) may be a treatment option. But there is still no consensus on the ideal timing for this therapy (Strosberg et al. 2021).

Everolimus can be utilized when previous therapies are contraindicated or after progression (Yao et al. 2016). Patients with progressive metastatic disease and NET G3 are suitable for cytotoxic chemotherapy with capecitabine, dacarbazine, fluorouracil, oxaliplatin and temozolomide. The mostly prescribed chemotherapy regimens use capecitabine combined with oxaliplatin or temozolomide (Ferrarotto et al. 2013, Strosberg et al. 2015). Most current clinical recommendations regarding G-NET care are based on retrospective studies due to the scarcity of prospective randomized trials – which are difficult to perform for G-NET.

This study aims to describe a cohort of patients with G-NET followed at the Brazilian National Cancer Institute (INCA) during the past decades. We sought to highlight differences in presentation, treatment and survival outcomes among the three subtypes. The study adheres to the strengthening the reporting of observational studies in epidemiology (STROBE) statement (https://www.strobe-statement.org/). INCA is part of the Brazilian public health system (SUS) and acts as the public cancer center of reference for all cancer types, including rare diseases. INCA also plays a role in formulating cancer-related recommendations and policies through the Brazilian Ministry of Health.

Materials and methods

Results

Population characteristics

From July 2000 to October 2022, 116 patients with G-NEN were identified. No case of misdiagnosed G-NET was identified – Fig. 1. The median age at diagnosis was 55 years (17–87 years), with 66 (56.9%) female and 50 (43.1%) male. No significant difference was observed in terms of age between groups (Table 2).

View Full Size

Study’s included cases flowchart.

Citation: Endocrine Oncology 5, 1; 10.1530/EO-24-0063

• Download Figure
• Download figure as PowerPoint slide

Table 2

Population characteristics according to pathology classification.

	All patients (n = 116)	G1 (n = 51)	G2 (n = 31)	G3 (n = 4)	NEC (n = 11)	P
Age (years) (median)	55 (17–87)	56 (17–87)	49 (32–72)	66 (60–74)	69 (39–84)
Gender	116	51	31	3	12
Female	66 (56.9%)	33 (64.7%)	20 (64.5%)	1 (33.3%)	5 (41.7%)	0.290
Male	50 (43.1%)	18 (35.3%)	11 (35.5%)	2 (66.7%)	7 (58.3%)
Tumor size (mm) (median)
Endoscopy	6.0 (1–80)	5 (3–68)	6.0 (3–80)	31.5 (3–60)	40 (20–50)
Pathology	7.5 (1–80)	7 (2–75)	9.5 (3–22)		65 (60–80)
Tumor site	114	50	31	3	12
Antrum	6 (5.3%)	3 (6%)	0	1 (33.33%)	2 (16.7%)	0.001
Cardia	6 (5.3%)	2 (4%)	0	1 (33.3%)	3 (25%)
Body	74 (64.9%)	34 (68%)	22 (71%)	1 (33.3%)	5 (41.6%)
Fundus	8 (7%)	2 (4%)	1 (3.2%)	0	2 (16.7%)
Multifocal	20 (17.5%)	9 (18%)	8 (25.8%)	0	0
Clinical classification (n)	76	42	20
Type 1	60 (78.9%)	32 (76.2%)	16 (80%)	−	−	0.525
Type 2	5 (6.6%)	2 (4.8%)	2 (10%)	−	−
Type 3	11 (14.5%)	8 (19%)	2 (10%)	−	−
Initial symptoms (n)	116	51	31	3	12
Abdominal pain	40 (34.5%)	18 (35.3%)	11 (35.5%)	1 (33.3%)	3 (25%)	0.014
Asymptomatic	32 (27.5%)	19 (37.2%)	8 (25.8%)	0	0
Anemia	18 (15.5%)	8 (15.7%)	4 (12.9%)	1 (33.3%)	0
Weight loss	17 (14.7%)	2 (3.9%)	6 (19.4%)	1 (33.3%)	7 (58.4%)
GI bleeding	8 (6.9%)	4 (7.8%)	2 (6.5%)	0	1 (8.3%)
GI obstruction	1 (0.9%)	0	0	0	1 (8.3%)

GI, gastrointestinal. Bold indicates statistical significance.

A second malignancy was present in 22 patients and concomitant gastric adenocarcinoma was the most common, observed in five (4.3%) cases.

Regarding presentation at diagnosis, 40 (34.5%) patients had abdominal pain, 18 (15.5%) had anemia and 17 (14.7%) had weight loss as initial symptoms. Thirty-two (27.5%) were asymptomatic and had G-NETs incidentally diagnosed during an upper GI endoscopy routine examination (Table 2). Eight patients presented with upper GI bleeding (6.9%). In four of them, the bleeding was tumor-related and the G-NET width ranged from 20 to 75 mm; three of them had actinic endoscopic findings due to radiation therapy directed to a previously diagnosed malignancy (mainly gastric adenocarcinoma); one patient was diagnosed with esophageal varices due to chronic liver disease. Finally, one patient presented with initial GI obstruction (0.9%) due to a 65 mm pyloric G-NET. This patient was referred to an emergency surgery with subtotal gastrectomy and lymphadenectomy.

Presentation according to clinical classification was not different among the three groups, as described in Table 3.

Table 3

Initial symptoms according to clinical classification.

Initial symptoms (n)	Type 1 60	Type 2 5	Type 3 11	P
Abdominal pain	20 (33.3%)	3 (60%)	7 (63.6%)	0.23
Asymptomatic	22 (36.7%)	1 (20%)	3 (27.3%)
Anemia	15 (25%)	−	−
Weight loss	2 (3.3%)	1 (20%)	−
GI bleeding	1 (1.7%)	−	1 (9.1%)
GI obstruction	−	−	−

GI, gastrointestinal.

