76. Automating immunogenomic tumor board decision-making for neoantigen cancer vaccine design

Advancements in immunogenomics and immuno-oncology have enabled the development of neoantigen vaccines, offering personalized cancer therapies by targeting cancer cell-specific somatic mutations. These mutations produce neoantigens that, when presented on tumor cells by MHC molecules, can elicit a robust and specific immune response. To date, there are 108 interventional studies listed on clinicaltrials.gov that explore the use of cancer vaccines. We have supported a number of these trials through the creation of bioinformatic pipelines, tools and procedures for the identification of patient-specific neoantigen candidates. Final prioritization of neoantigen candidates relies on manual review by an Immunogenomics Tumor Board (ITB) that meets weekly, increasing turnaround time and presenting a barrier to scaling.

Addressing this challenge, we introduce a machine learning-based approach to automate the selection of neoantigens peptides. We implemented a random forest model to train and test on existing ITB results from 21 patients and 1,324 peptides, including 297 peptides prioritized for personalized vaccine inclusion. This model aims to use features such as mutation position, driver gene status, tumor variant allele frequency, RNA expression, and other features to automatically predict whether a peptide will be accepted, rejected, or require further review for the vaccine. The model achieved an 88.89% sensitivity and 86.4% specificity, with an area under the curve of 0.933. By integrating this model into the vaccine development pipeline, we foresee a significant reduction in the time required to transition from patient sample collection to vaccine manufacturing, thereby enhancing the efficiency and scalability of personalized cancer vaccine production.