ATIM-19. CATEGORIZING IMMUNE RESPONDERS WITH FUSION METRICS AND SIMULATION FOR ASSOCIATION TO SURVIVAL AND PROGRESSION FREE SURVIVAL WITH IMMUNE RESPONSE IN HLA-A2+ PATIENTS WITH GBM FROM A PHASE 2 TRIAL OF DENDRITIC CELL (DC) IMMUNOTHERAPY (ICT-107)

BACKGROUND: Detailed confirmatory testing and analysis verifies and strengthens the association between clinical outcomes and immune response of HLA-A2+ patients enrolled in a randomized phase 2 trial of ICT-107. METHODS: 77 HLA-A2+ patients, randomized 2:1, received ICT-107 (autologous DCs incubated with 6 synthetic peptide CTL epitopes targeting GBM tumor/stem cell-associated antigens, including the four HLA-A2-restricted antigens HER-2, TRP-2, gp100, and IL-13Ralpha2) or matching control (un-incubated DC). Multimer testing was performed on a subset of these patients. The pioneering analysis heuristic of fusion metrics used in conjunction with Monte Carlo simulation was used to identify multimer immune responders. P-values between dependent variables and multiple overall survival (OS) or progression-free survival (PFS) metrics was performed using log-rank test and Fisher's exact test. RESULTS: HLA-A2+ patients consistently continued to show evidence of immune response being associated with both OS and PFS. Multimer immune responders independently confirmed the ELISpot immune responder associations between assignment group (p=0.0308), and initial OS and PFS (p=0.0043 and 0.0352, respectively). Combining ELISpot and multimer responders strengthened or maintained associations with all OS and PFS metrics. Notable significant associations were determined when data was stratified by treatment group in both treatment and placebo subgroups, leading to speculation of the possible positive effects of DCs alone. This finding supports changing the placebo in the Phase III study from dendritic cells to PBMCs. CONCLUSIONS: The robust associations identified between OS and PFS with immunologic response, explored using both multimer and ELISpot analysis to determine immune response with fusion metrics in a Monte Carlo setting, provide support for the efficacy of ICT-107 to induce peptide-specific T cell responses in HLA-A2+ patients.