Somatostatin Analogs for Pancreatic Neuroendocrine Tumors: Any Benefit When Ki‐67 Is ≥10%?

Abstract

Background

Long‐acting somatostatin analogs (SSAs) are the primary first‐line treatment of well‐differentiated advanced gastroenteropancreatic neuroendocrine tumors (NETs), but data about their efficacy in pancreatic NETs (panNETs) with Ki‐67 ≥10% are still limited.

Materials and Methods

To assess the clinical outcomes of advanced, nonfunctioning, well‐differentiated panNETs with Ki‐67 ≥10% receiving first‐line long‐acting SSAs in a real‐world setting, we carried out a retrospective, multicenter study including patients treated between 2014–2018 across 10 centers of the NET CONNECT Network. The primary endpoints were time to next treatment (TNT) and progression‐free survival (PFS), whereas overall survival (OS) and treatment safety were secondary endpoints.

Results

A total of 73 patients were included (68 grade [G]2, 5 G3), with liver metastases in 61 cases (84%). After a median follow‐up of 36.4 months (range, 6–173), the median TNT and PFS were 14.2 months (95% confidence interval [CI], 11.6–16.2) and 11.9 months (95% CI, 8.6–14.1) respectively. No statistically significant difference was observed according to the somatostatin analog used (octreotide vs. lanreotide), whereas increased tumor grade (hazard ratio [HR], 4.4; 95% CI, 1.2–16.6; p = .04) and hepatic tumor load (HR, 2; 95% CI, 1–4; p = .03) were independently associated with shortened PFS. The median OS recorded was 86 months (95% CI, 56.8–86 months), with poor outcomes observed when the hepatic tumor burden was >25% (HR, 3.4; 95% CI, 1.2–10; p = .01). Treatment‐related adverse events were reported in 14 patients, most frequently diarrhea.

Conclusion

SSAs exert antiproliferative activity in panNETs with Ki‐67 ≥10%, particularly in G2 tumors, as well as when hepatic tumor load is ≤25%.

Implications for Practice

The results of the study call into question the antiproliferative activity of somatostatin analogs (SSAs) in pancreatic neuroendocrine tumors with Ki‐67 ≥10%. Patients with grade 2 tumors and with hepatic tumor load ≤25% appear to derive higher benefit from SSAs. Prospective studies are needed to validate these results to optimize tailored therapeutic strategies for this specific patient population.

Introduction

Long‐acting somatostatin analogs (SSAs) are the primary first‐line treatment of advanced gastroenteropancreatic (GEP) neuroendocrine tumors (NETs) [1–3]. Initially developed as antisecretory agents for the palliation of hormonal symptoms associated with NETs, two phase III clinical trials (the PROMID and the CLARINET studies) [4, 5] have demonstrated their antiproliferative activity in well‐ or moderately differentiated NETs.

The double‐blind, placebo‐controlled PROMID trial [4] randomized 85 patients diagnosed with well‐differentiated, metastatic midgut NETs (mostly grade [G]1) to receive either octreotide long‐acting repeatable (LAR) 30 mg every 28 days or placebo. A significant improvement in the median time to progression (TTP) from 6 months in the placebo arm to 14.3 months in the SSA arm (hazard ratio [HR], 0.34; p < .001) was reported. More recently, the double‐blind, placebo‐controlled CLARINET trial [5] randomized 204 patients who had metastatic, OctreoScan‐positive, hormonally nonfunctioning GEP‐NETs with a Ki‐67 index less than 10% to receive lanreotide depot 120 mg every 4 weeks or placebo. After a median study drug exposure of 24 months, the median progression‐free survival (PFS) was not reached in the lanreotide arm versus 18 months in the placebo arm (HR, 0.47; p < .001). In a recent meta‐analysis [6], a significant cumulative benefit in terms of disease control was shown for SSAs in comparison with placebo (HR, 0.41; p < .01), and an excellent safety profile was reported.

