To the Editor


Somatostatin analogs (SSAs) are the mainstay of treatment for patients with advanced, well-differentiated NETs (WDNETs) expressing somatostatin receptors [1, 2]. The antitumor effects of SSAs in WDNETs are in part mediated through the inhibition of the PI3K/AKT/mTOR and MAPK pathways [3]. In retrospective studies, the use of the antidiabetic compound metformin in diabetic patients with advanced WDNETs was associated with better clinical outcomes when combined with standard SSAs plus/minus everolimus [4,5,6,7]. However, no prospective evidence exists to support metformin use in combination with SSAs in advanced WDNET patients with or without diabetes mellitus (DM). Based on these premises, we conducted MetNET-2, a first-in-human, phase Ib clinical trial that investigated the safety and antitumor activity of experimental metformin in combination with standard lanreotide autogel (ATG) in both diabetic and non-diabetic patients with advanced GI or thoracic WDNETs. The primary study objective was to assess the safety of the experimental treatment, as defined as the incidence of serious adverse events (SAEs). Study Methods are reported in Additional file 1.

Between April 2016 and April 2019 we enrolled a total number of 20 patients, whose characteristics are summarized in Additional file 2: Table S1. Of these, six patients (30%) had a prior diagnosis of DM. Median duration of exposure to the experimental treatment was 15.9 months (IQR range, 5.8–21.0). Study drug exposure is reported in Additional file 3: Fig. S1 and Additional file 4: Table S2. With only one treatment-related SAE (5%, acute renal failure), MetNET-2 met its primary endpoint demonstrating the safety of metformin plus SSAs. The renal SAE was likely to be multifactorial (G2 hypertension NDR, G1 urolithiasis NDR), with a possible contribution of metformin-related pre-renal kidney injury from G1 diarrhea. Of note, this toxicity was reversible after temporary interruption of metformin administration and the initiation of anti-hypertensive therapy. The most common any-grade treatment-related AEs (trAEs) were diarrhea (75%), hyperglycaemia (55%), asthenia (40%) and hypercholesterolemia (40%) (Table 1). Grade 3 trAEs occurred in 15% of patients, and consisted of acute renal failure (n = 1; 5%), diarrhea (n = 1; 5%) and abdominal pain (n = 1; 5%). No grade ≥ 4 trAEs were reported. trAE incidence was not significantly different in diabetic versus non-diabetic patients (Additional file 5: Table S3). In addition, no trAEs led to the discontinuation of Lanreotide ATG or metformin. Treatment-emergent AEs are reported in Additional file 6: Table S4.

Table 1 Most common treatment-related adverse events (trAEs)
Full size table

The antitumor activity of the experimental treatment is summarized in Additional file 7: Table S5, Additional file 8: Fig. S2 and Additional file 9: Fig. S3. ORR was 10% (95% CI 1–32%) and DCR was 85% (95% CI 62–96%). With a median follow-up of 39 months (95% CI 28 months-NE), median PFS was 24 months (95% CI 16-NE months) (Fig. 1A), with 12-month and 24-month PFS probability of 75% (95% CI 58–97%) and 49% (95% CI 31–77%), respectively. Median TTP was 26 months (95% CI 17-NE months) (Fig. 1B). Diabetic status was not significantly associated with PFS (Additional file 10: Fig. S4). With the exception of non-functioning tumor status, which was associated with a lower risk of disease progression, none of the other clinico-pathological characteristics showed an association with PFS (Additional file 11: Table S6).

Fig. 1
figure 1

Kaplan–Meier curves for progression-free survival (PFS) (A) and time-to-progression (TTP) in the MetNET2 study cohort (B). Box plots depicting changes in the indicated metabolic parameters during the experimental treatment (C–H). The p value in panels C–H refers to the paired t test for the indicated comparisons

Full size image

Then, we explored the potentially prognostic role of tumor genomic alterations, as evaluated through NGS analysis (Additional file 12: Fig. S5). Genomic alterations were not differently distributed between diabetic and non-diabetic patients (Additional file 13: Table S7). We found no statistically significant PFS differences between patients harboring any tumor genomic alteration and patients with wild-type genomic profiles (mPFS: 24 months [95% CI 14-NA] vs. 26 months; [95% CI 5-NA]; HR 0.61, 95% CI 0.19–1.96, p = 0.42). Interestingly, patients harboring tumor alterations in genes involved in DNA repair showed a trend towards worse PFS when compared to patients without alterations in DNA repair genes (median PFS 13 months [95% CI 11-NA] vs. 27 months [95% CI 20-NA]; HR 2.74; 95% CI 0.85–8.81; p = 0.09) (Additional file 14: Fig. S6A), whereas FGFR4 gene polymorphisms or ATM gene alterations were not associated with patient PFS (Additional file 14: Fig. S6B–D).

In the whole 24 months follow-up period, we observed no significant changes in any of the metabolic parameters evaluated (Fig. 1C–H; Additional file 15: Table S8). The HOMA-IR index was reduced 3 months after treatment initiation, whereas patient BMI and plasma cholesterol levels were reduced within 6 months (Fig. 1C–H; Additional file 15: Table S8). Patients experiencing higher early reduction of HOMA-IR index and plasma cholesterol concentration showed a trend towards better PFS (p = 0.055 and p = 0.086, respectively, Additional file 16: Table S9).

In conclusion, metformin plus lanreotide ATG is safe, well tolerated and active in both non-diabetic and diabetic patients with WDNETs of the GI or thoracic tract. A precocious reduction of HOMA-IR index and plasma cholesterol may predict higher clinical benefit from this treatment. Larger, prospective clinical trials should be conducted to investigate if adding metformin to SSAs results in outcome improvement when compared to SSAs alone in this clinical setting.