Impact of ramucirumab plus erlotinib on circulating cell-free DNA from patients with untreated metastatic non-small cell lung cancer with EGFR-activating mutations (RELAY phase 3 randomized study)

BACKGROUND: An exploratory, proof-of-concept, liquid biopsy addendum to examine biomarkers within cell-free DNA (cfDNA) in the RELAY phase 3, randomized, double-blind, placebo-controlled study was conducted. RELAY showed improved progression-free survival (PFS) with ramucirumab (RAM), a human immunoglobulin G1 vascular endothelial growth factor receptor 2 antagonist, plus erlotinib (ERL), a tyrosine kinase inhibitor, compared with placebo (PL) plus ERL. METHODS: Treatment-naïve patients with endothelial growth factor receptor (EGFR)-mutated metastatic non-small cell lung cancer were randomized (1:1) to RAM + ERL or PL + ERL. Plasma samples were collected at baseline, on treatment, and at 30-day post-study treatment discontinuation follow-up. Baseline and treatment-emergent gene alterations and EGFR-activating mutation allele counts were investigated by next-generation sequencing (NGS) and droplet digital polymerase chain reaction (ddPCR), respectively. cfDNA concentration and fragment size were evaluated by real-time polymerase chain reaction and the BioAnalyzer. Patients with a valid baseline plasma sample were included (70 RAM + ERL, 61 PL + ERL). RESULTS: TP53 mutation was the most frequently co-occurring baseline gene alteration (43%). Post-study treatment discontinuation EGFR T790M mutation rates were 54.5% (6/11) and 41.2% (7/17) by ddPCR, and 22.2% (2/9) and 29.4% (5/17) by NGS, in the RAM + ERL and PL + ERL arms, respectively. EGFR-activating mutation allele count decreased at Cycle 4 in both treatment arms and was sustained at follow-up with RAM + ERL. PFS improved for patients with no detectable EGFR-activating mutation at Cycle 4 vs. those with detectable EGFR-activating mutation. Total cfDNA concentration increased from baseline at Cycle 4 and through to follow-up with RAM + ERL. cfDNA fragment size was similar between treatment arms at baseline [mean (standard deviation) base pairs: RAM + ERL, 173.4 (2.6); PL + ERL, 172.9 (3.2)] and was shorter at Cycle 4 with RAM + ERL vs. PL + ERL [169.5 (2.8) vs. 174.1 (3.3), respectively; P<0.0001]. Baseline vs. Cycle 4 paired analysis showed a decrease in cfDNA fragment size for 84% (48/57) and 23% (11/47) of patient samples in the RAM + ERL and PL + ERL arms, respectively. CONCLUSIONS: EGFR-activating mutation allele count was suppressed, total cfDNA concentration increased, and short fragment-sized cfDNA increased with RAM + ERL, suggesting the additional anti-tumor effect of RAM may contribute to the PFS benefit observed in RELAY with RAM + ERL vs. PL + ERL. TRIAL REGISTRATION: ClinicalTrials.gov; identifier: NCT02411448