e14037 Background: Glioblastomas are the most common primary malignant brain tumors, comprising 2% of all adult cancers. Despite recent advancements in genetic classification and targeted therapies for specific mutations, there has been no significant improvement in overall life expectancy for the majority of patients. Last year, our research group reported promising results from a phase II prospective trial involving allogeneic dendritic cell vaccination. To date, six out of 37 reported cases remain alive without tumor recurrence, exhibiting a performance status of 90 to 100% according to the Karnofsky Performance Scale (KPS) three years after the initiation of vaccination. Methods: In this study, we focused on the immune cell infiltration observed at the time of tumor excision in enrolled patients. We utilized flow cytometry to quantify cells expressing membrane CD86, CD45, HLA-DR, PD-L1, CD11b, and CD14, in addition to assessing IDH-1 status, age, gender, KPS, gender, ECOG, symptomatology, and the number of vaccines administered during the trial. Our goal was to predict responsiveness in terms of overall survival using a neural network prediction algorithm with three hidden nodes and five seeds for improved reproducibility. We employed the random holdback method for algorithm validation. Results: From the 37 enrolled patients, we successfully obtained viable tumor cells from 19 frozen samples. The model achieved an RSquare value of 0.71 after training and 0.66 after validation (where 1 indicates a perfect model and 0 implies no improvement over a constant model). Similar RSquare values in both training and validation sets indicate strong predictive power. Positive predictors of overall survival included HLA-DR positivity, PD-L1 positivity, IDH-1 mutation status, the presence of focal deficits, and higher KPS scores, while negative predictors were CD86 expression, older age, symptoms of intracranial hypertension (ICH), and higher ECOG performance status. Conclusions: In conclusion, the neural network algorithm described herein enables the prediction of responsiveness to dendritic cell vaccination in terms of overall survival, based on clinical features and the profile of immune cell infiltration observed at the time of tumor excision.