3017 Background: Cancer vaccines have been able to induce T-cell responses in cancer patients but rarely demonstrated a correlation of immune responses with clinical benefit. Here, we present immunological results of a phase 1 study with IMA901, a therapeutic cancer vaccine based on multiple novel synthetic tumor-associated peptides (TUMAP) identified as being naturally presented in primary renal cell carcinoma (RCC) tissues. Methods: The HLA peptidome of 32 primary RCC samples was systematically investigated using a combination of mass spectrometry, gene expression profiling and in vitro human T-cell assays. 9 HLA-A*02- and 1 HLA-DR- restricted TUMAPs derived from 9 different tumor antigens were selected and designated IMA901. 28 HLA-A*02-positive stage III/IV RCC patients were enrolled in a single arm, multicenter study and received 8 vaccinations on days 1, 2, 3, 8, 15, 22, 36, and 64 each consisting of 4.5 mg IMA901 (including a HBV-derived viral marker peptide) and 75 μg GM-CSF as immune adjuvant. T-cell responses using IFN-γ ELISPOT and HLA multimer analysis and CD4+ Foxp3+ regulatory T cell (Treg) levels were measured in peripheral blood. Results: In vivo IMA901- induced specific T-cell responses were detected to the HBV marker peptide (52% of 27 evaluable patients), at least one TUMAP (74%) or multiple TUMAPs (30%). T-cell responses were detectable already at day 15, peaked subsequently and were sustainable until follow-up in the majority of patients. Most importantly, patients eliciting multiple responses to TUMAPs significantly showed a higher clinical benefit rate (SD+PR; p=0.018) and lower Treg levels at study onset (p=0.016). No correlation of HBV marker peptide responses with either clinical benefit or Treg levels was observed. Conclusions: IMA901 rapidly induced T-cell responses in a majority of advanced RCC patients. A clinical mode of action is strongly supported by the significant correlation of multiple T-cell responses with clinical benefit. CD4+ Foxp3+ Tregs seem to play an active role in limiting the broadness of T-cell responses. Furthermore, our data suggest that we can predict in vivo immunogenicity of cancer vaccine antigens by our in vitro drug discovery approach. No significant financial relationships to disclose.