Lead Investigator: Chiara Alessandra Cella, Istituto Europeo di Oncologia
Title of Proposal Research: Addressing the synergistic effect of metformin in patients with Pancreatic neuroendocrine tumors treated with sunitinib: a post-hoc analysis from the main pivotal trials.
Vivli Data Request: 7874
Funding Source: None
Potential Conflicts of Interest: CAC reports personal fees from BMS and Leo Pharma; and a research grant from IPSEN (institutional).
Any potential conflicts of interest will be declared in subsequent publication.
Dr. Fazio received personal fees from AAA, Hutchinson MediPharma, Merck, Novartis; has financial interest with 4SC, Astellas, Beigene, FIBROGEN, Incyte, IPSEN, MSD and NUCANA; and has received research grant form IPSEN (institutional).
Any potential conflicts of interest will be declared in subsequent publication.
Summary of the Proposed Research:
Neuroendocrine neoplasms (NENs) are a type of abnormal and excessive growth of tissue, accounting for about 0.5% of all newly diagnosed malignancies. The worldwide incidence, which is on the rise, possibly due to improved awareness, is approximately 5.86/100,000 per year with a prevalence in the United States estimated around <200,000, which makes it an orphan disease. Pancreatic neuroendocrine neoplasm (Pan-NENs), represent less than 3% of primary pancreatic tumors and involve the pancreatic islet cell, a pancreatic cell that produces hormones (e.g., insulin and glucagon) that are secreted into the bloodstream to help control the level of glucose (sugar) in the blood (also called islet of Langerhans). The mainstay of treatment is surgery with curative intent, if possible. When surgery has no indication, chronic medical management is required. Among the approved options, sunitinib (SUN) represents a cornerstone in the therapeutical strategy on Pan-NENs, having showed a progression-free survival (PFS) doubling rate over placebo. PFS refers to the time during which a patient shows no signs or symptoms of the growth or the spreading of a tumor. However, soon or later, the onset of intrinsic or acquired resistance still represents the limiting criteria for a durable response.
Metformin (MET) is one of the most popular oral glucose-lowering medications, widely considered to be the optimal initial therapy for patients with type 2 diabetes mellitus (DM2). Metformin has shown to exert an anti-tumor effect in solid tumors, including NENs. More in detail, MET has shown to inhibit tumor growth in Pan-NENs in preclinical models and is also associated with longer PFS in patients with Pan-NENs receiving everolimus, a drug used as an immunosuppressant to prevent rejection of organ transplants and as a targeted therapy in the treatment of several solid tumours. Additionally, MET has shown safety and efficacy in improving outcomes even when combined with sunitinib in renal cell carcinoma.
Neverthless, limited data are available on a possibile role of MET in patients with Pan-NENs treated with antiangiogenic targeted agents, like sunitinib.
Basically, the aim of this analysis is to evaluate if the association between the MET and SUN improves survival in patients with advanced Pan-NETs.
Statistical Analysis Plan:
The primary aim of this analysis is to evaluate the association between the use of metformin and PFS among advanced Pan-NETs patients treated with Sunitinib.
A total of 258 patients with Pan-NETs were treated with Sunitinib in the three included trials (66 from NCT00056693, 86 from NCT00428597 and 106 from A6181202) and will be considered in this analysis.
The median PFS among patients treated with Sunitinib, among the included patients, was approximately 11 months, overall.
We expect a 15% prevalence of patients with type II diabetes (T2D) and half of these T2D patients to be treated with metformin.
Therefore, we anticipate to observe:
– 219 patients without T2D;
– 19 patients with T2D, but no treated with metformin;
– 19 patients with T2D and treated with metformin.
Our primary hypothesis is that the use of metformin halves the hazard of progression in patients treated with Sunitinib.
This hypothesis is based on what was observed in patients with Pan-NENs who were given everolimus, where the risk of progression was significantly lower in patients with T2D treated with metformin than in patients with T2D not treated with metformin (HR: 0.49, 95% CI: 0.34–0.69) and in patients without diabetes (HR: 0.45; 95% CI, 0.32–0.62).
Under this premise, and assuming that T2D is not associated with PFS, we expect to observe a median PFS of approximately 20 months in patients with T2D and treated with metformin.
