Plasma biomarker dynamics following ramucirumab treatment and survival analysis after ramucirumab treatment failure in patients with advanced gastric cancer.

Abstract

79 Background: Ramucirumab (RAM) showed significantly improved survivals of gastric cancer in the second line with paclitaxel or single use. It is known that RAM causes pharmacodynamic changes of plasma VEGF-A family levels during treatment. However, its clinical significance is still unclear. The aim of this study is to reveal the changes of plasma biomarker levels until disease progression, and to examine associations between plasma biomarker levels at disease progression and survival after RAM treatment failure. Methods: Plasma samples were collected at three points: base line, day 8, and disease progression. Nine kinds of plasma biomarker related to angiogenesis: VEGF-A, C, D, PlGF, sVEGFR-1, 2, Neuropilin-1 (NRP1), Angiopietin-1, and SDF-1α, were measured by means of ELISA. The dynamics of plasma biomarkers were compared using the ANOVA. Patients were dichotomized by optimal cut-off value in each biomarker. Survival after RAM treatment failure was estimated by Kaplan-Maier methods and compared by log-rank test. Results: Plasma samples were collected from 21 patients. Median age was 67 y.o and female was dominant (57%). Plasma VEGF-A and PlGF levels sharply increased at day 8 and these higher plasma levels were sustained until disease progression: the median VEGF-A and PlGF levels at baseline, day 8, and disease progression were 20.2, 350.2, and 596.7 pg/ml for VEGF-A (p<0.001) and 10.5, 223.4, and 261.1 pg/ml for PlGF (p<0.001), respectively. Conversely, plasma NRP1 levels consistently decreased during treatment course: the median NRP1 levels at baseline, day 8, and disease progression were 444.3, 309.2, and 230.9 mg/ml (p=0.001). The median survival after disease progression was 3.1 months (95%CI 2.5-3.7). Patients with higher VEGF-A and NRP1 levels showed shorter survival with 4.5 vs. 0.9 months (HR 3.15 95%CI 1.6-6.4 p<0.001) for VEGF-A and 8.8 vs. 2.6 months (HR 2.92 95% CI 1.4-6.3 p=0.002) for NRP1. Conclusions: Our data suggest that not only a clue of mechanism of acquired resistance of RAM but necessity of new treatment strategy after RAM treatment failure. Further understanding of molecular correlates and clinical validation are warranted.