Summary Cancer cells can leverage several cell-intrinsic and -extrinsic mechanisms to escape immune system recognition. The inherent complexity of the tumor microenvironment, with its multicellular and dynamic nature, poses great challenges for the extraction of biomarkers of immune response and immunotherapy efficacy. Here, we use RNA-sequencing (RNA-seq) data combined with different sources of prior knowledge to derive system-based signatures of the tumor microenvironment, quantifying immune-cell composition and intra- and intercellular communications. We applied multi-task learning to these signatures to predict different hallmarks of immune responses and derive cancer-type-specific models based on interpretable systems biomarkers. By applying our models to independent RNA-seq data from cancer patients treated with PD-1/PD-L1 inhibitors, we demonstrated that our method to Estimate Systems Immune Response (EaSIeR) accurately predicts therapeutic outcome. We anticipate that EaSIeR will be a valuable tool to provide a holistic description of immune responses in complex and dynamic systems such as tumors using available RNA-seq data.

Immunotherapy with checkpoint blockers (ICBs), aimed at unleashing the immune response toward tumor cells, has shown a great improvement in overall patient survival compared to standard therapy, but only in a subset of patients. While a number of recent studies have significantly improved our understanding of mechanisms playing an important role in the tumor microenvironment (TME), we still have an incomplete view of how the TME works as a whole. This hampers our ability to effectively predict the large heterogeneity of patients' response to ICBs. Systems approaches could overcome this limitation by adopting a holistic perspective to analyze the complexity of tumors. In this Mini Review, we focus on how an integrative view of the increasingly available multi-omics experimental data and computational approaches enables the definition of new systems-based predictive biomarkers. In particular, we will focus on three facets of the TME toward the definition of new systems biomarkers. First, we will review how different types of immune cells influence the efficacy of ICBs, not only in terms of their quantification, but also considering their localization and functional state. Second, we will focus on how different cells in the TME interact, analyzing how inter- and intra-cellular networks play an important role in shaping the immune response and are responsible for resistance to immunotherapy. Finally, we will describe the potential of looking at these networks as dynamic systems and how mathematical models can be used to study the rewiring of the complex interactions taking place in the TME.

Purpose: Predictive biomarkers of immune checkpoint inhibitors (ICIs) efficacy are currently lacking for non-small cell lung cancer (NSCLC). Here, we describe the results from the Anti-PD-1 Response Prediction DREAM Challenge, a crowdsourced initiative that enabled the assessment of predictive models by using data from two randomized controlled clinical trials (RCTs) of ICIs in first-line metastatic NSCLC. Methods: Participants developed and trained models using public resources. These were evaluated with data from the CheckMate 026 trial (NCT02041533), according to the model-to-data paradigm to maintain patient confidentiality. The generalizability of the models with the best predictive performance was assessed using data from the CheckMate 227 trial (NCT02477826). Both trials were phase III RCTs with a chemotherapy control arm, which supported the differentiation between predictive and prognostic models. Isolated model containers were evaluated using a bespoke strategy that considered the challenges of handling transcriptome data from clinical trials. Results: A total of 59 teams participated, with 417 models submitted. Multiple predictive models, as opposed to a prognostic model, were generated for predicting overall survival, progression-free survival, and progressive disease status with ICIs. Variables within the models submitted by participants included tumor mutational burden (TMB), programmed death ligand 1 (PD-L1) expression, and gene-expression-based signatures. The best-performing models showed improved predictive power over reference variables, including TMB or PD-L1. Conclusion: This DREAM Challenge is the first successful attempt to use protected phase III clinical data for a crowdsourced effort towards generating predictive models for ICIs clinical outcomes and could serve as a blueprint for similar efforts in other tumor types and disease states, setting a benchmark for future studies aiming to identify biomarkers predictive of ICIs efficacy.

Immunotherapy with immune checkpoint blockers (ICB) is associated with striking clinical success, but only in a small fraction of patients. Thus, we need computational biomarker-based methods that can anticipate which patients will respond to treatment. Current established biomarkers are imperfect due to their incomplete view of the tumor and its microenvironment. We have recently presented a novel approach that integrates transcriptomics data with biological knowledge to study tumors at a more holistic level. Validated in four different solid cancers, our approach outperformed the state-of-the-art methods to predict response to ICB. Here, we introduce estimate systems immune response (easier), an R/Bioconductor package that applies our approach to quantify biomarkers and assess patients’ likelihood to respond to immunotherapy, providing just the patients’ baseline bulk-tumor RNA-sequencing (RNA-seq) data as input.

Cell-cell interaction networks are pivotal in cancer development and treatment response. These networks can be inferred from data, however this process often combines data from multiple patients, and/or creates networks on a cell-types level. It creates a good average overview of cell-cell interaction networks, but fails to capture patient heterogeneity and/or masks potentially relevant local network structures. We propose a mathematical model based on random graphs (called RaCInG) to alleviate these issues using prior knowledge on potential cellular interactions and patient's bulk RNA-seq data. We have applied RaCInG to extract 444 network features related to the tumor microenvironment, unveiled associations with immune response and subtypes, and identified cancer-type specific differences in inter-cellular signaling. Additionally, we have used RaCInG to explain how immune phenotypes regulated by context-specific intercellular communication affect immunotherapy response. RaCInG is a modular pipeline, and we envision its application for cell-cell interaction reconstruction in different contexts.