The Evolving Treatment Algorithm for Advanced Neuroendocrine Neoplasms: Diversity and Commonalities Across Tumor Types

LCNEC and PD-NEC We present a conceptual treatment schematic for systemic therapy based on tumor classi ﬁ cation for advanced NEN (Fig. There is currently no established standard treatment for the new classi ﬁ cation of G3 WD-NET, which have a signi ﬁ cantly worse response to platinum-based chemotherapy than G3 PD-NEC [15]. Systemic treatments for tumor control in G2 WD-NET should also be considered for G3 WD-NET, given their high SSTR2 expression [19] and consistent molecular pro- ﬁ le [32]. An analysis of sunitinib ef ﬁ cacy in a phase II trial reported preliminary evidence of activity in G3 WD-NET [57], indicating that targeted therapies may be appropriate in this clinical situation. Prospective randomized trials are required to determine the effectiveness of therapies for these patients.


INTRODUCTION
Neuroendocrine neoplasms (NEN) most commonly arise in the gastroenteropancreatic (GEP) system and lungs [1,2]. The term NEN encompasses both well-differentiated (WD) neuroendocrine tumors (NET) and poorly differentiated (PD) neuroendocrine carcinomas (NEC). Many symptoms associated with NEN are nonspecific, such as abdominal pain, diarrhea, flushing, and bowel obstruction, which may lead to delays in diagnosis [3]. The rates of NEN diagnosed each year have increased steadily since the early 1970s [1,2,4], possibly as a result of improvements in diagnostic techniques and the increased detection of asymptomatic early-stage disease [1].
NEN are currently classified by functional status, tumor cell differentiation and grade, somatostatin receptor (SSTR) expression, and, more recently, their genetic alterations. In this review, we present recent epidemiologic data and summarize the biologic characteristics of NEN. We also review pivotal clinical evidence regarding systemic treatments for advanced disease and provide a conceptual schematic of systemic treatments based on tumor classification for advanced NEN.

EPIDEMIOLOGY
Although the total incidence of malignant neoplasms has remained relatively stable over the past 40 years, the incidence of NEN has increased-by 6.4-fold between 1973 and 2012, from 1.09 to 6.98 per 100,000 persons [1]. This increase in incidence occurred across all sites and stages, although localized disease was markedly increased, from 0.21 to 3.15 per 100,000 persons [1]. The greatest increases in incidence for individual primary tumor sites occurred in the stomach (15-fold) and rectum (9-fold); the increased incidence at these sites may be associated with the increased use of endoscopic procedures in clinical practice [1]. The prevalence of NEN has also increased significantly in the past 20 years, from 0.006% in 1993 to 0.048% in 2012, which may be explained by the increased incidence of NEN and the increased diagnosis of early-stage disease [1]. For NEN diagnosed in the U.S. between 2000 and 2012, the highest incidences were for lung carcinoid tumors (1.49 per 100,000 persons), gastroenteropancreatic (GEP) sites (3.56 per 100,000), and unknown primary site (0.84 per 100,000) [1]. The small intestine (1.05 per 100,000 persons) and rectum (1.04 per 100,000) were the most common GEP primary sites, and the pancreas had an incidence of 0.48 per 100,000 [1].

FUNCTIONAL STATUS
Neuroendocrine cells can produce a variety of amine and peptide secretory products, which may produce clinical syndromes in NEN [6]. Classification of NEN as functional or nonfunctional is based on the presence or absence of symptoms related to these hypersecretory syndromes [7].
Carcinoid syndrome results from hypersecretion of vasoactive amines, including serotonin, histamine, tachykinins, and prostaglandins [7]. The cardinal features of carcinoid syndrome are diarrhea (frequency 60%-80%) and flushing (frequency 60%-85%), and a severe complication of carcinoid syndrome is carcinoid heart disease (frequency ≤20%). Carcinoid syndrome is a common functional syndrome in WD-NET originating in the ileum [8]. A recent analysis of the SEER registry linked with Medicare claims data identified 1,786 patients diagnosed with both NET and carcinoid syndrome between 2000 and 2011 in the United States; 19% of these patients received a diagnosis of carcinoid syndrome within 6 months of being diagnosed with NET [9]. Carcinoid syndrome was most common in patients with WD G1/2, distant metastatic disease originating in the small intestine (56%; 242/436) or the cecum (52%; 28/54).
Carcinoid syndrome occurring in patients with a pancreatic primary tumor is very rare [10], owing to the absence of serotonin-producing cells in the normal pancreas [9].