Although SSAs have a pivotal role in the treatment of GEP‐NETs, evidence on their efficacy in specific patient populations is still limited. Although subgroup analyses of the CLARINET trial [5] showed that lanreotide had a positive impact on the PFS of the 91 patients with pancreatic NETs (panNETs) included in the study (median PFS, not reached in the lanreotide arm vs 12.1 months in the placebo arm; HR, 0.58; p < .06), data supporting the efficacy of SSAs in panNETs with a Ki‐67 ≥10% are presently lacking.

This study aimed at assessing the clinical outcomes of patients with panNETs and a Ki‐67 ≥10% treated with long‐acting SSAs as first‐line antitumor therapy in a real‐world setting. Whereas time to next treatment (TNT) and PFS were chosen as primary endpoints, overall survival (OS) and treatment safety were used as secondary endpoints.

Subjects, Materials, and Methods

Study design and patient selection

A retrospective, real‐world study investigating the efficacy and safety of first‐line SSAs in panNETs with Ki‐67 ≥10% was conducted in 10 NET expert centers belonging to the NET CONNECT network (Beaujon University Hospital, Clichy, France; Erasmus Medical Center, Rotterdam, The Netherlands; Moffitt Cancer Center, Tampa, FL; Peking University Cancer Hospital, Beijing, China; Royal Free London NHS Foundation Trust, London, U.K.; The Christie NHS Foundation Trust, Manchester, U.K.; University Hospital Ramon y Cajal, Madrid, Spain; University of Bari, Italy; University of Warmia and Mazury, Olsztyn, Poland; Vall d'Hebron University Hospital, Barcelona, Spain). The study protocol was approved by the ethics or audit committees at each institution, where required. Informed consent was waived, because of the retrospective nature of the study.

Chart reviews and existing institutional databases were used to identify adult (>18 years) patients diagnosed with metastatic (American Joint Committee on Cancer stage IV) and histologically or cytologically proven panNETs with Ki‐67 ≥10%. Patients with poorly differentiated neuroendocrine carcinoma were excluded. Only patients who initiated treatment with first‐line octreotide LAR or lanreotide depot between 2014 and 2018 and had at least 1 year of follow‐up were eligible. Locoregional therapies prior to the initiation of the treatment with SSAs were not allowed. To limit referral bias, we excluded patients who initiated treatment with SSAs prior to referral to our institutions.

Data Collection

The following information was collected by reviewing patient medical records: demographics, date of diagnosis, location of metastases, volume of liver metastases (as assessed by study investigators through scan review), somatostatin receptor imaging status, Ki‐67 proliferative index, chromogranin A levels, occurrence of adverse events while on treatment with SSAs, and follow‐up data.

Statistical Analysis

Descriptive statistics were used for patient demographics and clinical and pathological data. TNT was defined as the time from initial SSA therapy to initiation of new systemic or locoregional treatment. PFS was defined as the time from SSA initiation to radiographic progression or death. Progression status was assessed based on a review of medical notes and scan reports and not based on formal RECIST radiographic readings. OS was measured from the date of initial diagnosis until death from any cause, or patients were censored at last known follow‐up. Survival curves were estimated using the Kaplan‐Meier method and compared by the log‐rank test. Multivariable analysis was performed using Cox proportional hazards regression. The assumption of proportionality was verified by log‐log plot. Only variables with a p < .2 at univariate analysis were included in the Cox multivariable model, and the HR was interpreted as the instantaneous relative risk of an event for an individual with the risk factor present compared with an individual with the risk factor absent. Exact 95% confidence intervals (CIs) were calculated for each proportion of interest. All tests were two‐sided, and statistical significance was declared when p ≤ .05. All statistical analyses were performed using MedCalc 12.7 (MedCalc Software, Ostend, Belgium).

