Single-dose neoadjuvant atezolizumab in resectable stage IA-IIIA non-N2 NSCLC: Tumor infiltrate analysis from the PRINCEPS trial

Phase 2 PRINCEPS trial (NCT02994576) showed that single dose neoadjuvant Atezolizumab (A, 1200 mg IV) was feasible in 30 stage IA-IIIA non-N2 NSCLC patients (treated arm, T). A comparative arm (C) was later added. We report tumor immune profiles and updated survival for T and C arms. C arm included untreated resected stage IA-IIIA non-N2 NSCLC patients. Surgery samples were profiled for residual viable tumor (RVT) and immune infiltrates. RVT cutoffs were defined with Maximally Selected Rank Test for DFS/OS (Table1). Immune infiltrates (N=316) were compared across groups by Mann Whitney U test. From 9/2019 to 8/2022, 30 patients were enrolled in C (2 non-NSCLC excluded). Mean age was 64 vs 66, squamous histology 13% vs 26%, PD-L1 neg 63% vs 46%, adjuvant ChT 20% vs 50% (T vs C). T arm tumors were enriched in CD45+ leukocytes, CD8+ T cells, HLA-DR+ CD163+ myeloids; C arm tumors showed CD115+ CD163+ macrophages and CD16Low immature neutrophils suggesting an immunosuppressive TME. In T arm, Responders (RVT≤60%) had higher lymphocyte, activated/exhausted PD-1+/TIGIT+ T cell and mature/activated CD16+ CD11bHigh neutrophil rates; Non-Responders (RVT>60%) showed more senescent CD57+ T cells and immunosuppressive (CD16-, CD163+) myeloids. In C arm, Spontaneous Regressors (RVT≤95%) had proliferating Ki67+ OX40+ T cells, CD69+ NK, exhausted CTLA4+ TIGIT+ T cells and pro-inflammatory CD16+ macrophages vs Non-Spontaneous Regressors (RVT>95%), enriched in highly activated Treg. In conclusion, neoadjuvant single dose A induced a CD8+/activated-myeloid profile that could relate to durable survival benefit. In T arm Responders showed activated/exhausted T cells and mature pro-inflammatory neutrophils, similar to Spontaneous Regressors in C arm; Non-Responders were enriched in senescent T cells and immunosuppressive myeloids. These results may indicate PD-1+/TIGIT+ exhausted T cells and pro-inflammatory myeloids as predictive biomarkers of A activity. Table 1. DFS: Disease-free Survival; OS: Overall Survival; MaxStat: Maximally Selected Rank Test Arm Treated [N=30] Comparative [N=28] Median follow-up months (95%CI) 80 (74.2-94.5) 50.6 (40.5-60.7) Overall 5 years DFS 76.3% 55.6% Overall 5 years OS 83.3% 58.7% MaxStat RVT cutoff predicting best DFS outcome 60% 95% MaxStat RVT cutoff predicting best OS outcome 60% 95% Number of patients achieving ≤RVT MaxStat cutoff for DFS 13 (Responders) 19 (Spontaneous Regressors) 5 years DFS in ≤RVT MaxStat cutoff 84.6% 61.8% 5 years DFS in >RVT MaxStat cutoff group 69.7% 41.7% DFS HR (95% CI) ≤RVT vs >RVT MaxStat cutoff groups 0.47 (0.12-1.82) 0.41 (0.13-1.37) DFS ≤RVT vs >RVT MaxStat cutoff groups log-rang p value 0.26 0.13 5 years OS in ≤RVT MaxStat cutoff group 100% 63.7% 5 years OS in >RVT MaxStat cutoff group 70% 44.4% OS HR (95% CI) ≤RVT vs >RVT MaxStat cutoff groups NA (no events) 0.52 (0.14-1.94) OS ≤RVT vs >RVT MaxStat cutoff groups log-rang p value 0.01 0.31