CELL DIFFERENTIATION AND GRADING
Grading of GEP NEN is assessed by mitotic rate and/or Ki-67 index [13]. According to the World Health Organization (WHO) classification, NEN are separated into WD and PD tumors based on their cell morphology [13,14]. PD tumors are further classified as small-cell or large-cell morphology [13]. Previously, the differentiation of GEP NEN was assumed to be directly correlated with grade-all low-to intermediate-grade (G1 or G2) NEN were classified as WD-NET, and all high-grade (G3) NEN were classified as PD-NEC [13]. However, G3 NEN are now known to be heterogeneous, comprising both WD-NET with elevated proliferative index (Ki-67 >20%) and PD-NEC ( Fig. 1A) [15,16]. Accordingly, a new category of G3 WD-NET has been recognized in the WHO classification of tumors of endocrine organs, which refers specifically to pancreatic NEN [17]. Although these tumors have a somewhat worse prognosis than G2 WD-NET, they are less aggressive than G3 PD-NEC [17]. Distinguishing G3 WD-NET from G3 PD-NEC based on cell morphology can be difficult and may require correlation with clinical history, imaging findings, and molecular phenotype [18,19].
Preliminary evidence suggests that lung NEN, like GEP NEN, may include G3 WD-NET with a Ki-67 index exceeding 20% (or mitotic count >10 per 2 mm 2 ), and these patients have a poorer prognosis than those with G2 WD-NET.

GENETIC LANDSCAPE
Most NEN are sporadic; fewer than 5% of GEP NEN have inherited mutations [6]. No activating mutations have been identified, with inherited mutations resulting in a loss of tumor suppressor genes. Somatic alterations are identified in 1%-50% of GEP NEN; these typically involve mutations, loss of heterozygosity, and chromosomal changes that lead to activated signaling pathways such as PI3K/AKT/mTOR [6]. In pancreatic NET, the MEN1 gene is somatically inactivated in 45% of patients, alterations in mTOR pathway genes are evident in 15% [17], and loss of DAXX or ATRX occurs in 43% [29]. The loss of DAXX/ATRX is associated with chromosome instability and reduced survival [30]. The genetic alterations in pancreatic PD-NEC are considerably different from those in pancreatic WD-NET; there are often mutations in TP53 and RB1, whereas MEN1, ATRX, and DAXX are not altered [17,31]. Conversely, RB1 and TP53 mutations have not been identified in pancreatic WD-NET [31]. The genetic landscape of pancreatic G3 WD-NET is more similar to that of G1/G2 NET than to PD-NEC [18]. It has been proposed that G3 WD-NET arises from progression of G1/G2 WD-NET and is a biologically different entity from G3 PD-NEC (Fig. 1B) [32]. Differential diagnosis of G3 WD-NET from G3 PD-NEC may be assisted by application of molecular markers TP53, RB1, ATRX, and DAXX [18] in conjunction with Ki-67 proliferation index and SSTR 2A expression [19].  Evolving Treatment Algorithm for Advanced NEN Lung carcinoid tumors display a very low somatic mutation rate (0.4 per Mbps), whereas SCLC and LCNEC have the highest rates (>7 per Mbps) in human tumors [20]. TP53 and RB1 mutation and inactivation are rare in TC (<5%) but more frequent in AC (20%). TC and AC may contain MEN1 mutations in sporadic cases (40%). SCLC and LCNEC are driven by inactivating mutations in the RB and TP53 genes and never harbor MEN1 somatic mutations [20]. Although double genetic inactivation of TP53 and RB is universal in SCLC [20], it is not in LCNEC, with TP53 altered in 78% of cases and RB1 in 38% [33]. Current data support the notion that lung carcinoid tumors are genetically distinct from high-grade lung tumors and are therefore not early progenitors of SCLC or LCNEC, whereas TC and AC have no genetic segregation and appear to derive from the same progenitor [20]. Molecular markers for WD G3 lung NET have not been examined at this time.

