Advances in the Treatment of Gastroenteropancreatic Neuroendocrine Carcinomas: Are we Moving Forward?

Abstract

Poorly differentiated gastroenteropancreatic neuroendocrine carcinomas are aggressive neoplasms of challenging clinical management. A small proportion of patients with early-stage disease may achieve long-term survival, but the majority of patients present with rapidly lethal metastatic disease. Current standard of care still follows the treatment paradigm of small cell lung cancer, a far more common G3 neuroendocrine neoplasm, although emerging molecular and clinical data increasingly question this approach. In this article, we will briefly summarize epidemiology and prognosis of gastroenteropancreatic neuroendocrine carcinomas to emphasize the very low incidence, aggressive nature, and orphan status of this tumor entity. We will also discuss the current pathological classification and its limitations, as well as recent data on their differential biological background compared with small cell lung cancer, and its potential implications for patients care. Then, we will review the standard of care of systemic therapy, basically focused on platinum-based cytotoxic chemotherapy, including some recent randomized trials providing evidence regarding efficacy of irinotecan vs etoposide platinum doublets. Finally, we will present a comprehensive overview of novel therapeutic strategies in current clinical development, including recently reported data on immunotherapy, tumor-agnostic therapies (microsatellite instability, high tumor mutational burden, NTRK and RET gene fusions, BRAF or KRAS inhibitors), and additional treatment strategies targeting other tumor vulnerabilities (ie, Notch pathway, novel targets for radioligand therapy), and provide some insights regarding unmet needs and future perspectives to improve patient's care and prognosis.

Graphical Abstract

Graphical Abstract.

Essential Points.

• Poorly differentiated gastroenteropancreatic neuroendocrine carcinomas (GEP NECs) are rare tumors with limited treatment options and a dismal prognosis

• Some patients with early-stage GEP NECs may achieve long-term survival with surgery or multimodality therapy, but most present or eventually develop rapidly lethal metastatic disease

• Systemic chemotherapy (etoposide- vs irinotecan-platinum doublets) is the mainstay of care in the metastatic setting, though responses are generally short lasting

• Standard of care shall shift from the small cell lung cancer treatment paradigm because emerging data points out major biological differences between GEP and lung NECs, and even within different GEP sites, suggesting site-specific approaches should be explored

• The role of immunotherapy in GEP NECs is still a matter of debate; better understanding of the immune landscape of tumors and their interaction with the host is needed to improve patient selection and most adequate strategy

• More than 20% of GEP NECs harbor potentially druggable molecular alterations (involving DNA damage repair, cell-cycle regulation, Notch or MAPK pathways, NTRK or RET gene fusions, among others) that should shift the treatment paradigm toward a more personalized, molecularly driven therapy that could eventually improve patient's survival

High-grade (G3, Ki-67 > 20%) neuroendocrine neoplasms (NENs) are rare aggressive tumors with a dismal prognosis (1–9). Only 9% are extrapulmonary, and 3% are of gastroenteropancreatic (GEP) origin. The majority are poorly differentiated small or large cell neuroendocrine carcinomas (NECs), but a small subset is well-differentiated tumors (neuroendocrine tumors [NETs]) and have a different biology and less ominous clinical course. Diagnosis and treatment are challenging because of the lack of high-quality data. Most patients are diagnosed with advanced disease, and systemic therapy is the mainstay of care. Treatment strategies are commonly adopted from those of high-grade lung NECs given their histological and clinical resemblance, with platinum/etoposide-based regimens recommended in the first-line setting (1–3). Recent evidence, however, indicates some significant differences in the molecular background of these tumors according to the anatomical site of origin, with potentially relevant therapeutic implications. Indeed, whereas TP53 and RB1 genomic alterations are almost universal in small cell lung carcinomas (SCLC), this is not the case for GEP NECs, particularly the more common large cell subtype (4, 5). Frequently mutated genes in GEP NECs include TP53, APC, KRAS, and BRAF, whereas RB1 copy number alterations are more common than mutations (<15%). MEN1, ATRX, DAXX, and SETD2 mutations may be typically identified in G3 GEP NETs, but a small proportion of well-differentiated tumors may also harbor TP53 mutations. More importantly, more than 20% of GEP NECs have potentially targetable genomic alterations (eg, DNA damage repair, BRAF/KRAS mutations, microsatellite instability)(5). Therefore, optimal therapy for GEP NECs may be site-specific and different from SCLC. But the biological background is still poorly understood, and further fundamental research is needed for progress to be made. Specific GEP NEC trials will be conducted taking into account the organ of origin, histological subtype, and proliferation index, and will incorporate comprehensive tumor molecular profiling to identify predictive biomarkers that foster personalized medicine. The aim of this manuscript is to review the state of the art of treatment strategies for patients with advanced high-grade GEP NECs, to discuss novel therapeutic approaches recently assessed in clinical trials including immunotherapy, radioligand and other targeted therapies, and to provide some insights regarding unmet needs and future perspectives to improve patient's care and prognosis.

 

Epidemiology and Prognosis

High-grade (G3, Ki-67 > 20%) NENs are rare aggressive tumors with a dismal prognosis. The incidence of high-grade NENs is 0.6 to 1 new cases/10⁵ habitants/year according to the largest published US and European population-based registries (6–8). Only 9% are extrapulmonary, one-third of which are of GEP origin (9). Regarding risk factors, there are also some differences with lung primaries because GEP NECs are not related to tobacco exposure (with the only exception being esophageal NECs) and some primary sites are related to human papillomavirus (rectal and anal canal). The great majority of lung NECs have small cell morphology (95%), whereas gastrointestinal NECs are more commonly (61%) of large cell (LC) morphology, except for esophageal and anal canal primaries. Less than 20% are diagnosed with stage I-II disease, and up to 65% of patients present metastatic disease at diagnosis. Brain metastases are common for lung NECs but occur in <2% of GEP NECs. Prognosis is overall poor but is better for GEP than lung primaries and large vs small cell NECs (5-year overall survival [OS] of 32%, 15%, and 6% for LC-GEP, LC-lung, and any site small cell NECs, respectively) (10). Within the GEP tract, most common primary sites are the pancreas (19%), colon (18%), esophagus (18%), and rectum (16%). About 25% to 30% of patients undergo surgical resection with curative intent, and their OS at 5 years ranges from 20% to 40% (7, 11). Prognosis within the GEP tract also varies by primary tumor location, with small bowel and esophageal NECs having the best and worst survival rates, respectively (9, 12). A small subset of G3 NENs (∼10%) are well-differentiated tumors (G3 NETs) and have a different biology and less ominous clinical course (13).

