Prediction of Immune Checkpoint Inhibition Response

This technical report presents a methodology for predicting responses to immune checkpoint inhibitors (ICIs) using tumor RNA sequencing data. The study outlines the challenges in predicting ICI therapy responses and emphasizes the importance of biomarkers such as tumor mutational burden (TMB), microsatellite instability (MSI), infiltrating lymphocytes (TILs), and immune gene expression signatures. A suite of data processing and analysis pipelines has been developed to extract these features from RNA sequencing profiles, allowing for a more efficient and integrated prediction model. The report details the establishment of a variant calling pipeline for identifying somatic mutations from RNA-Seq data and compares the accuracy of TMB and eTMB measurements across various cancer types. Additionally, the report discusses the accuracy of RNA-Seq predictions for MSI status and the integration of multiple features into a single logistic regression model, which significantly enhances prediction accuracy for ICI therapy responses. The findings underscore the potential of the RNA-Seq Bio-IT model in clinical applications.