VP5-2021: IMpower133: Gene expression (GE) analysis in long-term survivors (LTS) with ES-SCLC treated with first-line carboplatin and etoposide (CE) ± atezolizumab (atezo)

Abstract

Exploratory analyses to characterise LTS (patients [pts] who survived ≥18 mo post randomisation) in IMpower133 showed that more LTS were treated with atezo (anti–PD-L1) + CE than placebo (PBO) + CE (Liu et al. ESMO 2020). Here, we report the association of differential GE (single genes, T-effector [Teff] and B-cell signatures) and published SCLC subtypes (Gay et al. Cancer Cell 2021) in IMpower133 LTS. Treatment-naive pts with ES-SCLC were randomised 1:1 to four 21-day cycles of C (AUC 5 mg/mL/min IV, day 1) + E (100 mg/m2 IV, days 1-3) combined with atezo (1200 mg IV, day 1) or PBO, followed by maintenance atezo or PBO until disease progression or toxicity. Co-primary endpoints were PFS and OS. Differential GE was analysed using RNA-sequencing (RNA-seq) data in LTS and non-LTS. OS was assessed by Teff and B-cell gene signature expression in both arms. Distribution of recently identified neuroendocrine, non-neuroendocrine and immune-infiltrated subtypes were also assessed in LTS and non-LTS. 271 pts were in the RNA-seq biomarker-evaluable population, 253 of whom had sufficient follow-up for LTS classification (LTS RNA-seq BEP; LTS, n=64; non-LTS, n=189). Within the LTS RNA-seq BEP, significantly more LTS were in the atezo (n=39, 32%) vs PBO arm (n=25, 19%; P=0.027). In both arms within the LTS RNA-seq BEP, distributions of Teff- and B-cell−related genes and gene signatures (≥ median vs < median) were similar in LTS. Exploratory OS efficacy analyses in the LTS RNA-seq BEP suggest that benefit was observed in favour of atezo vs PBO across all Teff and B-cell expression subgroups (high [≥ median]: Teff, HR, 0.65 [95% CI, 0.43, 0.97]; B-cell, HR, 0.75 [95% CI, 0.51, 1.11], and low [< median]: Teff, HR, 0.75 [95% CI. 0.52, 1.11]; B-cell, HR, 0.66 [95% CI: 0.45, 0.97]). Among the LTS RNA-seq BEP in both arms, GE subtype distribution was broadly similar for LTS. In IMpower133, more LTS were treated with atezo + CE vs PBO + CE. Enhanced immune-related signaling was seen in LTS in both arms. Atezo and PBO had similar GE-defined subtype distribution, suggesting that subtype associations with outcome may be prognostic.