ImmuniT Platform for Improved Neoantigen Prediction in Lung Cancer

Background: Lung cancer remains the leading cause of cancer-related mortality, with many patients responding poorly to immunotherapy due to limited tumor recognition. Neoantigen-based strategies offer a promising solution, but current discovery methods often miss key targets, particularly those with low or heterogeneous expression. To address this, we developed ImmuniT, a three-phase platform for enhanced neoantigen discovery and validation. Methods: Under an IRB-approved protocol, patients with lung cancer consented to tumor collection for ex vivo processing to modulate antigen expression. Autologous T cells from matched blood were co-cultured with treated cancer cells to expand tumor-reactive populations. The nextneopi pipeline integrated mutational, transcriptomic, and HLA data to predict candidate neoantigens, which were validated using MHC epitope tetramer staining. Results: In five patient samples, ImmuniT identified a broader spectrum of neoantigens and induced stronger T cell activation in vitro compared to conventional approaches. Notably, in one case, two neoantigens missed by standard methods were confirmed to elicit tumor-specific T cell responses in both the tumor-infiltrating and peripheral compartments. Conclusions: These findings highlight ImmuniT’s potential to expand the repertoire of actionable tumor antigens and improve personalized immunotherapy strategies, particularly for patients with limited response to existing treatments. Note: This submission provides de-identified, summary-level results from NSCLC patient samples, including normalized RNA-seq counts, differential expression (DE) results, and copy number variation (CNV) calls. Raw sequencing files (FASTQ/BAM/VCF) are not deposited due to informed consent limitations; see README. Case-series of 5 lung cancer samples; processed-only GEO submission with per-sample normalized counts and CNV calls; study-level DE results. *************************************************************** Raw files for human/patient samples were not submitted to GEO due to concerns about submitting personally identifiable sequence data for open access. ***************************************************************