Gastrin was measured in 76 patients (65.5%), of which 65 (85.5%) had levels above the normal range. Similarly, vitamin B12 was measured in 70 patients (60.3%), with levels below normal in 29 (41.4%). Therefore, G-NET clinical classification was possible in 76 patients (65.5%). Among the 76 cases properly classified by subtype, G-NET type 1 was the most frequent with 60 cases (78.9%), followed by G-NET type 3 with eleven cases (14.5%) and G-NET type 2 with only five cases (6.6%) – Fig. 1.

Survival

The median follow-up for the entire cohort was 69.7 (0–243.8) months. Thirty-three (28.4%) patients died after a median survival of 18.8 (0.1–243) months. G-NET was the cause of death in 14 cases (42.4%). Among the 14 deceased, eight (57.1%) cases were classified as NEC, four (28.6%) as NET-G2 and two (14.3%) as NET-G3.

Death related to a second malignancy occurred in seven cases (21.2%) and from unknown cause in 12 patients (36.4%). The median overall survival according to the pathological classification were 84.5 months for G1, 73.4 months for G2, 17.4 months for G3 and 6.2 months for NEC (Fig. 2).

View Full Size

The median overall survival according to pathology classification.

Citation: Endocrine Oncology 5, 1; 10.1530/EO-24-0063

• Download Figure
• Download figure as PowerPoint slide

Discussion

To our knowledge, the present study reports the largest G-NET cohort to this date from a single Brazilian cancer center. Given the G-NEN rarity and the need of real-life reports from reference cancer centers (specially from low/middle-income countries), our study adds to the literature as we believe that the presented data might be of use due to the long follow-up observed. We believe that these data may serve as a benchmark for future policies and building of evidence in the field of this rare disease. Our findings highlight the importance of having providers from multiple specialties working collaboratively to define the appropriate approach to treat G-NET. Multiple specialties working together in the same institution is particularly important in the Brazilian public health system, where many patients would not have the ability to manage multiple referrals and appointments at different facilities due to out-of-pocket expenses with food, transportation and procedures.

G-NET management relies on the pathological report of tumor differentiation, grade, extent of invasion and presence of poor prognostic features in addition to the subtype definition (Burkitt & Pritchard 2006, Sato et al. 2016, Dillon 2020, Exarchou et al. 2022a , b , Panzuto et al. 2023). Ideally, all these information should be available for patients with G-NET because they assist the healthcare team in defining the appropriate treatment and correlates with disease-specific survival (Ferrarotto et al. 2013, Sato et al. 2016). In our cohort, 34.5% of patients did not have subtype classification due to lack of serum gastrin evaluation, while 18.1% did not have Ki-67 index documented due to irregular public funding for specialized blood tests and immunohistochemistry. These challenges are often observed in limited resources settings, and lead to gaps in the availability of many tests and markers recommended as essential by international guidelines and treatment protocols. Our findings reinforce the need to create treatment protocols that take into account limited resources settings, such as many LMICs. Despite that, it is noteworthy that overall survival among our sample is similar to that described by cancer centers from high-income countries.

Among those patients with type 1 G-NET (8.3%) treated with surgical resection, one had high-risk criteria (endoscopy tumor size of 75 mm) and was treated with total gastrectomy and lymphadenectomy. The other three patients could have been treated with less invasive procedures as they had G1 G-NET smaller than 10 mm, confined to mucosa or submucosa. However, these disease characteristics were better defined after surgery with the surgical specimen evaluation. Interestingly, of the eleven patients classified as type 3 G-NET, the majority 63.6% (7/11) were treated with endoscopic resection. The tumors were confined to the submucosa, with sizes ranging from 5 to 10 mm; five patients had tumor grade G1 (with Ki-67 2%) and one had G2 (with Ki-67 5%). Three patients had more than one endoscopic resection and had stable disease at last follow-up visit, emphasizing the need for individualized patient evaluation and treatment planning.

Among the 14 patients (42.4%) who died due to the G-NET outcomes, nine patients had NEC, two had NET G3 and three had muscle layer invasion, reinforcing the importance of high risk features increasing the risk of death. Due to the prolonged follow-up, death due to unknown causes happened in 12 patients (36.4%). Seven patients (21.2%) died of second malignancy, highlighting the need for better practices of promotion of healthy habits and cancer screening among long-term survivors of cancer.

Our cohort results mirror many western studies outcomes. However, this study has limitations. Most of them are related to the study’s retrospective nature. The lack of essential information limits the assessment of tumor classification and its relation to clinical main outcomes. In addition, it was not possible to evaluate the underlying gastric pathology due to the lack of investigation or report in the medical records, and gastrin was measured in 76 patients (65.5%), therefore, G-NET clinical classification was possible only in 76 patients (65.5%).

In conclusion, G-NEN is a rare neoplasm that includes different subtypes with different treatments and prognosis. This study highlights the need for individualized treatment planning and multispecialty collaboration for G-NEN management. These practices can help to tailor the treatment according to the tumors’ behavior, utilizing less aggressive treatments when possible.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding

This work did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.

Author contribution statement

SC collected and interpreted data and wrote the manuscript. BPV performed data analysis and interpretation and manuscript review. CBC, RG and RR assisted with manuscript review. DB conceived the study, performed data collection, statistical analysis, assisted with interpretation of the results, wrote and edited the manuscript.