BEAT-meso: A randomized phase III study of bevacizumab (B) and standard chemotherapy (C) with or without atezolizumab (A), as first-line treatment (TX) for advanced pleural mesothelioma (PM)—Results from the ETOP 13-18 trial.

LBA8002 Background: The currently approved frontline TXs for PM are the combination of ipilimumab/nivolumab or platinum plus pemetrexed. The addition of B to C has been shown to improve overall survival in a randomized clinical trial. While combined immunotherapy or single agent immunotherapy with C is superior to C alone, there is potential for a synergistic triple combination of C, B, and immunotherapy. Methods: BEAT-meso (NCT03762018) is an international open-label, 1:1 randomized phase III trial, stratified by histology and stage. The objective is to determine the efficacy and safety of adding A (1200 mg, Q3W until progression) to B (15mg/kg, Q3W until progression) and standard C (4-6 cycles of carboplatin AUC5 with pemetrexed 500 mg/m 2 , Q3W), as first-line TX for advanced PM. The trial is designed to detect an increase in the median overall survival (OS, primary endpoint) with the addition of A, aiming for a hazard ratio (HR) of 0.708, at 2.5% 1-sided alpha and 82% power (284 deaths, sample size 400 patients (pts)). In the pre-specified interim efficacy analysis (80% of the events, 01/2023), boundary was not crossed, and the trial continued to completion. Secondary endpoints include progression-free survival (PFS), objective response rate (ORR), disease control rate, duration of response (DoR), adverse events (AEs) assessed by CTCAE v5.0 and symptom-specific and global quality of life (QoL). Results: Between 04/2019 and 03/2022, a total of 400 pts was randomized, 200 per arm. The median age was 70 years, 79% were male, 50% were former smokers, 65% had ECOG performance status 1 and 78% had epithelioid histology. At a median follow-up of 35 months (m) (as of 1/09/2023), median OS was 20.5m [95% CI: 17.5-23.3] in the ABC and 18.1m [15.7-20.9] in the BC arm (deaths: 145 & 150; HR ABC vs BC =0.84; [0.66 - 1.06], 2-sided stratified p=0.14, ITT final analysis). PFS was significantly longer in ABC with median 9.2m [8.1-10.9] vs 7.6m [6.9-8.3] in BC (HR=0.72; [0.59 - 0.89], 2-sided stratified p=0.0021). Histology shows a significant TX interaction for both PFS and OS. The OS HR is 0.51 [0.32-0.80] for non-epithelioid and 1.01 [0.77-1.32] for epithelioid (interaction p=0.012). In an exploratory analysis, post-progression OS was significantly different between the two arms, adjusted for post-progression TX (HR=0.76; [0.58 - 0.99]). The ORR was 55% in ABC and 49% in BC (p=0.27), while median DoR was 8.2m [6.8-9.7] in ABC and 5.6m [4.8-7.0] in BC (p=0.0041). Global QoL change was not significantly different between the two arms. Grade≥3 TX-related AEs occurred in 55% of pts in ABC and 47% of pts in BC (grade 5: 7 and 1 pt, respectively). Conclusions: The significant increase in median PFS with the addition of A did not translate into a significant increase in median OS. ABC demonstrated superiority over BC in non-epithelioid cases. Clinical trial information: NCT03762018 .