Results

Demographics and Tumor Characteristics

A total of 73 patients diagnosed with nonfunctioning, well‐differentiated, metastatic panNETs treated with first‐line SSAs were included in the study. Demographic variables and baseline clinicopathological characteristics are summarized in Table 1. Thirty‐nine patients (53%) were male, and the mean age at SSA initiation was 58 (±12) years. The majority (68/73) of tumors were classified as G2 according to the World Health Organization 2019 classification (Klimstra et al. 2019), and all G3 tumors were well‐differentiated [7]. The Ki‐67 labeling index was ≤15% in 52 of 73 patients (71%). Pathological analyses were performed on biopsy and surgical samples in 52 and 21 cases, respectively. Seventy‐one tumors were sporadic, whereas two occurred in the context of type 1 multiple endocrine neoplasia syndrome. Liver metastases were diagnosed in 61 patients (84%), with hepatic tumor load ≤25% and >25% in the 38% and 43% of the cohort, respectively. The hepatic tumor volume was not reported in 19% of cases. The somatostatin receptor imaging (single‐photon emission–computed tomography or positron emission tomography–computed tomography) was positive or heterogeneous (coexistence of both positive and negative tumor lesions) in 92% and 7%, respectively, of the study population. Octreotide LAR or lanreotide depot were administered in 35 and 38 patients, respectively, and the median time from diagnosis until SSA initiation was 1.5 months (range, 0–145). All patients had metastatic disease at start of treatment with SSAs, with 61 of them (84%) having metastases at diagnosis or within 12 months of diagnosis (synchronous) and the remaining (16%) developing metastasis after 12 months since the original diagnosis (metachronous). No patients received above‐label dosages of SSAs. Primary tumor surgery was performed before SSA initiation in 21 (29%) patients, and the median frequency of follow‐up imaging during the treatment with SSAs was 3 months (range, 1–12). At SSA progression, 43% of the patients received cytotoxic treatments including chemotherapy or peptide receptor radionuclide therapy (PRRT).

Treatment Outcomes

Patients received a median of twelve 28‐day treatment cycles. Reasons for discontinuation included radiographic tumor progression (n = 55), clinical progression (n = 4), and death (n = 2). Three patients were lost to follow‐up, whereas 9 patients remained on treatment at the time of data analysis.

All enrolled patients were evaluable for TNT and PFS. After a median follow‐up duration of 36.4 months (range, 6–173), the median TNT was 14.2 months (95% CI, 11.6–16.2 months), whereas the median PFS was 11.9 months (95% CI, 8.6–14.1 months; Fig. 1). Stratified by tumor grade (Fig. 2A, B), the median TNT was 14.7 months (95% CI, 12.8–18.5 months) for patients with G2 panNETs and 5.2 months (95% CI, 3–8.7 months) for those with G3 panNETs (p = .0001). Similarly, the median PFS was significantly prolonged in patients with G2 panNETs as compared with those having G3 panNETs (12.4 months vs. 4 months; p = .0007). As depicted in Figure 2C and D, patients with hepatic tumor burden ≤25% showed longer PFS (median: 15 months; 95% CI, 12.2–19.8 months) and TNT (median: 19 months; 95% CI, 13.1–22.6 months) as compared with those with higher liver tumor load, showing a median PFS and TNT of 9.7 months (95% CI, 5.1–12.4 months; p = .04) and 12.8 months (95% CI, 5.7–15.8 months; p = .07), respectively. As shown in supplemental online Figure 1, in patients who developed metastatic disease after 12 months from tumor diagnosis, the median TNT and PFS were 22.6 months (95% CI, 12.8–37.7 months) and 18 months (95% CI, 12.4–34.3 months), respectively, versus 14.2 months (95% CI, 11.6–16.2 months; p = .12) and 11 months (95% CI, 6.6–13.8 months; p = .05) in subjects developing metastasis within 12 months since diagnosis. There was no difference in PFS and TNT according to somatostatin receptor imaging status, serum chromogranin levels at treatment initiation, and administered SSA (octreotide vs. lanreotide). By multivariable analysis, both increased tumor grade (HR, 4.4; 95% CI, 1.2–16.6; p = .04) and hepatic tumor load >25% (HR, 2; 95% CI, 1–4; p = .03) remained independently associated with shortened PFS.

Overall Survival

At the time of data cutoff, 24 patients had died and 49 patients were alive. The median OS was 86 months (95% CI, 56.8–86 months). Estimated rates of OS at 1, 3, and 5 years were 93% (±3%), 75.4% (±5.4%), and 57.9% (±7.7%), respectively (Fig. 3A). As depicted in Figure 3B, a hepatic tumor burden >25% predicted poor OS (median OS: 86 months vs. 56.8 months; p = .001), and this prognostic effect was retained after multivariable analysis (HR, 3.4; 95% CI, 1.2–10; p = .01). All other investigated variables (tumor grade, somatostatin receptor imaging status, serum chromogranin levels at treatment initiation, presence of bone metastases, synchronous/metachronous occurrence of metastases) did not show any prognostic impact. The median OS was 86 months (95% CI, 57–86 months) in patients with G2 tumors and 56.8 months (95% CI, 19.6–56.8 months) in patients with G3 tumors.

