Abstract 5476: Veracyte Biopharma Atlas for colorectal cancer: Combining multi-parameter approach and machine learning to capture the complexity of the tumor immune contexture

Abstract

Abstract (a) Understanding the immune contribution has deeply modified the standard of care of patients, especially with the advent of immunotherapies (ITs). However, ITs are still efficient only on a minority of the patients, especially for colorectal cancer (CRC), a high-incidence cancer. One hypothesis to explain the heterogeneity of the response to IT is that the presence or absence of immune factors within the tumor micro-environment (TME) may critically impact the efficacy of treatments. In this context, a deeper molecular understanding of the TME could be key to stratify patients. Here, we describe a proof of concept of the Veracyte Biopharma Atlas for CRC based on a multi-omics approach which integrates a unique range of assays allowing to decipher the immune contexture of tumors and offer a personalized and dynamic “fingerprint” of the interactions between the tumor and the immune system. (b) Veracyte Biopharma Atlas characterizes the tumor microenvironment by analyzing information from 3 different modalities: (i) proteomics (Brightplex® multiplex immunohistochemistry panels), (ii) transcriptomics (RNA sequencing) and (iii) genomics (somatic mutations by exome sequencing). This deep characterization of the tumor and its micro-environment requires the integration of thousands of features. The innovative Veracyte Biopharma Atlas approach relies on linear (multimodal factor analysis) and non-linear (self-organizing map) approaches with graphical representation of the data to integrate this multi-omics information. (c) Gene expression profiling, genomics analysis and proteomics by multiplex spatial technology were performed on 47 CRC FFPE samples to characterize the immune contexture of this disease. Multimodal analysis was used to derive a unique atlas map highlighting clusters distinguished by complex features and their immune tumor contexture. Indeed, as examples, the atlas map allows the identification of a group of patients displaying high immune cell infiltration, including T cell and macrophages. They are also characterized by an immune genes signature. By contrast, another group of patients is poorly infiltrated by immune cells but exhibits an elevated M2/M1 ratio, favorable to tumor development. (d) This proof of concept lays the foundations of an atlas for CRC that will be extended to samples associated with clinical data to build a map of the disease. Later-on, by projecting data from new specimens onto this referential map, it will become possible to establish an Immunogram of the tumor microenvironment of individual patients to define the most appropriate treatment but also facilitate the development of new therapies. Citation Format: Laurent Vanhille, Alboukadel Kassambara, Chafik Hamdad, Margaux Mercadal, Alexia Papadopoulos, Pernelle Outters, Théo Vasse, Vanina Leca, Jérôme Galon, Jacques Fieschi. Veracyte Biopharma Atlas for colorectal cancer: Combining multi-parameter approach and machine learning to capture the complexity of the tumor immune contexture. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5476.