Pathological Classification: Implications for Patient's Care

The wide anatomical distribution of NENs has led to the development of multiple, not always consistent, organ-dependent classifications of these neoplasms, giving place to complex and sometimes confusing nomenclature. The World Health Organization (WHO) has progressively evolved toward a universal classification based on tumor differentiation and grade based on proliferation rate, and the most recent 2022 WHO Classification has, for the first time, included both endocrine and neuroendocrine tumors also of nonendocrine organs (14). Briefly, based on tumor differentiation, NENs are first stratified in well-differentiated tumors (NETs) and poorly differentiated carcinomas (NECs). Then, based on proliferation index, NETs are further subdivided in grade 1 (Ki-67 < 3%), grade 2 (Ki-67 3-20%), and grade 3 (Ki-67 > 20%), whereas NECs are always grade 3. Although this could seem quite simple and straightforward, morphological criteria to discriminate NETs from NECs are not well standardized, and significant discrepancies exist even among expert pathologists, who recommend considering additional clinical and molecular data to help in the differential diagnosis of these 2 entities. A prior or coexisting NET component would suggest a G3 NET, whereas a coexisting adenocarcinoma or squamous cell carcinoma would suggest a NEC or a mixed neuroendocrine-nonneuroendocrine neoplasm (MINEN). NECs do not resemble their nonneoplastic cell counterparts and have a solid growth pattern, often with abundant geographic necrosis and severe cell polymorphism, high nuclear to cytoplasm ratio, and frequent mitoses. NECs generally have reduced or absent hormone secretion and somatostatin receptor expression (though not always) (15–17). There are 2 different morphological subtypes of NECs, small cell and large cell NECs, and increasing evidence suggests they may have different pathogenetic pathways and clinical behavior. Whereas the small cell morphology is almost universal in lung NECs (95%), large cell is the more common GEP NEC (61%), particularly in small and large bowel NECs (9). The use of a panel of immunohistochemical markers is recommended for diagnosis. Positivity for cytokeratin and at least 1 general neuroendocrine marker (2 highly advisable for LC-NECs) is required for the diagnosis of NECs. The most commonly used are synaptophysin, chromogranin A, and INSM1 because enolase and PGP9.5 are highly nonspecific and may be misleading. CD56 should be interpreted with caution and is not to be used as a single marker. Very importantly, if morphological criteria for NENs are not met, positivity of general neuroendocrine markers is not sufficient to establish the diagnosis of NECs. The WHO classification also indicates that lack of pRb (protein encoded by the RB1 gene) immunostaining or overexpression or global loss of p53 are common in NECs and may help in the differential diagnosis of LC-NECs from G3 NETs. However, recent next generation sequencing data challenge this recommendation in GEP NECs because pRb mutations are particularly much less commonly encountered than in lung primaries (5).

Adequate pathological and molecular assessment is essential for optimal patient care. Indeed, discriminating G3 NETs from NECs (particularly LC-NECs) has relevant biological, prognostic, and therapeutic implications (16, 18). NET G3 have generally molecular alterations common to low-grade NETs, including loss of DAXX or ATRX expression (pancreatic NETs) and lack of p53/pRb alterations. More often, they can secrete hormones and express somatostatin receptors (15–17) and may thus benefit from radioligand therapy (RLT), and they have better survival than NECs but respond poorly to platinum-based chemotherapy (19). Thus, management of G3 NETs tends to resemble that of high-G2 NETs and will not be discussed in this review.

Molecular Profiling of GEP NECs

Small cell NECs, more commonly encountered within the GEP tract in the esophagus, gall bladder, and anal canal, are generally molecularly more homogeneous and, similarly to SCLC, are often characterized by biallelic inactivation of TP53 and RB1. On the contrary, large cell NECs are molecularly more complex and heterogeneous, often resembling nonneuroendocrine tumors of similar anatomic site (ie, KRAS mutations in pancreatic NECs, BRAF mutations in right colon NECs) (4, 5, 20). In fact, up to 40% of large cell NECs may harbor a nonneuroendocrine component, suggesting a common clonal precursor. Differential diagnosis of large cell NECs from G3 NETs is often challenging, and molecular profiling may be an important aid in this context. Progression from a NET to an NEC is extremely rare because NECs generally originate de novo or from a nonneuroendocrine epithelial cancer through a phenomenon known as lineage plasticity, as has been documented in lung or prostate adenocarcinomas under selective pressure of targeted therapies (anti-EGFR tyrosine kinase inhibitors or antiandrogens) (4).

The WHO classification of digestive tumors (2019) and classification of endocrine and neuroendocrine tumors (2022) point out TP53 and RB1 mutations as main differential molecular markers to discriminate NECs from high-grade NETs (14, 21). However, recent data indicate the incidence of these mutations in GEP NECs, particularly in those of large cell morphology, is much lower than that observed in lung NECs, and they may also occur in a small proportion of G3 NETs (Fig. 1). Indeed, in 1 of the largest reported molecularly profiled series of G3 GEP NENs (N = 181), TP53 mutations were observed in 64% of NECs vs 14% of NETs, whereas RB1 mutations were rare (14% of NECs) but copy number alterations (losses or deletions) affecting RB1 were common in both NECs (34%) and NETs (31%) (5). Other frequently encountered gene copy number losses were ARID1A (35%), ATM (31%), and ESR1 (25%), and common gene amplifications were found in MYC (51%) and KDM5A (45%). Of note, gene mutations associated with the nonneuroendocrine counterparts could better discriminate NECs from NETs, including BRAF (20% vs 3%) or KRAS (22% vs 0%) mutations, particularly in certain anatomic sites (ie, BRAF mutations occur in 49% of colon NECs and 70% if located in the right colon; FBXW7 mutations were seen in 25% of rectal cancers). In fact, a 4-gene classifier including TP53, APC, KRAS, and BRAF mutations was found to yield 92% sensitivity, 61% specificity, and 83% positive predictive value for discerning large cell NECs from G3 NETs. Moreover, specificity (89%) and positive predictive value (96%) were further improved if 5 additional genes were included in a 9-gene model (TP53, APC, KRAS, BRAF, ATRX, DAXX, MEN1, MYOB5B, and SMAD2). These are interesting findings that, if validated, could be potentially implemented in clinical practice.

Figure 1.

Common genomic alterations in GEP NECs and G3 NETs. Bar plots depict the frequency of samples harboring any genomic alteration (point mutations, copy number gains or losses, amplifications or deletions) for the most frequently mutated genes in small cell NECs, large cell NECs, and G3 NETs of GEP origin (5). G3 NET, grade 3 neuroendocrine tumor; GEP: gastroenteropancreatic; NEC: neuroendocrine carcinoma.

Very interestingly, microsatellite instability (MSI) has been documented in ∼5% to 12% of GEP NECs and even in ∼3% of G3 NETs (5, 18, 22–24). As in colorectal adenocarcinoma, cooccurrence with BRAF mutations is common. Because both MSI and BRAF mutations are potentially targetable with drugs approved for tumor agnostic indications, their routine testing should be considered in patients with advance disease, at least in those of colonic origin. Overall, at least 20% of NECs may have these and other druggable molecular alterations (ie, deficient DNA damage repair) and, thus, molecular profiling is highly recommended if available in this setting.

