Modified FOLFOX plus/minus nivolumab and ipilimumab vs FLOT plus nivolumab in patients with previously untreated advanced or metastatic adenocarcinoma of the stomach or gastroesophageal junction: The randomized phase 2 IKF-S628 Moonlight trial of the Arbeitsgemeinschaft Internistische Onkologie (AIO)

4050 Background: FOLFOX plus nivolumab has become standard of care for first-line therapy of patients (pts) with esophagogastric adenocarcinomas (EGA). The AIO-STO-0417 trial (Moonlight) is a multi-cohort treatment optimization trial that evaluates the combination of FOLFOX alone (Arm B) vs FOLFOX plus nivolumab (nivo) and ipilimumab (ipi) administered in parallel (Arm A/A1) or sequentially (Arm A2) and FLOT plus nivolumab administered in parallel (Arm C) for 1L-treatment of metastatic or advanced inoperable Her-2 negative EGA. The aim is to generate signals whether a. dual checkpoint inhibition or b. a triplet chemotherapy is beneficial in the context of nivolumab therapy for this disease. Methods: Pts were randomized 1:1 to Arm A (mFOLFOX q2w plus nivo 240 mg q2w + ipi 1 mg/kg q6w administered in parallel) or B (mFOLFOX alone). In a subsequent phase pts were randomized 1:2 to Arm A1 (identical to Arm A) or A2 (three cycles of mFOLFOX followed by nivo q2w + ipi q6w, with optional repetition). In a final phase, all pts were allocated to single Arm C (FLOT q2w + nivo q2w). The primary endpoint was progression-free survival (PFS) based on the ITT population for Arm A vs Arm B and PFS rate at 6 months (PFS@6) for Arms A2 and C. Main secondary endpoints were PFS and ORR. Here, we present results for Arms A/A1, A2 and B. Results: A total of 262 pts were enrolled, Arm A (n=60), Arm B (n=60), Arm A1 (n=30), Arm A2 (n=60) and Arm C (n=52, results presented elsewhere). Baseline characteristics were comparable in all arms. Overall PD-L1 expression (CPS ≥1) was low with 41% and balanced between Arms A/A1, A2 and B. Analysis of pooled Arms A/A1 (n=90) showed an increase in toxicity (pts with AEs grade ≥3 88% vs 65% in Arm B and 75% in Arm A2, treatment related SAEs grade ≥3 A/A1 41% vs B 18% vs A2 17%) but not in activity. Arm A/A1 compared with Arm A2 was more effective in terms of PFS@6 (48% vs 30%), median PFS 5.8 vs 4.0 months and objective response rate (ORR) 46% vs 30%. Conclusions: Albeit the small number of pts in each cohort we conclude that in the first-line setting of metastatic EGA. a) Chemo plus dual checkpoint inhibition administered in parallel is associated with an increase in toxicity but not activity. b) Although associated with lower toxicity the use of sequential chemo followed by IO monotherapy is insufficient. The relatively low numbers of patients with PD-L1 CPS ≥1 may have impacted these results. Clinical trial information: NCT03647969 .