What are the Unmet Needs in Managing Digestive Neuroendocrine Carcinomas? Authors: Halfdan Sorbye, Mairéad G McNamara, Margot ET Tesselaar, Mauro Cives, Stian Knappskog, Aurel Perren, Jonathan Strosberg Published: 2026 ABSTRACT Purpose of ReviewThis review highlights the major unmet needs in diagnosis and treatment of digestive neuroendocrine carcinomas (NEC) and describes the current and evolving evidence on these aspects.Recent FindingsImproved pathological criteria are needed to accurately diagnose NEC and distinguish this entity from high-grade well-differentiated neuroendocrine tumours as well as from adenocarcinoma. Studies on the molecular mechanisms underlying the extreme aggressiveness of the disease are ongoing. Data are lacking on optimal selection of patients with locoregional disease for primary tumour surgery and use of adjuvant or neoadjuvant treatment. For patients with metastatic disease, there has been no improvement in systemic treatment over the last two decades and immunotherapy has limited benefit. Novel antigen-directed strategies targeting DLL3 are evolving.SummaryDigestive NEC are among the most aggressive cancers and prognosis remains poor. There have been only minor improvements over the last two decades. To improve diagnosis and therapy, the unmet needs discussed in this review must be addressed. Introduction Digestive neuroendocrine carcinomas (NEC) are high-grade, aggressive malignancies with a high propensity for early distant metastases and a poor prognosis, even when clinically localised [1–4]. As these are rare cancers, very limited data exist to guide diagnosis and treatment, and therefore there are several major unmet needs. Digestive NEC are frequently misdiagnosed histologically, and better criteria for diagnosing NEC and distinguishing NEC from adenocarcinomas with neuroendocrine differentiation and neuroendocrine tumours grade 3 (NET G3) are needed [5]. The molecular mechanisms behind the extreme aggressiveness of this disease are unknown, but such data are likely needed before major therapeutic improvements can evolve. The use of primary tumour surgery for locoregional disease is highly controversial, and high recurrence rates after surgery have led to recommendations advocating adjuvant and/or neoadjuvant therapy, although definitive evidence is missing [1, 2, 6]. The therapeutic benefit of palliative chemotherapy remains limited, with minimal improvement in clinical outcomes over the past two decades, and immunotherapy appears to offer only modest additional benefit. Novel therapies are however evolving, such as antigen-directed strategies targeting delta-like ligand-3 (DLL3). This review addresses the following major unmet needs for digestive NEC: Improving criteria for diagnosing NEC and distinguishing NEC from NET G3 and adenocarcinomas with neuroendocrine differentiation.Understanding the molecular mechanisms behind the aggressiveness of NEC.Role of surgery for locoregional disease.Role of adjuvant/neoadjuvant therapy for locoregional disease.Improving the benefit of palliative chemotherapy and role of targeted therapy.Use of immune checkpoint inhibitors.Role of antigen-directed strategies including T-cell engagers (TCEs), antibody-drug conjugates (ADCs) and chimeric antigen receptor T-cells (CAR-T). Improving criteria for diagnosing NEC and distinguishing NEC from NET G3 and adenocarcinomas with neuroendocrine differentiation. Improving criteria for diagnosing NEC and distinguishing NEC from NET G3 and adenocarcinomas with neuroendocrine differentiation. Understanding the molecular mechanisms behind the aggressiveness of NEC. Understanding the molecular mechanisms behind the aggressiveness of NEC. Role of surgery for locoregional disease. Role of surgery for locoregional disease. Role of adjuvant/neoadjuvant therapy for locoregional disease. Role of adjuvant/neoadjuvant therapy for locoregional disease. Improving the benefit of palliative chemotherapy and role of targeted therapy. Improving the benefit of palliative chemotherapy and role of targeted therapy. Use of immune checkpoint inhibitors. Use of immune checkpoint inhibitors. Role of antigen-directed strategies including T-cell engagers (TCEs), antibody-drug conjugates (ADCs) and chimeric antigen receptor T-cells (CAR-T). Role of antigen-directed strategies including T-cell engagers (TCEs), antibody-drug conjugates (ADCs) and chimeric antigen receptor T-cells (CAR-T). Improving Criteria for Diagnosing NEC and Distinguishing NEC from NET G3 and Adenocarcinomas with Neuroendocrine Differentiation Histopathological diagnosis of neuroendocrine neoplasms (NEN) and separation of high-grade NEN into NEC and NET G3 has critical implications for therapy [3]. Practically, histopathological diagnosis often depends on small biopsies that are exhausted quickly to determine morphological differential diagnoses. While the histopathological definition of NEN is straightforward according to the WHO classification, based on neuroendocrine morphology and expression of at least two neuroendocrine markers, neuroendocrine differentiation is not binary, but rather exists along a continuum, varying in both extent and intensity [7]. NEC may show very low expression of neuroendocrine markers [7], and adenocarcinomas expressing neuroendocrine markers may be described as adenocarcinomas with neuroendocrine differentiation (neuroendocrine features). In the presence of well-preserved adenocarcinoma morphology, a clinical course more consistent with adenocarcinoma has been reported despite high intensity of the neuroendocrine marker synaptophysin [8]. If tumour cells show both neuroendocrine and exocrine differentiation by immunohistochemistry, these tumours are designated as amphicrine carcinoma [9]. The clinical course and the diagnostic distinction between such amphicrine carcinomas, adenocarcinomas with neuroendocrine differentiation and mixed non-neuroendocrine neuroendocrine neoplasms (MiNEN) is unclear. In summary, our available set of neuroendocrine markers is of help, but the ultimate neuroendocrine marker is yet to be discovered, if it exists at all. Morphological criteria are also central for the separation of NET G3 from NEC, which is difficult even among NET experts [5]. Morphology accounts for a larger diagnostic weight than immunohistochemical or molecular markers. Judgement of morphology, however, is largely dependent on the experience of examining pathologists [10]. Artificial intelligence (AI)-augmented digital pathology may support less experienced pathologists in the future. Ancillary tests only help in a subset of cases. While immunohistochemical RB1 loss is a very strong argument for a NEC diagnosis, only 50% of digestive NEC harbour such a loss [11]. In pancreatic primaries, mutations of MEN1, DAXX and ATRX genes argue in favour of NET G3, but such a mutational analysis is futile in half of pancreatic NET and in most NET G3 of other sites due to their lack of these driver mutations [12]. DNA methylation profiles have shown the capacity to distinguish NET G3 from NEC [13–15]. These tools are not widely available, and the method has not been validated in independent cohorts and across multiple sites. Separation of NEC into small-cell and large-cell subtype is based on morphology. While this separation is reproducible in lung NEC, its application to digestive NEC is problematic as around 30% show an intermediate phenotype. A molecular subclassification into ANHPY subtypes based on transcription factor expression may be more reproducible and meaningful in the future [7]. Practically, the lack of tissue in small biopsies often limits the analytic process. Here, mutational analysis of liquid biopsies has the potential to avoid a second tissue biopsy. Liquid biopsy in high-grade NEN has shown an excellent capacity to detect driver mutations and potential therapy targets [16]. Liquid-biopsy-based methylation analysis may be a future option supporting this process. In summary, improved methods of diagnosing high-grade NEN are needed for a clear separation from adenocarcinomas with neuroendocrine differentiation and for a better distinction between NEC and NET G3. Liquid biopsies and AI-based digital pathology are two promising tools for improving diagnosis. To close the gap between the knowledge of molecular heterogeneity of NEC and clinical use of ANHPY subclassification, large retrospective analyses correlating clinical outcomes and therapy responses in these groups are needed. Understanding the Molecular Mechanisms Behind the Aggressiveness of NEC Compared to patients with adenocarcinomas at the same site, patients with NEC are much more likely to have metastatic disease at diagnosis, experience recurrence after surgery and derive less benefit from palliative chemotherapy. While new molecular data are emerging, the mechanisms underlying the aggressive phenotype of digestive NEC remain obscure. A better understanding of these mechanisms is needed to improve future treatment of NEC. The fraction of NEC with underlying pathogenic germline variants is relatively high (4.5%), compared to most cancer types [17], but most NEC cases do not seem to have germline genetics as a major underlying cause. Studies of somatic mutations have revealed key driver mutations that are largely the same as in adenocarcinomas from the same primary site. For digestive NEC, this includes driver mutations in TP53, BRAF, KRAS, APC as well as copy number alterations affecting genes like MYC, RB1 and CCNE1 [14, 18]. Thus, it seems unlikely that the phenotypical differences can be explained by somatic mutations. A recent study on MiNEN further underlined this concept, since no somatic mutations were found to cause the evolutionary split between NEC cells and adenocarcinoma cells [19]. While there are some clear differences between mutations typically found in NEC and NET G3, there is a large overlap [14, 18, 20, 21] further supporting that somatic mutations are not the main drivers of NEC phenotype. Looking beyond driver mutations, mutational signatures as footprints of mutational processes [22] have been widely used to describe underlying forces of tumour evolution. For NEC, few studies have been performed [14, 23], but data so far indicate no signatures with known aetiology [24] to be systematically enriched in NEC. Other genetic features (e.g., structural rearrangements) may be important to determine the phenotype, but this remains to be investigated in detail. Studying the transformation from NET G3 to aggressive cases with a NEC-like genetic profile will be important [25]. Notably, NEC tumours reveal high ploidy. In a study of 152 digestive NEC cases, the majority had a ploidy > 3 and seemed to have undergone whole genome duplication [18]. Furthermore, in MiNEN the ploidy is usually higher in the NEC than in the adenocarcinoma component [19]. Whether high ploidy is involved in shaping the NEC phenotype or simply represents a passive consequence of other NEC-specific features, is an open question. For future exploration, it is highly relevant to build on data from pilot cases, revealing differences in epigenetic landscapes [15] and transcriptomic programmes [26] between NEC and adenocarcinomas to decipher the cellular programmes specifically deregulated in NEC. Given that the number of NEC cases is a limiting factor in most studies, the use of MiNEN and MiNEN organoids as conceptual models for such analyses, cancelling out differences between individuals, would be highly