MSI is an established predictive marker of response to immunotherapy because it is the consequence of deficient DNA mismatch repair that leads to highly mutated and immunogenic tumors. A high tumor mutational burden (TMB-H) may also occur through other mechanisms and is also associated with sensitivity to immunotherapy. Although up to 20% of G3 NENs have been reported to be TMB-H, this seems less common in those of GEP origin, with median TMB generally below 10 mutations/megabase (5, 25).

Disruption of epigenetic regulation also plays an important role in GEP NECs, with the AT-Rich Interaction Domain 1A (ARIDA1A), histone lysine methyltransferase 2 (KMT2), and histone lysine demethylase (KMD) family genes being the most frequently affected (5, 26–29). Methylation profiles seem to differ by primary tumor site, with highly methylated tumors located in the stomach, intestine, and gallbladder (30–32). Interestingly, high methylation levels are associated with MSI phenotype, as frequently methylated genes in MSI NECs include MLH1 (23). Other commonly methylated genes include MGMT and RASSF1A, particularly in foregut NECs. Methylation analysis also clearly discriminated NECs from adenocarcinomas in different primary sites, such as the pancreas or colorectum (31, 33, 34). Phylo-epigenetic and cell-type signature features derived from alpha, beta, acinar, and ductal adult cells suggest an exocrine cell of origin for pancreatic NECs, thus separating them in cell lineage from other pancreatic NENs of endocrine origin. Genes differentially methylated in colorectal NECs (co-NECs) were involved in immune cell differentiation, cytoskeleton dynamics and cell polarity, DNA damage, or apoptosis. Of note, genes typically expressed in neuroendocrine cells, such as somatostatin and other neurotransmitters, were highly methylated in colorectal adenocarcinomas. Moreover, EGFR methylation was also more commonly encountered in co-NECs than in adenocarcinomas, and EGFR methylation status and its related protein expression was directly correlated with response and acquired resistance to BRAF inhibitors in patient-derived xenograft models of co-NECs. Finally, recently published data by Kawasaki et al have also reported mutations in epigenetic modifier genes and promoter demethylation of key neuroendocrine transcription factors in GEP NEC organoids, further supporting the relevance of epigenetic alterations in the development of these neoplasms and as a potential therapeutic target that deserves to be further explored (35).

Standard of Care of GEP NECs: How Far Can We Go With Chemotherapy?

Surgery with curative intent is usually recommended for patients with locoregional disease because large retrospective series suggest it significantly improves survival regardless of primary tumor site and perioperative therapies (36). In patients with important comorbidities or in anatomical sites where surgical resection is associated with high morbidity (ie, anal canal, esophagus), radiotherapy may be the local treatment of choice, generally combined with chemotherapy (2). Based on the high risk of systemic relapse after primary tumor resection, platinum-based adjuvant chemotherapy is generally indicated following surgery in patients with localized NECs, although evidence to support this common practice is rather weak. Moreover, retrospective data suggest the potential added value of adjuvant chemotherapy may differ depending on pathological features (greater benefit in T3-T4, node positive, or positive resection margins) and primary tumor site (greater benefit reported in esophageal and colorectal primaries; no clear benefit in gastric, small bowel, or pancreatic NECs) (37–41). Consistent with this, treatment patterns generally differ by age, stage, morphology, and primary tumor location (36). Some experts/guidelines also support the use of neoadjuvant chemotherapy with or without radiotherapy and resection in selected patients with locally advanced disease, potentially enabling less aggressive surgery or avoiding major surgery in high-risk patients who experience early progression (1, 42). However, no randomized trials have properly assessed to date the role of adjuvant or neoadjuvant chemotherapy and/or radiotherapy in this setting. Nevertheless, a relevant proportion of patients with locoregional GEP NECs achieve long-term survival generally with bi- or trimodality therapy, with 5-year survival rates of 40% to 50% for patients with gastric, pancreatic, or colorectal primaries, and of 25% for those with esophageal NECs (9). Long-term cure therefore seems achievable in some patients.

In summary, surgery is generally recommended for fit patients with localized or locally advanced resectable GEP NECs, and adjuvant platinum-etoposide may be considered. Exceptions include anatomical sites where surgical resection is highly morbid, such as esophageal or anal canal NECs, in which case chemoradiation is the preferred therapeutic approach. There is less consensus regarding optimal management of resectable distal rectal NECs. Neoadjuvant chemotherapy or chemoradiotherapy may be considered before surgery, but treatment decisions should be individualized and discussed within a multidisciplinary team.

However, more than 60% to 70% of patients with GEP NECs present metastatic disease at diagnosis and have a very poor prognosis, with a median OS of up to 12 months with best available therapy, and of barely 1 month in patients who only receive supportive care (12, 13, 43). Evidence to support treatment recommendations in GEP NECs is scarce, and clinical guidelines have been primarily based on the more solid data available from SCLC studies, even though there are important etiological, clinical, and molecular differences between these 2 entities, as has been discussed (4, 9). The standard of care for patients with advanced unresectable or metastatic GEP NECs is systemic chemotherapy, provided the patients have adequate performance status and organ function. Cisplatin or carboplatin in combination with etoposide are the most widely used chemotherapy regimens, although objective response rates (ORRs; 28%-48%) are generally lower than those reported for SCLC (9, 12, 43–45). Carboplatin is generally preferred over cisplatin because it has similar antitumor activity and better toxicity profile, with no nephro- nor neurotoxicity (12). Response to chemotherapy is lower in NECs with Ki-67 < 55% (15% vs 42%), although they have better survival than patients with Ki-67 ≥ 55% (14 vs 10 months) (12). Metastatic colorectal NECs have a particularly dismal prognosis, with more than 50% experiencing early disease progression with standard chemotherapy and a median survival of about 8 months. Some experts advocate the use upfront of fluoropyrimidine-based regimens with irinotecan or oxaliplatin in large cell GEP NECs, particularly if they are associated with a non-neuroendocrine component (MINEN). An ongoing randomized trial (FOLFIRINEC) is currently assessing FOLFIRINOX vs platinum-etoposide in NECs of GEP or unknown origin. Two randomized trials conducted in Asian population with GEP NECs have compared cisplatin and etoposide with cisplatin and irinotecan, demonstrating similar ORR (53%-63%), progression-free survival (PFS; 5.1-6.4 months), and OS (10.2-12.5 months) for both treatment regimens (Table 1) (46, 47). Substituting irinotecan for etoposide is therefore an alternative treatment option for these patients that is associated with less myelosuppression although a greater incidence of diarrhea. More recently, a small, randomized trial conducted in the United States compared platinum and etoposide vs CAPTEM in non–small cell GEP-NENs (including NETs and NECs), and was closed for futility, concluding CAPTEM was better tolerated but not associated with increased efficacy (Table 1) (48). No subgroup analysis has been provided yet by tumor differentiation or Ki-67 index in this trial.

