Tumor Microenvironment Evaluation for Gastrointestinal Cancer in the Era of Immunotherapy and Machine Learning

Abstract

A dynamic and mutualistic interplay between tumor cells and the surrounding tumor microenvironment (TME) triggered the initiation, progression, metastasis, and therapy response of solid tumors. Recent clinical breakthroughs in immunotherapy for gastrointestinal cancer conferred considerable attention to the estimation of TME, and the maturity of next-generation sequencing (NGS)-based technology contributed to the availability of increasing datasets and computational toolbox for deciphering TME compartments. In the current review, we demonstrated the components of TME, multiple methodologies involved in TME detection, and prognostic and predictive TME signatures derived from corresponding methods for gastrointestinal cancer. The TME evaluation comprises traditional, radiomics, and NGS-based high-throughput methodologies, and the computational algorithms are comprehensively discussed. Moreover, we systemically elucidated the existing TME-relevant signatures in the prognostic, chemotherapeutic, and immunotherapeutic settings. Collectively, we highlighted the clinical and technological advances in TME estimation for clinical translation and anticipated that TME-associated biomarkers may be promising in optimizing the future precision treatment for gastrointestinal cancer.

Keywords: chemotherapy; gastrointestinal cancer; immunotherapy; machine learning; tumor microenvironment.