Abstract Through somatic mutations, cancer cells accumulate cancer-specific mutated peptides, commonly referred to as neoepitopes or neoantigens, that can be recognized by T cells. Due to their cancer-specific nature, neoantigens are a highly attractive area of study for developing cancer immunotherapies. By comparing whole-exome sequencing of patient tumor tissue and matched normal tissue, researchers have been able to identify cancer-specific neoepitopes. However, the filters that can accurately delineate immunogenic neoepitopes from non-immunogenic remain elusive. Current methods, such as the pipeline developed by the Tumor Neoantigen Selection Alliance (TESLA), typically select short peptides based on hypothesized rules of immunogenicity. Importantly, only a small percentage of these epitopes have been reported to be recognized by T cells through experimental validation assays. In this study, we systematically tested the neoantigen prediction capabilities of our in-house pipeline, Identify-Prioritize-Validate (IPV), and the TESLA pipeline. Whole-exome sequencing from 11 head and neck squamous cell carcinoma patients was used to identify somatic variants. With IPV, variants were filtered and ranked primarily based on variant allele frequencies in RNA and DNA and 20mer peptides were generated from each mutation. In parallel, the TESLA pipeline was used to rank 8- 12-mer peptides based on a combination of HLA binding and stability predictions as well as expression levels of source antigens. Pools of the top 10 peptides from both methods were generated. To enable the direct comparison of TESLA and IPV while accounting for the disparate peptide lengths produced by the pipelines, two additional pools were generated: one pool of 20mers from the TESLA pipeline and one pool of 8-12mers from IPV. The four peptide pools were individually cultured in vitro with patient-matched PBMCs and immune recognition was measured through IFNg and IL5 using fluorospot assays. The IPV pipeline consistently outperformed the TESLA pipeline in predicting neoantigens that elicited an immune response. 5 of the 11 patients showed significant immune recognition of the IPV 20mer peptide pool, only 2 patients responded to the 20mer TESLA pool, while no patients responded to the TESLA and IPV 8-12mer pools. Through this data, we prove the efficacy of IPV, and its filters, and highlight the importance of utilizing 20mer peptides over shorter peptides. Finally, we deconvoluted the positive IPV responses and were able to provide the sequences of novel immunogenic neoepitopes. Our work underscores the improvement in the predictive ability of IPV in comparison to typical pipelines that output short peptides based on parameters derived from a priori knowledge and demonstrates the ability of IPV to be utilized as a tool to identify novel neoantigens. Citation Format: Leila Yasmine Chihab, Zeynep Kosaloglu-Yalcin, Jason Greenbaum, Aaron Miller, Stephen Schoenberger, Bjoern Peters. Experimental comparison of two pipelines to identify neo-epitopes in head and neck cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2719.