Table 1.

Randomized trials of chemotherapy in GEP NECs

Trial and author	N	Primary site	Treatment	Primary Endpoint	ORR (%)	PFS (median)	OS (median)
First-line therapy
TOPIC NEC Morizane et al	170	GEP (100 GI, 70 HBP)	CDDP + etoposide CDDP + irinotecan	OS	54.5% 52.5%	5.6 mo 5.1 mo	12.5 mo 10.9 mo
Zhang et al	66	GEP	CDDP + etoposide CDDP + irinotecan	ORR	63.2% 61.5%	6.4 mo 5.8 mos	11.3 mo 10.2 mo
Second-line therapy
NET-02 McNamara et al	58	GEP (40) Other EP NECs (11) UKP NECs (7)	Nal-IRI + 5FU docetaxel	6 mo	10.3% 10.3%	3 mos (31% 6-mo PFS) 2 mo (14% 6-mo PFS)	9.0 mo 5.0 mo
PRODIGE 41 (BEVANEC) Walter et al	126	GEP UKP	FOLFIRI + Beva FOLFIRI	6 mo OS	25.5% 18.3%	3.7 mo 3.5 mo	7.0 mo (52.5% 6- mo OS) 8.9 mo (58.2% 6-mo OS)

Abbreviations: CDDP, cisplatin; EP-NEC, extrapulmonary neuroendocrine carcinoma; GEP, gastroenteropancreatic; GI, gastrointestinal; HBP, hepatobiliopancreatic; IRI, irinotecan; NEC, neuroendocrine carcinoma; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.

The efficacy of chemotherapy in the refractory setting (second line and beyond) is very limited and survival is short (∼9 months with best available therapy). Although second-line regimens have not been extensively evaluated either, emerging data from recent trials consolidate infusional 5-FU and irinotecan (standard or pegylated liposomal formulation) (FOLFIRI or nal-IRI-5FU) as the treatment of choice in this setting (Table 1) (49, 50). The PRODIGE 41-BEVANEC phase 2 trial randomized 126 patients with advanced NECs of GEP or unknown origin, after failure to first-line platinum etoposide, to receive FOLFIRI (control arm) vs FOLFIRI with bevacizumab (experimental arm) (25). The trial reached its primary endpoint (6-month OS >50% in the experimental arm), but the addition of bevacizumab did not seem to improve PFS nor OS over FOLFIRI alone. A smaller phase 2 randomized, noncomparative study (NET-02) included 58 patients with extrapulmonary NECs (40 were of GEP origin) that were allocated to receive nal-IRI-5FU vs docetaxel as second-line therapy on progression to first-line platinum-based therapy (49). Liposomal irinotecan-5FU, but not docetaxel, met the primary endpoint with a 31% 6-month PFS rate that exceeded the prespecified threshold for efficacy. ORR were similar in both arms, and OS also favored the nal-IRI-5FU arm. This outcome is consistent with that reported in the BEVANEC trial for the FOLFIRI control arm, and both studies provide the most solid evidence available to date on salvage chemotherapy in NECs. Other treatment options include oxaliplatin- or temozolomide-based regimens, although evidence derives from retrospective series or tumor registries (2, 12, 43, 45).

In conclusion, carboplatin and etoposide is the preferred first-line treatment option for patients with metastatic GEP NECs. Irinotecan and cisplatin are an alternative treatment regimen that may be considered for patients with poor bone marrow reserve or who at risk of recurrent or severe infections because it has demonstrated similar efficacy and is associated with less myelosuppression. In the second-line setting, on failure of platinum-based chemotherapy, FOLFIRI is the preferred chemotherapy schedule based on best available evidence. Alternative regimens in the refractory setting include combinations of fluoropyrimidines with oxaliplatin or temozolomide, although efficacy is rather limited for any therapy in this context.

Theragnostics in NECs: Does it Play any Role?

¹⁷⁷Lutetium-DOTATATE is currently approved for the treatment of advanced, progressive, somatostatin receptor-positive well-differentiated G1-2 GEP NETs, though emerging evidence suggests it could also be effective in some G3 NENs. Consistent data from several noncontrolled studies reported ORR in G3 NETs or low-G3 NECs (Ki-67 < 55%) of 42% to 43%, with immediate disease progression more common in lowG3-NEC than in G3 NETs (26 vs 7%), and much higher in NECs with Ki-67 > 55% (45%) (51). PFS was 19, 11, and 4 months for G3 NETs, lowG3-NECs, and highG3-NECs, and OS 44, 22, and 9 months, respectively. Therefore, RLT may be considered in G3 NETs or lowG3-NECs homogeneously expressing somatostatin receptors.

Two ongoing randomized trials are currently assessing the role of somatostatin receptor-targeted radioligand therapy in somatostatin receptor-positive, low G3 GEP NENs. The NETTER-2 trial (NCT03972488) aims to determine if Lu177-oxodotreotide in combination with standard doses of long-acting octreotide (30 mg) prolongs PFS in advanced G2-3 GEP NETs (Ki-67 10%-55%), compared with treatment with high-dose (60 mg), long-acting octreotide. The COMPOSE trial (NCT04919226) is currently recruiting patients with G2-3 GEP NETs (Ki-67 15%-55%) that are randomized to 177Lu-edotreotide vs investigator's choice of standard chemotherapy (CAPTEM or FOLFOX) as first- or second-line therapy in the advanced setting. Both trials will provide relevant information regarding safety and efficacy of peptide receptor radionuclide therapy (PRRT) relative to other treatment options in this setting and will help define the role of PRRT in the therapeutic algorithm of G2-3 GEP NENs.

Other potential theragnostic targets being explored in high-grade NENs include C-X-C motif chemokine receptor-4 and urokinase plasminogen activator receptor (uPAR) (52, 53). Overexpression of CXCR4 has been documented in SST2-negative, high-grade NENs. 68Ga-Pentixafor show reliable assessment of CXCR4 expression, thereby identifying potential candidates for CXCR4-targeted RLT with 177Lu/90Y-pentixather (52). The main pitfall of this approach is that CXCR4 is also highly expressed in hematopoietic stem cells, making myelotoxicity a dose-limiting factor for therapy.

uPAR is a cell membrane receptor part of the plasminogen activation system, which is physiologically involved in tissue reorganization events (ie, wound healing). uPAR interacts with urokinase plasminogen activator, integrins, and other proteins, and through cross-talk with tyrosine kinase receptors regulates the shift between tumor dormancy and proliferation and promotes cell motility and invasion and angiogenesis. It is overexpressed in many human cancers, including high-grade NENs and correlates with early invasion and metastasis and a poor prognosis. ⁶⁸Ga-NOTA-AE105 positron emission tomography (PET) was explored in a phase 2 trial to image uPAR in 99 patients with G1-3 NENs, and uPAR-positive lesions were seen in both low-grade (57% G1, 69% G2) and high-grade NENs (75%) (53). A high uPAR expression was associated with a worse PFS and OS. uPAR PET may thus be useful for risk stratification of patients in the clinic and as a potential new target for RLT in patients with NENs.

