Evaluation of circulating VEGF based biomarkers in INTEGRATE: A randomized phase II double-blind placebo-controlled study of regorafenib in refractory advanced oesophagogastric cancer (AOGC)—A study by the Australasian Gastrointestinal Trials Group (AGITG).

Abstract

64 Background: The INTEGRATE study evaluated activity of regorafinib (REG) v placebo (PBO) in 147 eligible patients with refractory AOGC. REG was highly effective in prolonging progression free survival (PFS). Differences between regions (i.e. Australia New Zealand/Canada (ANZ/CAN) vs Korea) were found in the magnitude of effect, but REG was effective across all regions and subgroups. We report on an exploratory analysis of VEGF biomarkers to identify predictive/prognostic markers. Methods: Protein biomarkers IL8, VEGF-A,-B,-C-D, soluble(s)VEGFR-1,-2-3 were analysed in plasma at baseline (BL) by multiplex immunoassays (Bio-Plex,BioRad) or ELISA (Abnova). Spearman statistics were used to quantify correlations between markers. Wilcoxon Rank-Sum tests were used to compare markers across regions. The prognostic and predictive value of markers was determined using cox proportional hazards analysis of PFS. Results: There were moderate-to-strong correlations between BL levels of IL8 and VEGF-C (ρ = 0.68), IL8 and VEGF-D (ρ = 0.66), VEGF-A and VEGF-C (ρ = 0.68), VEGF-A and sVEGFR-1 (ρ = 0.54); and a modest negative correlation between VEGF-D and sVEGRF-1 (ρ = -0.33). The regions differed according to BL levels of: VEGF-A (higher in ANZ/CAN; p = 0.0015), VEGF-B (higher in Korea; p = 0.0003), VEGF-D (higher in Korea; p <.0001), and sVEGFR-1 (higher in ANZ/CAN; p <.0001). Adjusting for treatment group, there were statistically significant negative associations between PFS and BL IL8 (p = 0.047), VEGF-A (p = 0.037) and sVEGFR-1 (p = 0.045). There was no convincing statistical evidence that any BL plasma biomarker modified the effect of REG. The effect of region on effectiveness of REG was maintained when evaluated in conjunction with BL biomarkers individually and in combination. Conclusions: Highplasma IL8, VEGF-A and sVEGFR-1 may be adverse prognostic factors. A predictive VEGF blood based biomarker remains elusive. A broader biomarker study including markers beyond the VEGF axis and tissue based markers is ongoing. Clinical trial information: ACTRN12612000239864.