Within the gastrointestinal tract, neuroendocrine neoplasms (NENs) are a heterogeneous family of malignancies with distinct manifestations and clinical outcomes, dependent on the site of origin, stage of disease, and pathologic grading (1). Pathologic grading uses a system elaborated in 2019 by the World Health Organization (2), which separates NENs into well-differentiated neuroendocrine tumors (NETs) and poorly differentiated neuroendocrine carcinomas (NECs). Well-differentiated NETs are further separated into grades 1 to 3 based on their replicative rate, as assessed by Ki-67 staining or mitotic index, whereas poorly differentiated NECs are universally grade 3.

The history of targeted therapies in NENs is paradoxically rife with success and failure. On the one hand, there are no genomic alterations that predict outcome with any approved standard therapy, making the utility of molecular profiling in standard practice rather modest. However, with the exception of alkylating chemotherapy for NETs arising in the pancreas (3,4), all available therapies in NETs are specific to molecular targets. In contrast, NECs lack actionable targets and are treated exclusively with cytotoxic chemotherapy as the standard of care.

Well-differentiated NETs can boast one of the oldest and best-validated targets by virtue of their somatostatin receptor (SSTR) expression. For decades, radiolabeled somatostatin analogues have been incorporated into diagnostic techniques for tumor localization (5), and receptor expression has been considered a biomarker for response (6). Although receptor expression was mandatory for inclusion in the more modern CLARINET trial of lanreotide in patients with gastroenteropancreatic NETs (GEP-NETs) (7), it was not required in the PROMID study of octreotide for midgut patients, though the majority of patients did have known SSTR expression (8). Therefore, SSTR expression as assessed with modern nuclear imaging (e.g., DOTATATE PET) is considered as supportive of, but
not required for, therapy with conventional somatostatin analogues, such as lanreotide or octreotide. In contrast, confirmation of receptor expression is required prior to peptide receptor radionuclide therapy (PRRT) such as ¹⁷⁷Lu-DOTATATE.

Although SSTR expression remains a putative predictive biomarker for conventional and radioactive somatostatin analogues, other targeted agents lack known predictors of response. Multiple inhibitors of vascular endothelial growth factor (VEGF) have been evaluated, though few of these agents have been studied in randomized trials. In pancreatic NETs, sunitinib was evaluated in a randomized study after a phase II trial using objective response rate as its primary endpoint only demonstrated volumetric reduction in NETs originating in the pancreas. Similarly, surufatinib, which has a somewhat broader spectrum of kinase inhibition, demonstrated significant reduction in the risk of progression or death when compared to best supportive care in patients with advanced, progressive panNET (SANET-p) (9). However, in a parallel study in patients with progressive extrapancreatic NETs (SANET-ep), surufatinib demonstrated similar efficacy (10), suggesting a disconnect between rates of tumor shrinkage and improvements in survival. The results of the SANET trials may not apply to the NET population in the United States given differences in the study populations, but relatively few patients with small bowel NETs were included and the patients were relatively lightly pretreated. Similarly, a randomized phase II study of pazopanib in patients with progressive extrapancreatic NET suggested improved progression-free survival compared to best supportive case, despite a 0% response rate in a single-arm phase II trial (11). Despite these clear benefits with VEGF-targeted tyrosine kinase inhibitors (TKIs), mutations in the genes of the VEGF pathway are rarely identified in NET, especially those arising outside the pancreas.

Similarly, the mechanistic target of rapamycin (mTOR) inhibitor everolimus was evaluated in a series of studies in patients with pancreatic and extrapancreatic NETs. In pancreatic tumors (12) and nonfunctional extrapancreatic tumors (13), everolimus demonstrated significant reduction in the risk of progression or death estimated to be 65% and 52%, respectively. Another randomized clinical trial of everolimus in patients with extrapancreatic NET and carcinoid syndrome suggested improved progression-free survival (PFS) in patients treated with everolimus as compared to placebo but failed to meet its prespecified endpoint of efficacy (14). However, mutations in the mTOR pathway are rare (particularly in extrapancreatic NETs [15]) and have not been associated with benefit to our knowledge.

Finally, it is worthwhile to comment on the use of immunotherapy in both well-differentiated NET and poorly differentiated NECs of the gastrointestinal tract. Studies of single-agent checkpoint inhibitors in well-differentiated NET patients have been largely disappointing, with response rates and PFS durations similar to those observed with placebo (16,17,18). Results have been similarly challenging for patients with poorly differentiated NEC (18,19), although there has been some enthusiasm based on the results of the dual anti-CTLA-4 and anti-PD-1 blockade in rare tumors (DART SWOG 1609) study (20), which suggested that the combination of ipilimumab and nivolumab could induce responses in 8 of the 18 patients with G3 NENs of any primary site treated in the study. Results from larger studies of this combination in more homogeneous patient populations are eagerly anticipated to potentially impact standard of care.