Novel Therapeutic Strategies

Immunotherapy

The immune system plays a pivotal role in cancer prevention, development, and progression, and immune checkpoint inhibition (ICI) has changed the treatment paradigm of many cancer types over the past decade, including some high-grade NENs such as Merkel cell carcinoma (54) or small cell lung cancer (55, 56). Nevertheless, the role of immunotherapy remains controversial in G3 digestive NENs.

Single-agent PD-1 or PD-L1 blockade (ie, pembrolizumab, spartalizumab, avelumab) has shown to be mostly ineffective (ORR 3%-7%, PFS 1.8-4 months, OS 5.1-7 months) in molecularly unselected, heavily pretreated patients with advanced GEP NECs (57–59) (Table 2). The only exception was toripalimab, which reported an ORR of 20% in G2high-G3 NENs (primarily GEP NECs), with a median duration of response of 15.1 months (60). Interestingly, ORR was significantly higher in patients with PD-L1 expression ≥10% or TMB-H than PD-L1 < 10% (50.0% vs 10.7%, P = .019) and TMB-low patients (75.0% vs 16.1%, P = .03), and 3 of the 8 (37.5%) responders harbored ARID1A mutations. ARID1A gene product is a member of the switching/sucrose nonfermentable complex involved in chromatin remodeling and interacts with the DNA mismatch repair system. ARID1A alterations have been associated with MSI and TMB-H in several cancer types and have been also correlated with improved outcomes after immune checkpoint blockade (61). PD-L1 expression and TMB has been positively correlated to tumor grade and are associated with a poorer survival in NENs (62).

Table 2.

Clinical trials assessing immunotherapy in GEP NECs

Trial and author	N	Primary site	Treatment	Primary endpoint	ORR (%)	PFS (median)	OS (median)
First-line therapy
NICE NEC trial Riesco-Martinez et al	38	GEP or UKP G3 NENs	Nivolumab + CBDCA + etoposide	12-mo OS	54.1%	5.7 mo (43.2% 6-mo PFS, 17.5% 12-mo PFS)	13.9 mo (54% 12-mo OS, 44% 18-mo OS)
Second-line therapy and beyond
NIPINEC trial Walter et al	185	93 lung LC-NEC 92 GEP NECs	Nivolumab vs nivolumab + ipilimumab	ORR 8 wk	Nivo vs Nivo-Ipi: Lung: 7.3% vs 18.2% GEP: 7.1% vs 11.6%	1.8 vs 1.9 mo	7.2 vs 5.8 mo
DUNE trial Capdevila et al	33	C4: G3 GEP NENs	Durvalumab + tremelimumab	9-moOS	9.1%	2.4 mo	5.9 mo (9-mo OS 36%)
CA209-538 Basket trial Klein et al	29	NENs any site (13 G3, 10 GEP, 4 GEP NECs)	Nivolumab + ipilimumab	DCR	24% (G3 31%) (DCR 72%)	4.8 mo	14.8 mo
DART SWOG 1609 Basket trial Patel et al	32	NENs any site (18 G3, 15 GEP, 8 GEP NECs)	Nivolumab + ipilimumab	ORR	25% (G3 44%)	4 mo (G3 6-mo PFS 44%)	11 mo
Mulvey et al Chan et al	14 (A) 22 (B)	EP-NECs (6 + 16 GEP)	Part A: Pembro Part B: Pembro + CT (IRI or PAC)	ORR	A: 7% B: 5%	A: 1.9 mo B: 2.0 mo	A: NR B: 4.9 mo
Vijayvergia et al	29	G3 EP-NENs (19 NECs, 24 GEP)	Pembrolizumab	ORR	3.4%	2.2 mo	5.1 mo
Yao et al	116	21 GEP-NECs	Spartalizumab	ORR	4.8%	1.8 mo	6.8 mo
Lu et al	40	G2-3 NENs (Ki67 > 10%) (32 GEP, 32 NEC)	Toripalimab	ORR	All patients 20% GI NEN 13% Pan NEN 22% Other primaries 38% PD-L1 > 1 50% TMB-H 75%	2.5 mo	7.8 mo
AVENEC trial Fottner et al	29	G3 NENs (21 GEP, 19 NEC)	Avelumab	DCR	6.9% (DCR 27.6%)	4.0 months	7.0 mo (NEC 4.7 mo, G3 NET NR)

Abbreviations: CBDCA, carboplatin; CDDP, cisplatin; CT, chemotherapy; C4, cohort 4; DCR, disease control rate; EP-NEC, extrapulmonary neuroendocrine carcinoma; GEP, gastroenteropancreatic; LC-NEC, large-cell neuroendocrine carcinoma; NEC, neuroendocrine carcinoma; NEN, neuroendocrine neoplasm; NET, neuroendocrine tumors; ORR, objective response rate; NR, not reported or not reached; OS, overall survival; Pembro, pembrolizumab; PFS, progression-free survival; TMB-H, high tumor mutational burden; UKP, unknown primary.

Dual blockade using anti–PD-1 and anti–CTLA-4 monoclonal antibodies (MAbs) has demonstrated increased efficacy compared with single-agent ICI in several tumor types such as melanoma, renal cell carcinoma, MSI colorectal cancer, and certain subsets of non-SCLC, among others. The combination of nivolumab (anti-PD1 Mab) and ipilimumab (anti-CTLA4 MAb) showed promising activity in small cohorts of patients with high-grade NENs enrolled in 2 basket trials conducted in patients with rare cancers, DART SWOG 1609 and CA209-538 (63, 64). The DART trial reported an ORR of 44% in 18 patients with high-grade nonpancreatic NECs (including lung primaries and 8 GEP NECs), with a 6-month PFS rate of 44% (63). The CA209-538 documented an ORR of 31% in an additional cohort of 13 patients with high-grade NENs (4 GEP NECs) (64).

However, more recent data from the DUNE trial suggested only modest activity for durvalumab and tremelimumab in a cohort of 33 patients with G3 GEP NENs, with an ORR of 9.1% and a 9-month OS rate of 36.1%, although it surpassed the preestablished OS futility threshold (primary endpoint) (65). Of note, 10 (30.3%) patients were alive 12 months after treatment initiation. Outcomes in this cohort did not substantially differ by tumor differentiation (NET vs NEC) and Ki-67 index, and PD-L1 combined score did not correlate with treatment efficacy.