Patients’ demographic, disease and treatment characteristics,
overall and stratified by the three groups (‘‘T2D with Metformin”, ‘‘T2D w/o Metformin”, No T2D) will be reported. T- test for continuous variables, Chi-square test for categorical variables and Cochran–Armitage’s trend test for ordinal variables will be used to compare the distribution of the evaluated characteristics between groups.
The primary outcome will be the PFS, while secondary outcomes will be the overall survival (OS) and the objective response rate (ORR).
The PFS and OS functions will be estimated using the Kaplan-Meier method, and stratified log-rank test will be used to assess differences between groups. The analysis will be stratified by trial to account for the independence of each study included.
Stratified chi-square test will also be used to evaluate the association between metformin and ORR.
A stratified Cox regression model will be used to estimate the effect of metformin on PFS and on OS, in terms of HR and 95% CI, while a stratified log-binomial regression model will be used to estimate the relative risk (RR), and 95% CI, of objective response.
Multivariable Cox and log-binomial models will be used to adjust the effects of interest for patient characteristics that will result as significantly associated with metformin use.
All analyses will be performed using SAS software v.9.4 (SAS
Institute, Cary, NC, USA) and R version 3.4.1.
Considering a 5% two-sided type I error for the log-rank test, the power to reject the null hypothesis of no association between metformin use and PFS will be approximately 52%, 61% and 69% in case the observed PFS events will be, respectively, 111 (i.e. the actual sum of PFS events reported in the sunitinib arm in the phase III study published in the NEJM and the PFS events reported in the included phase II and phase IV trials), 150 and 200 (i.e. assuming that new PFS events that occurred after the publication of the original articles will be available).
Missing data not will be imputed and will be excluded from the analysis.
Power analysis was performed using PASS software v15.0 (NCSS, Kaysville, UT, USA).
To be more conservative, assuming a situation in which we observe only 10 patients with T2D and treated with metformin, and with all other assumptions remaining the same, the power to reject the null hypothesis of no association between metformin use and PFS will be approximately 37%, 43% and 49% in case the observed PFS events will be, respectively, 111, 150 and 200.
Requested Studies:
A Phase II Study Of The Efficacy And Safety Of SU011248 In Patients With Advanced Unresectable Neuroendocrine Tumor
Data Contributor: Pfizer Inc.
Study ID: NCT00056693
Sponsor ID: RTKC-0511-015
A Phase III Randomized, Double-Blind Study Of Sunitinib (SU011248, SUTENT) Versus Placebo In Patients With Progressive Advanced/Metastatic Well-Differentiated Pancreatic Islet Cell Tumors
Data Contributor: Pfizer Inc.
Study ID: NCT00428597
Sponsor ID: A6181111
Raymond E, Kulke MH, Qin S, Yu X, Schenker M, Cubillo A, Lou W, Tomasek J, Thiis-Evensen E, Xu JM, Croitoru AE, Khasraw M, Sedlackova E, Borbath I, Ruff P, Oberstein PE, Ito T, Jia L, Hammel P, Shen L, Shrikhande SV, Shen Y, Sufliarsky J, Khan GN, Morizane C, Galdy S, Khosravan R, Fernandez KC, Rosbrook B, Fazio N. Efficacy and Safety of Sunitinib in Patients with Well-Differentiated Pancreatic Neuroendocrine Tumours. Neuroendocrinology. 2018;107(3):237-245. doi: 10.1159/000491999. Epub 2018 Jul 10. PMID: 29991024.
Data Contributor: Pfizer Inc.
Study ID: NCT01525550
Summary of Results:
Unfortunately, we were unable to bring our analysis to completion due to the following reasons:
- Because of the equivocal classification of patients with or w/o diabetes, we decided to dichotomize the study population into “metformin users” and “non-metformin users”
- We could not use any data from Phase II (Kulke 2008) dataset, since we know that there were three metformin users in the dataset (thanks to your response to our inquiry: “Three patients (metformin users) were hidden by low frequency redaction in the “conmeds” dataset”), but we don’t know who they were and, consequently, if they developed or not the events of interest.
- once having dichotomized the patients, we found that the number of “metformin users” (n = 12) was lower than the original sample size calculated in the statistical plan (n = 19)
- Although we had to exclude the Phase II patients, we tried to perform our analyses, without reaching any conclusive results: there was a weak trend indicating a better PFS for the metformin users and, conversely, a weak trend indicating a worse OS in the same group (no statistical significance).
Please find linked a table and two figures summarizing the results.