valuable. Role of Surgery for Locoregional Digestive NEC The role of surgery in the management of patients with a locoregional digestive NEC remains uncertain. Some guidelines state that surgical resection may be considered for fit patients with localised, resectable disease, but emphasise individualised decision-making, particularly when major surgery is required [1]. Other expert groups have not reached consensus on the value of surgery but suggest that resection may be appropriate after systemic therapy or chemoradiation, especially in exceptional responders or in very early-stage disease [2]. For localised pancreatic NEC, guidelines explicitly describe the role of surgery as controversial and discourage an upfront surgical approach [6]. Unlike digestive adenocarcinomas, approximately 60–80% of digestive NEC are diagnosed when metastatic [3, 4]. Reported surgical rates for locoregional NEC vary widely: 95% for colon NEC, 60–70% for gastric, pancreatic, rectal, or gallbladder NEC, and only 25% for esophageal NEC [27]. In many cases, the diagnosis of NEC is established only after resection, as preoperative biopsies may be misleading. Retrospective studies suggest that surgery can be curative in selected patients with locoregional NEC, although recurrence rates appear higher than those observed in digestive adenocarcinomas. Series report 5-year survival rates of roughly 40% for stage II disease and around 25% for stage III disease [4]. Surgery was the only prognostic variable that significantly affected survival in stages I-II patients (HR 0.63) and showed a strong trend in stage III (HR 0.77) patients [27]. Survival after surgery varies by primary site: colon NEC show the most favourable 5-year survival (40%), while gallbladder, biliary, and esophageal NEC have the poorest outcomes (20%) [4]. For pancreatic NEC, 5-year survival after surgery has been reported to be up to 38% [28]. Other studies report a 3- to 5-year survival rate of 35–45% following surgical resection in locoregional digestive NEC [29–32]. Existing studies and current evidence have substantial limitations. Pathological re-evaluation frequently reveals misclassification, including cases initially labelled as NEC subsequently diagnosed as NET G3 [5]. NEC is often conflated with MiNEN or with adenocarcinoma with neuroendocrine differentiation. Many studies lack essential information such as Ki-67, use of adjuvant/neoadjuvant therapy or preoperative FDG-PET imaging. Outcomes are often not stratified by T-stage, N-stage, treatment modality or primary tumour site. High-quality data are needed to better define when surgery is beneficial or detrimental in patients with locoregional digestive NEC. Routine neuroendocrine marker staining in preoperative biopsies of poorly differentiated tumours may improve diagnostic accuracy. Prognostic factors should be clarified by primary site and stage. Additional data on the role of preoperative FDG-PET are required. Prospective randomised trials would be ideal to establish the role of adjuvant and neoadjuvant therapy, although feasibility remains challenging in this rare disease. Role of Adjuvant/Neoadjuvant Therapy for Locoregional Digestive NEC The high recurrence rate after primary tumour resection in patients with a locoregional digestive NEC [27], together with the substantial proportion of margin-positive resections [33] supports consideration of adjuvant and/or neoadjuvant therapy. Although international guidelines endorse these approaches [1, 2, 6], the available evidence remains limited and a better understanding of how to use these approaches is needed. Neoadjuvant chemotherapy is increasingly used in patients with resectable locoregional NEC. Its primary aims are early treatment of micro-metastatic disease, tumour downstaging and real-time assessment of chemosensitivity. The high response rates observed with platinum/etoposide regimens [3, 34] support the use of upfront systemic therapy in carefully selected patients. Responders after 2–4 cycles may proceed to surgery with curative intent, following FDG-PET exclusion of more advanced disease, whereas non-responders are generally redirected toward definitive non-surgical management. Although prospective data are lacking, retrospective studies suggest that neoadjuvant chemotherapy may improve surgical outcomes and increase the likelihood of margin-negative resection [33, 35, 36]. More evidence is needed to clarify its impact on survival and to determine progression rates during neoadjuvant treatment. As patients with localised NEC may be initially misdiagnosed as adenocarcinoma, some patients will not be considered for neoadjuvant therapy. To date, neoadjuvant targeted therapies remain investigational. TCEs targeting DLL3 represent a promising future option based on encouraging activity in the metastatic setting [37]. Chemoradiation is frequently used in patients with locoregional oesophageal or rectal NEC [27], often as part of initial treatment for a presumed adenocarcinoma [38]. Chemoradiation may serve as neoadjuvant therapy to improve local control and facilitate resection, or as definitive treatment in patients as part of organ preservation strategies or for those who are not surgical candidates or refuse surgery. Adjuvant chemotherapy is generally recommended after radical resection of locoregional NEC due to the high risk of systemic recurrence, even in early-stage disease. Platinum/etoposide regimens are most commonly used, typically four to six cycles [2, 33]. Because of the rarity of the disease, prospective studies are lacking. Most retrospective analyses suggest a benefit of adjuvant therapy, although the magnitude of benefit may vary by stage and primary tumour site [1, 2]. The effect appears greatest