Generating counterfactual explanations of tumor spatial proteomes to discover effective strategies for enhancing immune infiltration

TL;DR

This paper introduces a machine learning framework that uses spatial omics data to generate counterfactual tumor microenvironment modifications, aiming to identify strategies that enhance T-cell infiltration in solid tumors.

Contribution

It develops a novel counterfactual optimization method leveraging spatial omics and deep learning to predict tumor perturbations that could improve immune infiltration.

Findings

Identified perturbations that increase T-cell infiltration in melanoma, colorectal, and breast tumors.

Demonstrated the framework's ability to suggest combinatorial therapeutic strategies.

Applied the method across multiple cancer types with promising results.

Abstract

The tumor microenvironment (TME) significantly impacts cancer prognosis due to its immune composition. While therapies for altering the immune composition, including immunotherapies, have shown exciting results for treating hematological cancers, they are less effective for immunologically-cold, solid tumors. Spatial omics technologies capture the spatial organization of the TME with unprecedented molecular detail, revealing the relationship between immune cell localization and molecular signals. Here, we formulate T-cell infiltration prediction as a self-supervised machine learning problem and develop a counterfactual optimization strategy that leverages large scale spatial omics profiles of patient tumors to design tumor perturbations predicted to boost T-cell infiltration. A convolutional neural network predicts T-cell distribution based on signaling molecules in the TME provided by imaging mass cytometry. Gradient-based counterfactual generation, then, computes perturbations predicted to boost T-cell abundance. We apply our framework to melanoma, colorectal cancer liver metastases, and breast tumor data, discovering combinatorial perturbations predicted to support T-cell infiltration across tens to hundreds of patients. This work presents a paradigm for counterfactual-based prediction and design of cancer therapeutics using spatial omics data.