Safety

The side effects considered at least possibly related to treatment are listed in Table 2. Overall, any‐grade adverse events (AEs) were reported in 14 of 73 patients. Diarrhea was the most common toxicity and developed in 7 patients (13%). Other AEs included pancreatic insufficiency with steatorrhea (4%), cholangitis with biliary acute pancreatitis (4%), and joint pain (4%).

Discussion

To our knowledge, this is the first study to investigate the antiproliferative activity of SSAs in the first‐line treatment of nonfunctioning, well‐differentiated, sporadic or hereditary, advanced panNETs with Ki‐67 ≥10%. Among the 73 patients retrospectively enrolled in our multi‐institutional series, the SSAs octreotide LAR and lanreotide depot were associated with a median TNT of 14.2 months (95% CI, 11.6–16.2) and a median PFS of 11.9 months (95% CI, 8.6–14.1). Although lower grade and limited hepatic tumor load predicted better outcomes, similar efficacy results were observed in patients receiving either octreotide or lanreotide.

Based on the results of the PROMID and CLARINET trials [4, 5], SSAs currently constitute the primary treatment for patients with NETs. Nevertheless, given the characteristics of the patients included in both trials, little is known regarding the impact of SSA therapy in patients with more rapidly proliferative NETs. The aim of our study was therefore to retrospectively assess the real‐world efficacy of SSAs among patients with panNETs whose Ki‐67 exceeds the 10% cutoff used as an inclusion criterion in the CLARINET study [5]. Although cross‐study comparisons should be drawn with caution, the median PFS of 11.9 months recorded in our study compares favorably with prior literature in the same disease setting. Indeed, in a retrospective analysis of seven patients harboring panNETs with a Ki‐67 ≥10%, first‐line octreotide LAR was associated with a median TTP of 4 months (95% CI, 1.4–6.6 months) [8]. A median PFS of 10.3 months (95% CI, 2.1–18.4 months) was also reported in a single‐institution series including 20 patients with well‐differentiated GEP‐NETs and a Ki‐67 exceeding 10% who received lanreotide primarily in the first‐line setting [9]. Moreover, in a subanalysis of a phase II trial of lanreotide in Japanese patients with advanced NETs, the median PFS observed among the two patients having tumors with Ki‐67 ≥10% was only 12 weeks [10]. The heterogeneity of primary tumor site, metastatic pattern, and treatment setting along with the low number of patients analyzed in these studies should be taken into account when interpreting our findings.

Systemic treatment options active in controlling tumor growth in patients with well‐differentiated panNETs also include radiolabeled SSAs, everolimus, sunitinib, and chemotherapy. Although the NETTER‐1 study [11] investigated ¹⁷⁷Lu‐DOTATATE plus octreotide LAR in patients with advanced, OctreoScan‐positive, midgut NETs progressive on standard‐dose SSAs, the NETTER‐2 study (NCT03972488) is currently assessing the efficacy of PRRT plus octreotide LAR versus high‐dose octreotide LAR in a population of patients somehow similar to that analyzed in our study (well‐differentiated GEP‐NETs with Ki‐67 composed between 10% and 55%). In this regard, a study of 33 patients pretreated for advanced, somatostatin receptor–positive GEP neuroendocrine neoplasms (any degree of differentiation) with a Ki‐67 of at least 15% showed that ¹⁷⁷Lu‐DOTATATE was associated with an encouraging median PFS of 23 months (95% CI, 14.9–31 months) [12]. Data on mTOR inhibitors or antiangiogenic agents in the treatment of rapidly proliferating NETs are scant, and low number of patients hinder comparisons with our study. Recently, a phase II study investigated everolimus at 10 mg daily in nine patients with ¹⁸FDG‐positive G2 panNETs (Ki‐67 was ≥10% in 6 patients), reporting a median PFS of 14 months (95% CI, 7–24 months) [13]. In a small series of 18 patients with panNETs treated with sunitinib, the median PFS was 13 and 8 months when the Ki‐67 was <10% and ≥10%, respectively [14]. Notably, the efficacy of first‐line chemotherapy in panNETs with Ki‐67 ≥10% has been specifically investigated in a retrospective multicenter study of the French group of endocrine tumors [15]. Seventy‐four patients were enrolled, and a median PFS of 7.2, 7.5, and 7.2 months was recorded for patients receiving streptozocin‐, platinum‐, or dacarbazine/temozolomide‐based regimens, respectively. Although the PFS observed in our study compares favorably with that reported in the French series, selection biases may hinder a reliable comparison between the two cohorts. Moreover, more promising data have been reported by other groups for the capecitabine and temozolomide (CAPTEM) regimen in a disease setting similar to that explored in the current work [16]. In this context, no detrimental association has been described between a Ki‐67 ≥10% and treatment outcomes in patients receiving CAPTEM [17].