in patients with high-risk features such as lymph-node involvement [27]. Even patients with node-negative disease are often considered for adjuvant treatment. More data are needed to identify subgroups that derive the most benefit. The high rate of early progression in patients with metastatic NEC [3], raises the question of whether platinum/etoposide is the optimal adjuvant regimen, or whether adenocarcinoma-based regimens tailored to the primary site may be more effective. Retrospective studies are subject to bias, as patients with poorer prognostic features such as regional lymph-node disease may be more likely to receive adjuvant therapy, while those with comorbidities or poor performance status may be less likely to receive it. Although evidence is largely retrospective and extrapolated from other high-grade malignancies, combined-modality therapy appears to improve outcome in selected patients [2]. The benefit likely depends on factors such as T and N stage and primary tumour site. For oesophageal NEC, definitive chemoradiation has been associated with better outcomes than upfront surgery [39] and the value of surgery following chemoradiation remains unclear. The general benefit of chemoradiation for resectable rectal NEC not threatening the meso-rectal fascia remains uncertain. Table 1 shows possible treatment strategies for locoregional digestive NEC. Table 1Possible treatment strategies for locoregional digestive NECClinical StageStrategiesResectableUpfront surgeryFollowed by adjuvant chemotherapy (platinum/etoposide, 4–6 cycles)Consider adjuvant chemoradiation in case of R1/R2 resectionNeoadjuvant chemotherapy ±chemoradiation2–4 cycles platinum–etoposideReassess → if response: surgeryIf progression: switch to definitive chemoradiation or chemotherapy alone.Borderline resectableNeoadjuvant chemotherapy ±chemoradiationDownstaging (tumor reduction)Reassess resectabilityIf resectable: surgery UnresectablelocoregionalDefinitive chemoradiationConcurrent platinum-based chemotherapy Table 1Possible treatment strategies for locoregional digestive NECClinical StageStrategiesResectableUpfront surgeryFollowed by adjuvant chemotherapy (platinum/etoposide, 4–6 cycles)Consider adjuvant chemoradiation in case of R1/R2 resectionNeoadjuvant chemotherapy ±chemoradiation2–4 cycles platinum–etoposideReassess → if response: surgeryIf progression: switch to definitive chemoradiation or chemotherapy alone.Borderline resectableNeoadjuvant chemotherapy ±chemoradiationDownstaging (tumor reduction)Reassess resectabilityIf resectable: surgery UnresectablelocoregionalDefinitive chemoradiationConcurrent platinum-based chemotherapy Possible treatment strategies for locoregional digestive NEC Neoadjuvant chemotherapy ± chemoradiation Neoadjuvant chemotherapy ± chemoradiation Unresectable locoregional Improve the Benefit of Palliative Chemotherapy and Role of Targeted Therapy Unfortunately, no improvement in survival after palliative chemotherapy for advanced digestive NEC has been observed in the last two decades, highlighting a critical unmet need. Platinum/etoposide or platinum/irinotecan are the recommended options for treatment of patients with advanced digestive NEC in the first-line setting followed by 5-fluorouracil (5-FU)-based chemotherapy in second-line [1, 2, 34, 40, 41] (Fig. 1). A recent prospective multi-centre study that included 427 patients with advanced digestive NEC (NORDIC NEC 2) reported an OS of 7.4 months and a 3-year survival rate of 5% with palliative chemotherapy (first-line platinum/etoposide in 83% of cases) [3]. Survival was shorter than expected, likely due to the population-based design of the study, which included older patients and patients with poor performance status. In addition to performance status and age, Ki-67 ≥ 55%, elevated alkaline phosphatase and male sex were also associated with inferior survival [3]. Additional reported negative prognostic factors include presence of liver metastases and elevated lactate dehydrogenase [42]. In the NORDIC NEC 2 study, patients with a colorectal primary had worse PFS and OS [3]. It has been questioned whether platinum/etoposide is the optimal combination in patients with such a primary, especially for BRAF V600E mutated cases [43, 44]. Available evidence suggests that adenocarcinoma-targeting regimens may be effective in the treatment of digestive NEC. In a heterogeneous population of 35 patients with high-grade NEN, initial therapy with mFOLFIRINOX (5-FU/irinotecan/oxaliplatin) demonstrated an ORR of 77% [45]. In another retrospective study of 50 patients with digestive NEC (60% with colon or pancreatic primaries) treated with FOLFIRINOX (82% in later lines), median PFS and OS were 5.6 months and 8.9 months respectively [46]. Whether similar efficacy outcomes and improved tolerability may be achieved by doublet therapy instead of triplet therapy is currently unclear. A PFS of 3.5 months and OS of 8.9 months have been reported with 5-FU/irinotecan in the second-line setting [41]. The NORDIC NEC 2 study found primary resistance to palliative chemotherapy in 41% of patients with digestive NEC [3]. Molecular predictive factors for digestive NEC have been difficult to identify [47]. Expression of the retinoblastoma protein (Rb) has been suggested to be a candidate, but in a multicentre study including 185 patients (66% with Rb status) with advanced NEC (including lung primaries), it had no prognostic impact in patients treated with platinum/etoposide [48]. In small-cell lung cancer, the lung counterpart of digestive NEC, MYC expression has been reported to be elevated following the acquisition of platinum resistance [49]. Five molecular subtypes of NEC have been identified. They are defined by the transcriptional