Diego2021 - Improved Prediction of ICB Efficacy Across Multiple Cancer Types using Random Forests

In this study, the authors had developed a machine learning model to predict immune checkpoint blockade (ICB) response by integrating genomic, molecular, demographic and clinical data from a curated cohort (MSK-IMPACT) with 1479 patients treated with ICB across 16 different types of cancer. This model significantly outperformed the predictions based on Tumor Mutational Burden (TCB). This model uses two types of random forests, one uses 16 features and the other uses 11 features. These features are selected based on their permutation importance score. The model was deployed on docker to reproduce the results and the data was shared to promote FAIReR sharing of machine learning models.