The present study does not show statistically significant differences in clinical outcomes according to the type of SSA (octreotide vs. lanreotide). This is in line with prior literature, showing no difference in time to treatment discontinuation (TTD) between the two SSAs (median TTD of 19.2 and 17.5 months for octreotide and lanreotide, respectively) among 1,086 patients with pancreatic or extrapancreatic NET of any grade identified using a provider‐based claims database [18]. In contrast, a multivariable model showed that both liver tumor load (cutoff, 25%) and grading (G3 vs. G2) were significant, independent predictors of PFS in our cohort. This is partly consistent with the CLARINET trial (although in a different patient population) [5], as well as with a previous study by Kang et al. [9] that analyzed the efficacy of lanreotide monotherapy in a cohort of GEP‐NET patients including panNETs with Ki‐67 ≥10%.

The 5‐year OS rate of 57.9% recorded in our study for a cohort of G2/G3 panNETs with a Ki‐67 ≥10% compares favorably with that reported for either G1/G2 (54.7%) or G3 panNETs (23.2%) in a study analyzing 458 patients with metastatic panNETs from the Spanish National Tumor Registry [19]. This might suggest the inclusion of a population of patients with more indolent disease biology in our series and/or reflect the already described improvement of treatment outcomes for those patients who are treated in high‐volume centers [20]. By multivariable analysis, only the hepatic tumor load independently predicted OS in the present study. Although bone metastases have been previously described as a negative prognostic factor in patients with NETs [21], in our series their presence did not significantly affect clinical outcomes, possibly a consequence of low numbers of patients with bone neoplastic involvement (n = 13).

Although multiple series [22–25] have documented the use of SSAs among patients with G3 NETs, this is, to our knowledge, the first study to formally report the survival outcomes associated with SSAs in an albeit small cohort of patients with well‐differentiated, G3 panNETs. The rationale to treat patients with G3 panNETs with somatostatin‐based therapies relies on the expression of somatostatin receptors, which has been described in small series to occur in up to 80% of G3 GEP‐NETs [26, 27]. Nevertheless, based on the median PFS and TNT of 4 and 5.2 months, respectively, observed in our study, the role of SSA therapy in G3 panNETs appears quite limited.

In our series, SSAs showed a manageable toxicity profile, with a pattern of toxicities similar to that observed in previous studies. Rates of treatment‐related adverse events were lower than those reported by prospective clinical trials of SSAs in patients with NETs [4, 5]. Given its retrospective nature, our study might have underestimated the occurrence of toxicities.

Our study has several limitations. First, its retrospective, nonrandomized design does not allow us to exclude selection biases leading to the enrollment of patients with more favorable prognosis, therefore considered for treatment with SSAs. However, a substantial proportion (42%) of included patients displayed a liver tumor load greater than 25%, a figure that reflects the current “real‐life” practice among NET referral centers. Another possible bias of the present work is the lack of a centralized radiology review to assess disease progression of the included population. This limitation might be counterbalanced by the long‐standing experience with NET management of the participating centers, where progression is systematically reviewed in NET‐dedicated multidisciplinary meetings. The lack of standardized methods to assess liver burden and missing data for several patients also limit our work.