PIVOTAL CLINICAL EVIDENCE: TREATMENTS FOR TUMOR CONTROL IN ADVANCED NET
An overview of the patient characteristics and primary endpoint results from each prospective randomized clinical trial are provided in Table 1 [34][35][36][37][38][39][40]. Two key studies have demonstrated the antiproliferative effect of somatostatin analogues (SSAs) in advanced GEP NET: PROMID and CLARINET. The PROMID trial showed a significant prolongation in the median time to tumor progression with long-acting octreotide 30 mg every 28 days (q28d) versus placebo (14.3 vs. 6.0 months, HR = 0.34, 95% CI = 0.20-0.59, p = .000072) in patients with midgut NET. Patients with functional or nonfunctional tumors had similar responses, whereas the antiproliferative response was more pronounced in those with a low hepatic tumor load (≤10%) (p = .0023) [34]. The CLARINET trial demonstrated a significant prolongation in median progression-free survival (PFS) with lanreotide 120 mg q28d versus placebo (not reached vs. 18.0 months, HR = 0.47, 95% CI = 0.30-0.73, p < .001) [35]. Patients appeared to experience a PFS benefit from lanreotide treatment regardless of hepatic tumor volume or grade. Hazard ratios for disease progression in patients with primary tumor sites in the midgut (0.35, 95% CI = 0.16-0.80) and pancreas (0.58, 95% CI = 0.32-1.04) also appeared to favor lanreotide over placebo; however, the benefit of lanreotide for patients with a hindgut primary NET is unclear due to the small sample size of this subpopulation (n = 14) [35].
Everolimus is approved for the treatment of advanced, progressive, nonfunctional NET of the pancreas, GI tract, and lung [42,43], due to positive results from RADIANT-3 and RADIANT-4; a significant benefit in median PFS was demonstrated with everolimus in patients with advanced progressive pancreatic NET in RADIANT-3 (11.0 vs. 4.6 months, HR = 0.35, 95% CI = 0.27-0.45; p < .001) [37] and in patients with advanced progressive nonfunctional NET of the GI tract or lung in RADIANT-4 (11.0 vs. 3.9 months, HR = 0.48, 95% CI = 0.35-0.67, p < .00001) [38]. In RADIANT-4, the PFS benefit with everolimus compared with placebo was observed regardless of tumor grade or primary site [38]. Furthermore, health-related quality of life was preserved with everolimus treatment versus placebo [44].
RADIANT-2 compared everolimus and placebo, both in combination with long-acting octreotide 30 mg q28d, in patients with progressive, low-or intermediate-grade NET and a history of secretory symptoms attributed to carcinoid syndrome [36]. This trial did not lead to the approval of everolimus plus long-acting octreotide for functional NET. RADIANT-2 demonstrated a PFS benefit with everolimus versus placebo (median 16.4 vs. 11.3 months, HR = 0.77, 95% CI = 0.59-1.00, p = .026); however, the prespecified significance boundary of p = .0246 was not met [36]. A subgroup analysis indicated a potentially higher PFS benefit for everolimus in patients with NET originating in the colon (median 29.9 vs. 13.0 months, HR = 0.39) in comparison with those with NET originating in the small intestine (median 18.6 vs. 14.0 months, HR = 0.77) [36]. A subsequent analysis of the colorectal subgroup (n = 39) in RADIANT-2 confirmed this benefit in patients with hindgut NET, with a significant prolongation in median PFS with everolimus versus placebo (29.9 vs. 6.6 months, HR = 0.34, 95% CI = 0.13-0.89, p = .011) [45]. These findings suggest that the overall PFS benefit provided by everolimus may be lower in patients with slow-growing midgut NET and higher in those with the more aggressive hindgut NET. A RADIANT-4 subgroup analysis also provides support for a higher PFS benefit with everolimus in patients with nonmidgut NET (median 8.  RADIANT-3 [37] RADIANT-4 [38] Sunitinib phase III [39] NETTER-1 [40] (n = 85)

© AlphaMed Press 2018
Evolving Treatment Algorithm for Advanced NEN Treatment with capecitabine combined with temozolomide chemotherapy has shown efficacy in recent retrospective clinical studies of pancreatic NET, with an objective response rate (ORR) ranging from 53% to 70% and a median PFS of 16.5 to >18 months [49]. The efficacy of this treatment approach is undergoing validation in a phase II study of capecitabine + temozolomide versus temozolomide monotherapy in patients with advanced low-to intermediate-grade pancreatic NET (NCT01824875).