Consistent with results of the DUNE trial, the NIPINEC study (GCO-001) also showed in a larger patient population modest activity of dual blockade with nivolumab and ipilimumab in GEP NECs (66). This was a randomized noncomparative trial that included 185 patients with advanced, refractory large-cell pulmonary (N = 93) or GEP NECs (N = 92) on progression to 1 or 2 previous systemic treatment lines including platinum-based chemotherapy. These patients were randomly allocated to receive nivolumab or nivolumab plus ipilimumab. In this trial, combination of nivolumab and ipilimumab, but not nivolumab as single agent, reached its primary endpoint of ORR at 8 weeks >10%. Dual ICI induced higher ORR in lung large-cell NECs (18.2%) than in GEP NECs (11.6%). Global median PFS (1.9 months) and OS (5.8 months) were not encouraging, although further follow-up is needed to identify whether a proportion of patients achieve long-term benefit from this treatment strategy.

In summary, overall efficacy of dual immune checkpoint blockade seems lower in digestive than in lung NECs, although a small subset of patients may be long-term survivors. Thus, further efforts will be pursued to identify who these patients may be. Unfortunately, MSI was not assessed in any of these trials. The lack of treatment control and short follow-up of these studies further limits proper assessment of the real impact of this therapeutic approach on patient's long-term outcomes. Despite these modest results, however, and taking into account the limited efficacy of chemotherapy in this refractory setting, the National Comprehensive Cancer Network guidelines indicate that dual immune checkpoint blockade can be considered in NECs when the disease progresses following chemotherapy.

On the other hand, emerging evidence suggests that conventional chemotherapy induces an immunogenic cell death that can prime antitumor immunity within an immunosuppressive microenvironment and may be synergistic with immune checkpoint inhibition (67). This has been demonstrated in chemotherapy-näive patients with advanced SCLC because the addition of durvalumab or atezolizumab to first-line platinum-based chemotherapy improves survival and is now the standard of care for these patients (55, 56). In this context, the NICE NEC trial assessed this strategy in advanced untreated G3 NENs of GEP or unknown origin. The combination of nivolumab with carboplatin and etoposide induced an ORR of 54% and sustained response beyond 12 months in 6 of 38 patients (16%) with a median OS of 13.9 months (68). On the contrary, limited efficacy was reported with the combination of ICI (pembrolizumab) and chemotherapy (investigator's choice of either irinotecan or paclitaxel) in refractory extrapulmonary NECs (patients progressed to 1 to 3 prior lines of therapy) (69). Ongoing translational studies may help identify predictive or prognostic biomarkers to improve selection of patients most likely to benefit from this treatment strategy, and randomized trials would be required to properly assess the role of adding ICI to chemotherapy in this setting.

Finally, preclinical evidence suggests the genomic instability induced by drugs that target cell-cycle or DNA damage repair (DDR) regulators such as PARP, CHK1, or CDKs, may boost sensitivity to ICI, and these combinations are currently being explored in clinical trials in NECs. Other alternative immunotherapeutic approaches include bispecific antibodies or bispecific T-cell engagers that concurrently target CD3 in lymphocytes and specific tumor targets such as somatostatin receptors or DLL3, to redirect cytotoxic T cells against cancer cells (70). These therapies show promising activity in preclinical models and are currently in early stages of clinical development in NENs.

Other immunotherapeutic strategy in development, which may render immunologically “cold” tumors more sensitive to immune checkpoint blockade, are oncolytic viruses. Their antitumor activity is based on the ability to selectively replicate in tumor cells causing cell lysis, to amplify the initial therapeutic dose through replication in vivo, and to induce innate and adaptive antitumor immunity. A variety of native and, more commonly, genetically modified viruses have been successfully developed as oncolytic agents, such as T-Vec (talimogene laherparepvec) for melanoma or oncorine (H101) for head and neck tumors. Oncolytic viruses explored in NET preclinical models with some preliminary evidence of efficacy include GLV-1 h68, a genetically engineered vaccinia virus; HSV T-01, a third-generation herpes simplex; or Ad5fk^FWKT (CgA-E1A-miR122), a modified adenovirus with somatostatin motifs for selective infection of neuroendocrine tumor cells, and some early clinical trials in NEN patients are currently under way (71–73). Oncolytic virus may also be engineered to express different transgenic payloads, such as extracellular matrix-degrading enzymes, pro-inflammatory cytokines, antibodies (ie, ICI), and specific antigens that can expand their therapeutic potential. Still, many challenges posed by tumor heterogeneity, physical barriers, antiviral immunity, and the immunosuppressive tumor microenvironment need to be overcome to improve antitumor efficacy of oncolytic virotherapy.

Tumor Agnostic Therapy

Deficient mismatch repair and high mutational burden

Immune checkpoint inhibition has proven to be capable of eliciting a T cell–mediated cytotoxic effect that induces profound and durable tumor responses in certain subsets of patients with cancer. To identify this subset of tumors, several molecular determinants have been explored in specific tumor types with somewhat discordant results, including, among others, PD-L1 expression, RNA expression signatures, mutational burden, and patterns of lymphocytic infiltrates. But probably the most robust pan-tumor predictive biomarker to ICI is DNA mismatch repair deficiency (dMMR) or its phenotypic consequence, MSI. Mismatch repair deficiency results in an inability to recognize and repair spontaneous mutations during DNA replication and, thus, dMMR/MSI tumors harbor thousands of mutations (TMB > 50 mutations/megabase), which are highly immunogenic. Indeed, their typically prominent tumor lymphocytic infiltrates is consistent with an antitumor immune response. However, their active immune microenvironment is counterbalanced by immune inhibitory signals (PD-1, PDL1, CTLA-4, LAG-3, IDO) that lead to tumor immune scape. In this context, ICI can unlock a potent tumor-antigen specific immunological response leading to significant clinical benefit.

In 2017, pembrolizumab was granted accelerated approval by the US Food and Drug Administration (FDA) for patients with advanced dMMR/MSI solid tumors that have progressed from prior treatment and who have no satisfactory alternative treatment options. This was the first tumor agnostic indication ever approved in cancer. The approval was based on data from 149 patients with 15 types of MSI/dMMR solid tumors (90 patients had colorectal cancer, 59 had other tumor types) enrolled across 5 multicohort single-arm clinical trials (74). The global ORR was 39.6%, and responses were profound and durable, with median PFS and OS that had not been reached. Several years later (2021-2022), the European Medicines Agency also approved pembrolizumab for a more restricted spectrum of MSI/dMMR tumors, including colorectal, endometrial, gastric, small intestine, or biliary cancer on progression to at least 1 prior therapy. No specific data have been reported, however, in patients with dMMR/MSI NENs, which occurs in 3% to 12% of high-grade GEP NENs. MSI NECs are more commonly encountered in those of gastric or colorectal origin and are enriched in BRAF mutations and CpG island methylator phenotype.

