Immune response assessment in a colorectal cancer microenvironment using digital pathology analytics

Tumor infiltrating lymphocytes (TIL) have been associated with patients' clinical course and survival in different types of cancer. Digital immunohistochemistry (IHC) methods are used for automated quantitative assessment of TIL distribution in the tissue. This study presents a new method based on the principles of spatial hexagonal grid analytics by processing digital image analysis data to detect the tumor-stroma interface zone and to measure TIL density profiles within. Novel immune response indicators to quantify an absolute TIL density and it’s density gradient (immunogradient) in the direction from stroma to tumor in the interface zone were developed. The prognostic value of the indicators was tested in two independent colorectal cancer (CRC) patient cohorts, treated in health care institutions in Vilnius (Lithuania) and Nottingham (United Kingdom). The cytotoxic T cell (CD8) immunogradient (in both cohorts), the B cell (CD20) immunogradient, and the histological tumor growth pattern (in the Nottingham cohort) were independent prognostic features of overall patient survival and were more informative than the absolute TIL density in the tumor as well as conventional clinicopathological and molecular indicators. Combined prognostic scores were proposed: 1) the CD8–CD20 immunogradient score based on CD8 and CD20 IHC markers alone, and 2) the immuno-interface score which combines CD8 and CD20 IHC markers and the histological tumor growth pattern. These scores were strong prognostic factors for overall CRC patient survival, independently of the tumor microsatellite instability status