BACKGROUND: We have been conducting clinical trials of peptide vaccination targeting WT1 gene product against malignant gliomas. Given the results of phase II trial for recurrent malignant gliomas (Izumoto, Hashimoto et al., J Neurosurg 2008), we performed recursive partitioning analyses (RPA) (Carson et al., J Clin Oncol, 2007). We have also completed phase I trial of combined temozolomide (TMZ) and WT1 vaccination for newly diagnosed glioblastomas (GBM). METHODS: After the Kaplan-Meier analyses giving median PFS and OS, 70 patients with recurrent gliomas who entered into the phase II trial were classified into RPA classes, following Carson et al. Seven patients were enrolled into phase I study for newly diagnosed GBM. They underwent combined TMZ and WT1 vaccination just after the initial RT/TMZ. In both studies, immunological response was analyzed on WT1-specific T cells induction in peripheral blood, delayed type hypersensitivity as well as intra-tumoral immune cell analysis when surgery was indicated after vaccination. In the phase II study, predictive or prognostic indicators were extensively investigated. Although we have originally used RECIST criteria for MRI evaluation, we tried to directly compare RECIST, Macdonald, RANO and irRC using datasets from 50 GBM patients in terms of survival prediction. RESULTS: Median OS in 7 classes (I to VII) obtained by RPA showed the same tendency for OS ranking, and 6 of 7 classes (class II to VII) presented longer OS as compared with Carsons. Hazard ratio calculated for WT1 vaccination was 0.372, which was better than those with other agents used in previously published phase I/II studies. Seven patients with combined TMZ and WT1 vaccination tolerated well and showed no adverse events above CTCAE grade III, except for grade III lymphocytopenia in 71.4% of patients possibly due to preceded RT/TMZ therapy. PFS ranged from 5.2 to 37.1 months (median 31.5 months). We have confirmed the induction of WT1 specific T cells in the majority of patients of both studies. Patients showed longer survival statistically, who had stronger expression of WT1 protein on the tumor cells, DTH turning to be positive and methionine uptake decreased within 3 months. An evaluation by RANO was statistically correlated best with OS among those 4 criteria. CONCLUSIONS: Multivariate analysis with RPA showed the efficacy of WT1 vaccination with some predictive indicators. The following phase I study revealed the safe combination of TMZ and WT1, and throw light on the new multi-modal strategy. SECONDARY CATEGORY: Clinical Neuro-Oncology.