regulators ASCL1, NEUROD1, HNF4A, POU2F3, and YAP1, and show distinct molecular and clinical hallmarks [50]. The H subtype was HNF4A-dominated, and exhibited a gastrointestinal-like signature, wild-type RB1, and prevalent large-cell morphology. Interestingly, subtype H demonstrated unique neuroendocrine differentiation, with differential expression of CHGA/CD56, activation of the MYC pathway and mixed expression pattern of neuroendocrine and non-neuroendocrine genes, along with decreased sensitivity to platinum/etoposide therapy. Innovative trial designs incorporating individual NEC subtypes are eagerly awaited; their feasibility remains an issue. Tumour agnostic therapies targeting oncogenic drivers such as RET, NTRK fusions, BRAF or RAS mutations may be relevant for digestive NEC [51, 52]. Colon NEC are particularly enriched in BRAF V600E mutations and may benefit from BRAF inhibition and the addition of EGFR-inhibitors [53]. In BRAF V600E mutated colorectal NEC, BRAF splicing variants activate the RAF/MEK/ERK pathway contributing to acquired BRAF inhibitor resistance, therefore MEK or ERK may be potential targets to overcome BRAF resistance [54]. The BRAF-inhibitor encorafenib has been approved in combination with FOLFOX and cetuximab as first-line treatment for patients with any metastatic colorectal primary with BRAF V600E mutation [55]. Moving forward, liquid biopsy may hold promise as a less invasive technique for diagnosis, prognosis, response prediction, follow-up, and molecular analysis in digestive NEC [16, 56, 57]. Fig. 1Advanced digestive neuroendocrine carcinomas: unmet needs and treatment management Advanced digestive neuroendocrine carcinomas: unmet needs and treatment management Use of Immune Checkpoint Inhibitors in Digestive NEC The expectations for use of immunotherapy for digestive NEC were high, as addition of ICI to chemotherapy improves overall survival among patients with small-cell lung cancer [58]. In a large US cohort of metastatic digestive NEC, the percentage of patients receiving immunotherapy increased from 12% in 2018 to 34% in 2022 mainly given alongside chemotherapy [59]. However, ICI seem to have limited activity in unselected patients with advanced digestive NEC. In a report analysing two phase 2 trials of single-agent pembrolizumab, only 1/29 patients (3%) with extrapulmonary NEN G3 achieved an objective response [60]. A phase 2 trial of spartalizumab, a PD-1 inhibitor, included 21 patients with digestive NEC; PFS was 1.8 months with one response (5%) in a tumour with PD-L1 expression < 1% [61]. Early data on combination ICI (dual anti-CTLA-4 and anti-PD-1 blockade) appeared more promising. A study of ipilimumab and nivolumab (ipi/nivo) included a heterogeneous cohort of 19 patients with NEN G3 [62]. Five patients (26%) achieved a partial response; among them four digestive NEC. PFS was 2 months. Subsequent studies have shown lower response rates. The NIPINEC trial randomised two cohorts to receive nivolumab monotherapy or ipi/nivo [63]. ORR was low among 84 patients with advanced digestive NEC: 7% for nivolumab and 14% for ipi/nivo with median response duration of 22.3 months. PFS was 1.8 months and OS < 4 months in the digestive NEC cohort. The DUNE trial evaluated a combination of durvalumab (PD-L1 inhibitor) and tremelimumab (CTLA-4 inhibitor) in various NEN cohorts [64]. No correlation between PD-L1 expression and efficacy was observed. Among 18 patients with digestive NEC, 3 responses were reported (17%). A retrospective study of ipi/nivo in 34 patients with extrapulmonary NEN G3 reported ORR 15%, PFS 1 month and OS 5 months [65]. Among the 5 responders, duration of response was < 5 months in 3 patients. Combinations of ICI with other therapies have yielded results which are difficult to interpret. The phase 2 NICE-NEC trial evaluated nivolumab plus carboplatin/etoposide in chemo-naïve patients with NEN G3 of digestive or unknown primary [66]. The ORR was 57% with PFS 5.7 months and OS 13.9 months. Although these outcomes were considered encouraging, it is difficult to assess the contribution of nivolumab in a single-arm trial. A randomised phase 2 trial of platinum/etoposide with or without atezolizumab as first-line therapy for extrapulmonary NEC is currently accruing (NCT05058651). In the refractory setting, a single-arm study of ipi/nivo and cabozantinib in patients with extrapulmonary NEC reported PFS 2.7 months and responses in 2/16 patients (12.5%), both durable [67]. In summary, responses to single agent ICI therapy appear to be < 10%, whereas responses to combination ICI are modestly higher (approximately 15%). Although durable responses are occasionally observed, the PFS and OS in the refractory setting are exceptionally short suggesting that the impact of ICI is limited primarily to patients who achieve an objective response. At this time, there is insufficient information to suggest that particular primary sites or PD-L1 tumoral expression is predictive of response. For cases with microsatellite instability (MSI) (approximately 5% of total) or potentially patients with high tumour mutation burden, ICI is a reasonable option. There is insufficient information to support combination of ICI with other therapies. Future randomised data may indicate whether ICI in combination with first-line platinum and etoposide, improves OS among patients with digestive NEC. Role of Antigen-Directed Strategies Including T-Cell Engagers, Antibody-Drug Conjugates and Chimeric Antigen Receptor T Cells in Digestive NEC Novel antigen-directed strategies including TCEs, ADCs and CAR-T cells are emerging therapeutic options for patients with digestive NEC. While DLL3 is the most mature target, early-phase development is exploring agents