In 2020, the FDA granted pembrolizumab its second tumor agnostic indication for the treatment of advanced TMB-H solid tumors (≥10 mutations/megabase) that have progressed to prior therapy and with no satisfactory alternative treatment options, and also approved the FoundationOneCDx assay (Foundation Medicine, Inc.) as a companion diagnostic. Approval was based on the results of a planned retrospective analysis of 10 cohorts of patients included in the KEYNOTE-158 trial (NCT02628067) (75). A total of 105 patients (13%) had TMB-H tumors, including 5 of 87 (5.7%) patients with neuroendocrine tumors (tumor site, differentiation, and grade not specified). The global ORR was 29%, 28% among the 81 patients with non-MSI TMB-H tumors. Two of the 5 (40%) patients with TMB-H neuroendocrine tumors achieved an ORR. The median duration of response was not reached, with 50% of patients having response durations ≥24 months. Although TMB-H has been reported in up to 20% of G3 NENs, the precise incidence in GEP NECs is uncertain and seems less common than in other anatomic locations such as head and neck or lung NECs (5, 25).

Based on these data, pembrolizumab may be considered for the treatment of patients with advanced dMMR/MSI and/or TMB-high NECs that have progressed following prior treatment and have no alternative treatment options.

Neurotrophic tyrosine receptor kinase and RET gene fusions

Other molecular alterations that may be encountered in NECs and are efficiently targeted by approved tumor-agnostic drugs (larotrectinib, entrectinib) include neurotrophic tyrosine receptor kinase (NTRK) fusions (76, 77). NTRK gene fusions, resulting in chimeric TRK proteins, are oncogenic drivers in a diverse range of solid tumor types. Entrectinib is a potent, central nervous system-active TRK inhibitor approved for the treatment of patients with advanced solid tumors harboring an NTRK gene fusion. Approval was based on results of 3 single-arm clinical trials (ALKA, STARTRK-1 [NCT02097810] and STARTRK-2 [NCT02568267]) that enrolled 54 patients with NTRK-positive tumors (78). The ORR was 57% and in 45%, the patient response duration was 12 months or longer. The most common cancers included in these trials were sarcoma, non-SCLC, mammary analogue secretory carcinoma, breast, thyroid, and colorectal. Among the 5 patients with NENs enrolled in these trials, the ORR was 40% (79). However, relevant patients’ features of these NENs (tumor differentiation, grade, or primary site) have not been reported. More recently, a phase 1 trial with taletrectinib (AB-106/DS-6051b), a tyrosine kinase inhibitor with high affinity for ROS1 and NTRK1/2/3, was conducted in patients with neuroendocrine tumors, tumor-induced pain, or tumors harboring ROS1/NTRK rearrangements (80). Patients with NETs were eligible for this trial based on preclinical evidence suggesting TrkB (NTRK2) signaling pathway could be involved in the tumorigenesis, proliferation, and invasive nature of NETs. However, the ORR among the 12 included patients with molecularly unselected NENs was modest, with 1 patient (8.3%) with a heavily pretreated small bowel NEC achieving a confirmed partial response (PR), and a median PFS of 10.2 months (80). These results would only justify the use of NTRK inhibitors in NEN patients with NTRK rearrangements, although the incidence of NTRK fusions may be somewhat higher than in many other common malignancies. In fact, comprehensive genomic profiling identified NTRK fusions in 6 of 2417 NENs (0.3%), including patients with pancreatic (N = 2), uterine (N = 1), lung (N = 1), and unknown (N = 2) primary tumor sites. Three of them involved translocations of NTRK1 with unique fusion partners (GPATCH4, PIP5K1A, CCDC19). However, detailed information of the full study cohort regarding tumor differentiation, grade, and primary site was not provided.

RET inhibitors have also demonstrated efficacy in patients with a variety of RET fusion–positive solid tumors. RET is a proto-oncogene involved in many essential physiological and developmental functions, including spermatogonial stem cell maintenance, renal morphogenesis, and neural and neuroendocrine tissue development. Aberrant activation of RET through gain-of-function mutations is associated with the hereditary predisposition syndrome multiple endocrine neoplasia 2, which increases the risk of developing medullary thyroid cancer, pheochromocytoma, and parathyroid hyperplasia or adenoma. RET rearrangements occur in 10% to 20% of sporadic papillary thyroid cancer and in 1% to 3% of patients with non-SCLC, and are far less common in other tumor types (81). Pralsetinib is a selective RET inhibitor that potently inhibits RET kinases, including RET fusion proteins. The phase 1-2 ARROW trial enrolled 29 patients with 12 different RET fusion–positive solid tumor types, excluding thyroid and non-SCLC (82). ORR was 57%, and responses were observed regardless of tumor type or RET fusion partner, including 2 of 3 patients (67%) with RET-fusion positive neuroendocrine cancer. Median duration of response was 12 months, and median PFS and OS were 7 and 14 months, respectively. These data indicate RET could also be a tissue-agnostic target, although the incidence of RET fusions in NENs is uncertain and, to date, RET inhibitors are only approved for the treatment of lung and thyroid tumors harboring RET rearrangements.

MAPK signalling pathway

As in adenocarcinomas of similar anatomical location, aberrant activation of mitogen-activated protein kinase signaling pathway is common in GEP NECs. KRAS and BRAF mutations are frequently encountered in digestive NECs (24% and 29% of large cell NECs, respectively), particularly in those of pancreatic or colorectal origin (5). Of note, up to 70% of right colon NECs harbor BRAF V600E mutations. BRAF inhibition, alone or in combination with MEK or EGFR inhibition, has demonstrated antitumor activity in preclinical models and patients (case reports) (83–85). A recent study in patient's samples and patient-derived xenografts showed a distinctive methylome profile of co-NECs compared with colorectal adenocarcinomas, resulting in EGFR expression silencing (34). Thus, authors suggest BRAF-mutated co-NECs could potentially benefit from single-agent BRAF inhibition and support the addition of EGFR inhibition in the setting of acquired resistance. It should be pointed out, however, that these tumors often harbor a nonneuroendocrine component, and the combination of encorafenib and cetuximab is approved for the treatment of colorectal cancer (histology not specified). Moreover, in June 2022, the FDA approved the combination of dabrafenib (BRAF inhibitor) and trametinib (MEK inhibitor) for the treatment of adult and pediatric patients with advanced solid tumors with the BRAF V600E mutation. The efficacy of this combination was evaluated in 131 adult and 36 pediatric patients included in several open-label basket trials (BRF117019 [NCT02034110], NCI-MATCH [NCT02465060], CTMT212X2101 [NCT02124772]), and supported by results in COMBI-d, COMBI-v, and BRF113928 (86). The studies enrolled patients with 24 tumor types, including 2 patients with gastrointestinal MINEN and 2 patients with colon neuroendocrine carcinomas. The ORR was 41% in adults and 25% in pediatric patients. Duration of response was ≥24 months for 44% of patients. No ORR was documented in the few NEN patients included in the NCI-MATCH trial. Overall, these data support further specific clinical assessment in co-NECs of these therapeutic strategies.

Although traditionally considered “undruggable,” recent advances in targeting KRAS in a variety of solid tumors are encouraging, particularly KRAS G12C mutations, although no specific data have been reported in NECs (87). Referral of patients with KRAS-mutant tumors for enrollment in clinical trials is highly recommended.