PIVOTAL CLINICAL EVIDENCE: TREATMENTS FOR CARCINOID SYNDROME
Long-acting octreotide is approved for the treatment of severe diarrhea and flushing episodes associated with metastatic carcinoid tumors [50], based on the results of the randomized clinical trial by Rubin et al., in which complete (no rescue medication) or partial (rescue medication on >2 occasions for ≤5 days) treatment success was demonstrated in 66.7%, 71.4%, and 61.9% of participants treated with long-acting octreotide 10 mg, 20 mg, or 30 mg, respectively [51]. In the recent ELECT clinical trial, lanreotide was found to be effective for the control of carcinoid syndrome symptoms [52]. The adjusted mean percentage of days in which rescue octreotide was used during the 16-week double-blind phase was significantly lower for participants who received lanreotide than for those who received placebo (33.7% vs. 48.5%).
In early 2017, telotristat ethyl was approved for the treatment of carcinoid syndrome diarrhea in combination with SSAs in adults inadequately controlled with SSA therapy alone [53], based on the results of the TELESTAR trial, which assessed the safety and efficacy of telotristat ethyl + SSA therapy in patients with WD metastatic NET and carcinoid syndrome refractory to SSA therapy [54]. Telotristat ethyl at either dose (250 mg t.i.d. or 500 mg t.i.d.) was associated with statistically significant reductions in bowel movement frequency over time compared with placebo; 44% and 42% of participants treated with 250 mg t.i.d. and 500 mg t.i.d., respectively, were classified as responders versus 20% of patients receiving placebo [54].

EVIDENCE FOR TREATMENT OF G3 PD-NEC
Evidence supports the use of platinum-based chemotherapy in the first-line treatment for G3 PD-NEC regardless of primary site. Median PFS with this first-line treatment is approximately 5.0-5.8 months for patients with G3 PD GEP NEC, SCLC, and LCNEC [15, 55,56]. The ORR reported with first-line platinum-based chemotherapy is highest in SCLC (~67%) [55] compared with LCNEC (47%) [56]

EVOLVING TREATMENT ALGORITHM FOR ADVANCED NEN
We present a conceptual treatment schematic for systemic therapy based on tumor classification for advanced NEN (Fig. 2). There is currently no established standard treatment for the new classification of G3 WD-NET, which have a significantly worse response to platinum-based chemotherapy than G3 PD-NEC [15]. Systemic treatments for tumor control in G2 WD-NET should also be considered for G3 WD-NET, given their high SSTR2 expression [19] and consistent molecular profile [32]. An analysis of sunitinib efficacy in a phase II trial reported preliminary evidence of activity in G3 WD-NET [57], indicating that targeted therapies may be appropriate in this clinical situation. Prospective randomized trials are required to determine the effectiveness of therapies for these patients.
There is currently no established standard treatment for the new classification of G3 WD-NET, which have a significantly worse response to platinumbased chemotherapy than G3 PD-NEC. Systemic treatments for tumor control in G2 WD-NET should also be considered for G3 WD-NET, given their high SSTR2 expression and consistent molecular profile.

CONCLUSION
During recent years, the treatment algorithm for advanced NEN has evolved from one that is directed by primary site-specific classification to one that is directed by biologic classification, with overlapping treatments across the primary sites. Commonalities in biologic characteristics across primary tumor sites include functional/nonfunctional status, WD/PD status, high/low grade, SSTR+/SSTR− status, and genetic alterations, with biologic characteristics guiding treatment decisions in some circumstances. It is therefore important to understand the diversity of and commonalities in biologic characteristics across primary tumor sites as the conception of NET continues to evolve, with new classifications and clinical evidence becoming available.