targeting delta-like non-canonical Notch ligand-1 (DLK1), seizure-related homolog protein-6 (SEZ6) and B7 homolog-3 (B7-H3). DLL3 is an inhibitory Notch ligand aberrantly expressed on the surface of many neuroendocrine cancers, with limited expression in normal tissues [68]. DLL3 is expressed in two-thirds to three-quarters of digestive NEC, with enrichment in small-cell morphology tumours [69–72] and substantial concordance between primary tumours and metastases [69, 70]. Several bispecific (Obrixtamig, Tarlatamab) and trispecific (Gocatamig, Alveltamig) DLL3-directed TCEs have entered clinical development. The DLL3xCD3 TCE Obrixtamig (BI-764532) has generated the most compelling evidence of activity in digestive NEC to date. In a phase I study including 60 patients with extrapulmonary NEC, 45% had digestive primaries [73]. Activity of Obrixtamig was strongly associated with DLL3 expression, with a 40% ORR in DLL3-high tumours (≥ 50% positive tumour cells) versus 3.3% with DLL3-low tumours, and corresponding median duration of response of 7.9 and 2.8 months [37]. Notably, the ORR reached 50% in patients with DLL3-high digestive NEC. In the first-line DAREON-7 trial, Obrixtamig plus platinum/etoposide achieved an ORR of 75%, 89% DCR and PFS 7.7 months among 43 patients with DLL3-positive NEC, most with digestive primaries [74]. When combined with the PD-1 inhibitor Ezabenlimab, Obrixtamig also demonstrated activity in patients with pretreated DLL3-positive NEC [75]. Among 22 patients with extrapulmonary NEC, ORR was 29%, DCR 53% and PFS 3.9 months. Across trials, Obrixtamig showed a manageable safety profile, with cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) of mild-to-moderate grade. Obrixtamig is currently being evaluated as monotherapy in relapsed/refractory DLL3-high extrapulmonary NEC (DAREON-5) and in combination with first-line platinum/etoposide in a randomised phase III trial (DAREON-NEC-1). Clinical evidence for Tarlatamab remains preliminary in patients with digestive NEC. A complete radiographic response has been reported in a patient with metastatic oesophageal small-cell NEC [76]. In a retrospective series of 25 patients with high-grade NEN other than small-cell lung cancer, the ORR was 48%, increasing to 67% among DLL3-positive tumours [77]. Prospective evaluation of Tarlatamab in digestive NEC is underway (NCT06893783/NCT06816394/NCT07016230/NCT07061080). The phase III TARLANEC trial (NCT06937905) compares Tarlatamab with investigator-choice chemotherapy in pretreated DLL3-positive NEC. Gocatamig is a trispecific TCE that binds DLL3, CD3 and albumin, the latter providing half-life extension. In the phase I/II study, Gocatamig achieved a 46% ORR in 37 patients with pretreated non-prostate high-grade neuroendocrine malignancies [78]. Alveltamig, a biparatopic DLL3xDLL3xCD3 TCE, produced a 37.5% ORR in a randomised phase II study of 64 patients with pretreated extrapulmonary NEC; 11/19 responders had digestive primaries [79]. In DLL3-high tumours, ORR and PFS reached 56.3% and 8.4 months, supporting DLL3 expression as a therapeutic target and a predictive biomarker. A randomised phase III study (NCT07595913) is evaluating Alveltamig versus investigator-selected chemotherapy in previously treated patients with DLL3-positive extrapulmonary NEC. Clinical development of ADCs remains substantially less mature for digestive NEC. The first DLL3-directed ADC, Rovalpituzumab tesirine, demonstrated limited clinical activity and considerable toxicity in patients with NEC; its clinical development has been therefore discontinued [80]. SEZ6-directed topoisomerase-I ADC ABBV-706 demonstrated objective responses in 2/12 patients with digestive NEC (ORR 16.7%), providing proof-of-concept for ADC therapy in this population [81]. Next-generation ADCs include the DLL3-directed topoisomerase-I ADC IDE849, currently undergoing phase I/II evaluation in DLL3-positive malignancies (NCT07174583), and the B7-H3-directed ADC MHB088C in a phase II study of platinum-pretreated extrapulmonary NEC (NCT06951243). CAR-T therapy remains at an early stage of development in digestive NEC, with no published disease-specific data to date. The first-in-human SPECTRAL-1 trial (NCT07488923) is currently evaluating the potency-enhanced anti-DLL3 autologous CAR-T product ML261 in patients with relapsed/refractory malignancies. A phase I trial (NCT07509034) is evaluating autologous B7-H3 CAR-T cells in patients with small-cell lung cancer and extrapulmonary NEC. Conclusions Digestive NEC remain one of the most aggressive cancers. There have been minimal improvements in outcomes over the last two decades. There is a large unmet need to improve diagnosis and therapy. Improved pathological criteria for diagnosing NEC and distinguishing NEC from NET G3 are needed. A greater understanding of the molecular mechanisms underlying the aggressiveness of the disease is crucial. There is a paucity of data on how to select patients with locoregional disease for primary tumour surgery, adjuvant, and neoadjuvant treatment. For patients with metastatic disease, there has been no improvement in systemic treatment over the last two decades and immunotherapy has limited benefit. Novel antigen-directed strategies such as those targeting DLL3 are evolving. International collaborative prospective trials are warranted. Key References *Wang Z, Liu C, Zheng S, Yao Y, Wang S, Wang X, et al. Molecular subtypes of neuroendocrine carcinomas: A cross-tissue classification framework based on five transcriptional regulators. Cancer Cell. 2024;42(6):1106–25 e8.○ This large analysis of over 1000 NEC cases (31 primaries) revealed five intrinsic subtypes highlighting differing clinical and pathological presentations, genomic alterations, evolution, and response to treatment.