ABSTRACT
This chapter discusses a broad range of tissue phenomics use cases, including identification of scoring paradigms and signatures based on hypotheses derived from basic immune oncology research. It examines a radically different approach, largely hypothesis-free determination of prognostic descriptors by applying an unsupervised machine learning approach. Beck et al. used an unbiased data-driven approach to discover prognostically significant morphologic features in breast cancer. Tissue phenomics is the systematic discovery of quantitative descriptors for functional, morphological, and spatial patterns in histological sections that correlate with disease progression. The Immunoscore a groundbreaking diagnostic tool directly linked to the tumor microenvironment (TME) is the predictor of survival in colorectal cancer patients and has the potential to guide treatment approaches for colorectal cancer. Quantitative pathology approaches can improve significantly the prognostic value of TME-related features, like in the substratification of high-risk colorectal cancer cases building on the assessment of lymphovascular invasion, tumor budding, and nuclear grade.