Conclusions

SSAs exert antiproliferative activity against panNETs with Ki‐67 ≥10% and can be used instead of other more toxic treatments. There appears to be an association between lower tumor burden (<25%) and potential for benefit in this population. Prospective and preferably controlled studies are needed to validate these results to optimize tailored therapeutic strategies for this specific patient population.

Acknowledgments

The authors thank NET CONNECT for having made this study possible through logistical and organizational support provided by COR2ED. NET CONNECT is supported by an Independent Educational Grant from IPSEN. The program is therefore independent, and the content is not influenced by IPSEN and is under the sole responsibility of the experts. Elettra Merola, Teresa Alonso Gordoa, Panpan Zhang, Wouter Zandee, Faidon Laskaratos, Louis de Mestier, Angela Lamarca, Jorge Hernando, Wouter de Herder, Martin Caplin, Mauro Cives, and Rachel van Leeuwaarde are members of the NET CONNECT group. This research did not receive any specific grant from any funding agency in the public, commercial, or not‐for‐profit sector.

Author Contributions

Conception/design: Elettra Merola, Teresa Alonso Gordoa, Panpan Zhang, Wouter de Herder, Martin Caplin, Mauro Cives, Rachel van Leeuwaarde

Provision of study material or patients: Elettra Merola, Teresa Alonso Gordoa, Panpan Zhang, Taymeyah Al‐Toubah, Eleonora Pelle’, Agnieszka Kolasińska‐Ćwikła, Agnieszka Kolasińska‐Ćwikła, Faidon Laskaratos, Louis de Mestier, Angela Lamarca, Jorge Hernando, Jaroslaw Cwikla, Jonathan Strosberg, Wouter de Herder, Martin Caplin, Mauro Cives, Rachel van Leeuwaarde

Collection and/or assembly of data: Elettra Merola, Teresa Alonso Gordoa, Panpan Zhang, Taymeyah Al‐Toubah, Eleonora Pelle’, Agnieszka Kolasińska‐Ćwikła, Agnieszka Kolasińska‐Ćwikła, Faidon Laskaratos, Louis de Mestier, Angela Lamarca, Jorge Hernando, Mauro Cives, Rachel van Leeuwaarde

Data analysis and interpretation: Elettra Merola, Teresa Alonso Gordoa, Panpan Zhang, Mauro Cives, Rachel van Leeuwaarde

Manuscript writing: Elettra Merola, Teresa Alonso Gordoa, Panpan Zhang, Mauro Cives, Rachel van Leeuwaarde

Final approval of manuscript: Elettra Merola, Teresa Alonso Gordoa, Panpan Zhang, Taymeyah Al‐Toubah, Eleonora Pelle’, Agnieszka Kolasińska‐Ćwikła, Wouter Zandee, Faidon Laskaratos, Louis de Mestier, Angela Lamarca, Jorge Hernando, Jaroslaw Cwikla, Jonathan Strosberg, Wouter de Herder, Martin Caplin, Mauro Cives, Rachel van Leeuwaarde

Disclosures

Elettra Merola: ENETS excellence Academy fellowship 2017 (RF‐Grant awarded for another research project regarding neuroendocrine tumors); Louis de Mestier: Ipsen, Novartis, Pfizer (C/A), Novartis (RF); Angela Lamarca: Ipsen, Pfizer, Bayer, AAA, SirtEx, Novartis, Mylan, Delcath (Other‐travel, educational support), Merck, Pfizer, Ipsen, Incyte, AAA (H‐Speaker), EISAI, Nutricia Ipsen, QED, Roche (H‐Advisory), Knowledge Network and NETConnect Initiatives funded by Ipsen (Other‐Member); Jonathan Strosberg: Novartis (C/A), Ipsen, Lexicon (H); Wouter de Herder: Ipsen (RF), AAA/Novartis (SAB); Martin Caplin: Novartis, AAA, Ipsen, Pfizer (H, SAB); Mauro Cives: AAA, Ipsen (SAB). The other authors indicate no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

References

Author notes