*Venizelos A, Elvebakken H, Perren A, Nikolaienko O, Deng W, Lothe IMB, et al. The molecular characteristics of high-grade gastroenteropancreatic neuroendocrine neoplasms. Endocrine-related cancer. 2021;29(1):1-14.○ This study performed molecular characterisation of 152 digestive NEC, including the landscape of driver mutations, copy number changes and ploidy, and a comparison with NET G3.**Wermke M, Gambardella V, Kuboki Y, Felip E, Sanmamed MF, Alese OB, et al. Phase I Dose-Escalation Results for the Delta-Like Ligand 3/CD3 IgG-Like T-Cell Engager Obrixtamig (BI 764532) in Patients With Delta-Like Ligand 3+ Small Cell Lung Cancer or Neuroendocrine Carcinomas. J Clin Oncol. 2025;43(27):3021–31.○ This phase I dose-escalation study established the feasibility and preliminary antitumor activity of the DLL3×CD3 T-cell engager obrixtamig (BI 764532) in patients with heavily pretreated DLL3-positive small cell lung cancer, large cell pulmonary neuroendocrine carcinoma and extrapulmonary neuroendocrine carcinoma.**Capdevila J GV, Kuboki Y, Alese OB, Morgensztern D, Sayehli C, et al. Efficacy and safety of the DLL3/CD3 T-cell engager obrixtamig in patients with extrapulmonary neuroendocrine carcinomas with high or low DLL3 expression: Results from an ongoing phase I trial. J Clin Oncol. 2026;43(16S):3004.○ This study shows that activity of Obrixtamig was strongly associated with DLL3 expression, with a 40% ORR in DLL3-high tumours (≥50% positive tumour cells) versus 3.3% with DLL3-low tumours, and corresponding median duration of response of 7.9 and 2.8 months.*Sorbye H, Hjortland GO, Vestermark LW, Ladekarl M, Svensson J, Sundlöv A, et al. Characteristics and treatment outcome in a prospective cohort of 639 advanced high-grade digestive neuroendocrine neoplasms (NET G3 and NEC). The NORDIC NEC 2 study. Br J Cancer 2025:133(3):316–324.○ This large prospective cohort establishes the distinct clinical behaviour and treatment outcomes of advanced digestive NEC, highlighting the importance of accurate pathological classification and analysing predictive and prognostic factors for survival.*Walter TP, Mazieres J, Otto J, Lena H, Lepage C, Smith D, et al. Nivolumab +/- Ipilimumab in Patients With Pretreated Advanced Neuroendocrine Carcinoma: The GCO-001 NIPINEC Randomized Phase II Trial. J Clin Oncol. 2026;44(12):1120–9.○ This study shows that benefit of ICI is limited in unselected advanced digestive NEC. ORR was low: 7% for nivolumab and 14% for ipi/nivo with median response duration of 22.3 months. PFS was 1.8 months and OS <4 months. *Wang Z, Liu C, Zheng S, Yao Y, Wang S, Wang X, et al. Molecular subtypes of neuroendocrine carcinomas: A cross-tissue classification framework based on five transcriptional regulators. Cancer Cell. 2024;42(6):1106–25 e8.○ This large analysis of over 1000 NEC cases (31 primaries) revealed five intrinsic subtypes highlighting differing clinical and pathological presentations, genomic alterations, evolution, and response to treatment. *Wang Z, Liu C, Zheng S, Yao Y, Wang S, Wang X, et al. Molecular subtypes of neuroendocrine carcinomas: A cross-tissue classification framework based on five transcriptional regulators. Cancer Cell. 2024;42(6):1106–25 e8.○ This large analysis of over 1000 NEC cases (31 primaries) revealed five intrinsic subtypes highlighting differing clinical and pathological presentations, genomic alterations, evolution, and response to treatment. ○ This large analysis of over 1000 NEC cases (31 primaries) revealed five intrinsic subtypes highlighting differing clinical and pathological presentations, genomic alterations, evolution, and response to treatment. ○ This large analysis of over 1000 NEC cases (31 primaries) revealed five intrinsic subtypes highlighting differing clinical and pathological presentations, genomic alterations, evolution, and response to treatment. *Venizelos A, Elvebakken H, Perren A, Nikolaienko O, Deng W, Lothe IMB, et al. The molecular characteristics of high-grade gastroenteropancreatic neuroendocrine neoplasms. Endocrine-related cancer. 2021;29(1):1-14.○ This study performed molecular characterisation of 152 digestive NEC, including the landscape of driver mutations, copy number changes and ploidy, and a comparison with NET G3. *Venizelos A, Elvebakken H, Perren A, Nikolaienko O, Deng W, Lothe IMB, et al. The molecular characteristics of high-grade gastroenteropancreatic neuroendocrine neoplasms. Endocrine-related cancer. 2021;29(1):1-14.○ This study performed molecular characterisation of 152 digestive NEC, including the landscape of driver mutations, copy number changes and ploidy, and a comparison with NET G3. ○ This study performed molecular characterisation of 152 digestive NEC, including the landscape of driver mutations, copy number changes and ploidy, and a comparison with NET G3. ○ This study performed molecular characterisation of 152 digestive NEC, including the landscape of driver mutations, copy number changes and ploidy, and a comparison with NET G3. **Wermke M, Gambardella V, Kuboki Y, Felip E, Sanmamed MF, Alese OB, et al. Phase I Dose-Escalation Results for the Delta-Like Ligand 3/CD3 IgG-Like T-Cell Engager Obrixtamig (BI 764532) in Patients With Delta-Like Ligand 3+ Small Cell Lung Cancer or Neuroendocrine Carcinomas. J Clin Oncol. 2025;43(27):3021–31.○ This phase I dose-escalation study established the feasibility and preliminary antitumor activity of the DLL3×CD3 T-cell engager obrixtamig (BI 764532) in patients with heavily pretreated DLL3-positive small cell lung cancer, large cell pulmonary neuroendocrine carcinoma and extrapulmonary neuroendocrine carcinoma. **Wermke M, Gambardella V, Kuboki Y, Felip E, Sanmamed MF, Alese OB, et al. Phase I Dose-Escalation Results for the Delta-Like Ligand 3/CD3 IgG-Like T-Cell Engager Obrixtamig (BI 764532) in Patients With Delta-Like Ligand 3+ Small Cell Lung Cancer or Neuroendocrine Carcinomas. J Clin Oncol. 