Other Potentially Targetable Tumor Vulnerabilities

Notch signaling plays a suppressor function in NENs. Delta-like ligand 3 (DLL3), unlike other ligands, negatively regulate Notch signaling and is upregulated in aggressive NENs including GEP NECs, whereas its expression is almost absent in well-differentiated NETs or normal cells (70). This makes DLL3 a suitable target for directed therapies, including antibody-drug conjugates, adoptive cell therapy or radioligand therapy. rovalpituzumab tesirine (Rova-T), a DLL3-targeted antibody-drug conjugate consisting of a humanized anti-DLL3 IgG1 antibody linked to a pyrrolobenzodiazepine dimer toxin, reported promising antitumor activity in phase 1-2 trials conducted in SCLC, particularly in DLL3 high-expressing tumors. However, phase 3 trials assessing Rova-T as a first-line maintenance therapy following first-line platinum-based chemotherapy (MERU trial, NCT03033511) or comparing Rova-T to topotecan as a second-line therapy (TAHOE trial, NCT03061812) in patients with advanced SCLC failed to demonstrate a survival benefit or were even detrimental (88, 89). Moreover, Rova-T was associated with significant toxicity attributed to the pyrrolobenzodiazepine payload, including pleural and pericardial effusions, peripheral edema, pneumonitis, and thrombocytopenia, altogether leading to discontinuation of its clinical development. Rova-T was tested in 28 patients with GEP NECs (68% DLL3 positive) included in a phase 1-2 basket trial reported an ORR of 18% (50% in DLL3-high tumors), with a median PFS of 3.8 months and OS of 8.1 months (90).

DLL3-targeted agents include also bispecific T-cell engagers (AMG 119, BI 764 532) and chimeric antigen receptors that activate and redirect CD3-positive T lymphocytes to DLL3-expressing tumors (AMG 757), to induce a cytotoxic immune effect. Early results are promising in patients with SCLC, and these agents are now starting to be tested in other high-grade NENs. DLL3 is also currently being explored as a target for radioligand PET imaging (Zr-89-DFO-SC16.56) and therapy (linked to actinium or lutetium) in preclinical models of SCLC (70).

Other tumor vulnerabilities that may be therapeutically exploited include cell-cycle and DDR, epigenetic regulators or MYC family dysregulations (4). Germline mutations in DDR genes have been documented in up to 20% of extrapulmonary NECs and are associated with increased sensitivity to DNA-damaging drugs such as platinum agents. Tumor cells with impaired TP53 and/or RB1 function are particularly vulnerable to cell-cycle checkpoint inhibitors (ie, Aurora kinases or cyclin-dependent kinase inhibitors) or drugs targeting the DDR pathway (ie, CHK1 or PARP inhibitors), particularly in the context of DNA-damaging therapy. Different combinations of these agents are currently being explored in patients with NECs. MYC is amplified in up to 50% of GEP NECs and drugs targeting MYCN synthetic lethal partners (ie, EZH2, Aurora kinases) have shown promising activity in preclinical models and are currently in early stages of clinical development.

Conclusions and Future Perspectives

GEP NECs are a challenging, highly aggressive type of neoplasms that commonly present widespread disease at diagnosis and have a very poor prognosis. Although increasing evidence suggests relevant molecular differences relative to their lung counterparts, SCLC-type platinum-based chemotherapy remains the standard of care. Nevertheless, a lower rate of TP53 and RB1 alterations have been documented in GEP NECs and, consistently, they seem to be less responsive to platinum agents. Novel therapeutic approaches including immunotherapy (immune-check point inhibition and others), tumor agnostic therapies targeting oncogenic drivers (RET or NTRK fusions, BRAF or KRAS mutations), or strategies targeting other tumor vulnerabilities (eg, DNA damage repair, cell-cycle dysregulations, MYC amplification, Notch pathway) may benefit some patients but comprehensive tumor molecular profiling is not widely available in standard practice. Further research is needed to consolidate these strategies and better define suitable candidates that really benefit from these therapies.

Adequate preclinical models shall be developed to support fundamental research, as an improved understanding of the underlying biology of GEP NECs and its interaction with the host is critical for progress to be made. Moreover, quality of evidence to support clinical decisions in daily practice is rather poor, and specific GEP NEC trials should be conducted considering the organ of origin, histological subtype, proliferation index, and molecular background. Clinical trials should thus incorporate comprehensive tumor molecular profiling to identify predictive biomarkers and new actionable targets to improve efficacy in the era of personalized medicine. A higher enrichment of properly characterized NEN cohorts of patients in biomarker-driven basket trials should be pursued to generate more solid evidence to support tumor agnostic therapies in these patients. To this aim, international collaboration is highly encouraged as it is essential for progress to be made.

Abbreviations

co-NEC

colorectal neuroendocrine carcinoma

DDR

DNA damage repair

dMMR

DNA mismatch repair deficiency

FDA

US Food and Drug Administration

GEP

gastroenteropancreatic

ICI

immune checkpoint inhibition

LC

large cell

MAb

monoclonal antibody

MINEN

mixed neuroendocrine-nonneuroendocrine neoplasm

MSI

microsatellite instability

NEC

neuroendocrine carcinoma

NEN

neuroendocrine neoplasm

NET

neuroendocrine tumor

NTRK

neurotrophic tyrosine receptor kinase

ORR

objective response rate

OS

overall survival

PET

positron emission tomography

PFS

progression-free survival

RLT

radioligand therapy

SCLC

small cell lung cancer

TMB-H

high tumor mutational burden

uPAR

urokinase plasminogen activator receptor

WHO

World Health Organization

Funding

M.R. is partially funded by the AECC (CLSEN19003RIES). A.L.P. is funded by the Instituto de Salud Carlos III (PFIS; FI20/00131). C.C.P. was partially funded by CAM (PEJD-2016-PRE/BMD-2666). B.R.C. was partially funded by CAM (PEJD-2017-PRE/BMD-4981). B.S. is partially funded by the AECC (POSTDO46SOLD, Spain).

Disclosures

R.G.C. has provided scientific advice and/or received honoraria or funding for continuous medical education from AAA, Advanz Pharma, Amgen, Bayer, BMS, Boerhringer, Esteve, Hutchmed, Ipsen, Merck, Midatech Pharma, MSD, Novartis, PharmaMar, Pierre Fabre, Roche, and Servier, and has received research support from Pfizer, BMS, and MSD. B.A.P. received honoraria for medical advisory from AAA, Advanz. Merck, Servier, and Novartis. A.M. received support for medical education from IPSEN, Merck, Servier, leo-pharma, and AstraZeneca. M.C.R.M. received honoraria or funding for consultant of advisory role from Bayer, AAA, Servier, and Roche, for speaking from Servier, Roche, Merk, and Bayer, and has received research support from MSD and BMS.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed for this review article.