2025;43(27):3021–31.○ This phase I dose-escalation study established the feasibility and preliminary antitumor activity of the DLL3×CD3 T-cell engager obrixtamig (BI 764532) in patients with heavily pretreated DLL3-positive small cell lung cancer, large cell pulmonary neuroendocrine carcinoma and extrapulmonary neuroendocrine carcinoma. ○ This phase I dose-escalation study established the feasibility and preliminary antitumor activity of the DLL3×CD3 T-cell engager obrixtamig (BI 764532) in patients with heavily pretreated DLL3-positive small cell lung cancer, large cell pulmonary neuroendocrine carcinoma and extrapulmonary neuroendocrine carcinoma. ○ This phase I dose-escalation study established the feasibility and preliminary antitumor activity of the DLL3×CD3 T-cell engager obrixtamig (BI 764532) in patients with heavily pretreated DLL3-positive small cell lung cancer, large cell pulmonary neuroendocrine carcinoma and extrapulmonary neuroendocrine carcinoma. **Capdevila J GV, Kuboki Y, Alese OB, Morgensztern D, Sayehli C, et al. Efficacy and safety of the DLL3/CD3 T-cell engager obrixtamig in patients with extrapulmonary neuroendocrine carcinomas with high or low DLL3 expression: Results from an ongoing phase I trial. J Clin Oncol. 2026;43(16S):3004.○ This study shows that activity of Obrixtamig was strongly associated with DLL3 expression, with a 40% ORR in DLL3-high tumours (≥50% positive tumour cells) versus 3.3% with DLL3-low tumours, and corresponding median duration of response of 7.9 and 2.8 months. **Capdevila J GV, Kuboki Y, Alese OB, Morgensztern D, Sayehli C, et al. Efficacy and safety of the DLL3/CD3 T-cell engager obrixtamig in patients with extrapulmonary neuroendocrine carcinomas with high or low DLL3 expression: Results from an ongoing phase I trial. J Clin Oncol. 2026;43(16S):3004.○ This study shows that activity of Obrixtamig was strongly associated with DLL3 expression, with a 40% ORR in DLL3-high tumours (≥50% positive tumour cells) versus 3.3% with DLL3-low tumours, and corresponding median duration of response of 7.9 and 2.8 months. ○ This study shows that activity of Obrixtamig was strongly associated with DLL3 expression, with a 40% ORR in DLL3-high tumours (≥50% positive tumour cells) versus 3.3% with DLL3-low tumours, and corresponding median duration of response of 7.9 and 2.8 months. ○ This study shows that activity of Obrixtamig was strongly associated with DLL3 expression, with a 40% ORR in DLL3-high tumours (≥50% positive tumour cells) versus 3.3% with DLL3-low tumours, and corresponding median duration of response of 7.9 and 2.8 months. *Sorbye H, Hjortland GO, Vestermark LW, Ladekarl M, Svensson J, Sundlöv A, et al. Characteristics and treatment outcome in a prospective cohort of 639 advanced high-grade digestive neuroendocrine neoplasms (NET G3 and NEC). The NORDIC NEC 2 study. Br J Cancer 2025:133(3):316–324.○ This large prospective cohort establishes the distinct clinical behaviour and treatment outcomes of advanced digestive NEC, highlighting the importance of accurate pathological classification and analysing predictive and prognostic factors for survival. *Sorbye H, Hjortland GO, Vestermark LW, Ladekarl M, Svensson J, Sundlöv A, et al. Characteristics and treatment outcome in a prospective cohort of 639 advanced high-grade digestive neuroendocrine neoplasms (NET G3 and NEC). The NORDIC NEC 2 study. Br J Cancer 2025:133(3):316–324.○ This large prospective cohort establishes the distinct clinical behaviour and treatment outcomes of advanced digestive NEC, highlighting the importance of accurate pathological classification and analysing predictive and prognostic factors for survival. ○ This large prospective cohort establishes the distinct clinical behaviour and treatment outcomes of advanced digestive NEC, highlighting the importance of accurate pathological classification and analysing predictive and prognostic factors for survival. ○ This large prospective cohort establishes the distinct clinical behaviour and treatment outcomes of advanced digestive NEC, highlighting the importance of accurate pathological classification and analysing predictive and prognostic factors for survival. *Walter TP, Mazieres J, Otto J, Lena H, Lepage C, Smith D, et al. Nivolumab +/- Ipilimumab in Patients With Pretreated Advanced Neuroendocrine Carcinoma: The GCO-001 NIPINEC Randomized Phase II Trial. J Clin Oncol. 2026;44(12):1120–9.○ This study shows that benefit of ICI is limited in unselected advanced digestive NEC. ORR was low: 7% for nivolumab and 14% for ipi/nivo with median response duration of 22.3 months. PFS was 1.8 months and OS <4 months. *Walter TP, Mazieres J, Otto J, Lena H, Lepage C, Smith D, et al. Nivolumab +/- Ipilimumab in Patients With Pretreated Advanced Neuroendocrine Carcinoma: The GCO-001 NIPINEC Randomized Phase II Trial. J Clin Oncol. 2026;44(12):1120–9.○ This study shows that benefit of ICI is limited in unselected advanced digestive NEC. ORR was low: 7% for nivolumab and 14% for ipi/nivo with median response duration of 22.3 months. PFS was 1.8 months and OS <4 months. ○ This study shows that benefit of ICI is limited in unselected advanced digestive NEC. ORR was low: 7% for nivolumab and 14% for ipi/nivo with median response duration of 22.3 months. PFS was 1.8 months and OS <4 months. ○ This study shows that benefit of ICI is limited in unselected advanced digestive NEC. ORR was low: 7% for nivolumab and 14% for ipi/nivo with median response duration of 22.3 months. PFS was 1.8 months and OS <4 months.