Identification of histological features to predict MUC2 expression in colon cancer tissues

Abstract

Colorectal cancer (CRC) is the third-most common form of cancer among Americans. Like normal colon tissue, CRC cells are sustained by a subpopulation of \"stem cells\" that possess the ability to self-renew and differentiate into more specialized cancer cell types. In normal colon tissue, the enterocytes, goblet cells and other epithelial cells in the mucosa region have distinct morphologies that distinguish them from the other cells in the lamina propria, muscularis mucosa, and submucosa. However, in a tumor, the morphology of the cancer cells varies dramatically. Cancer cells that express genes specific to goblet cells significantly differ in shape and size compared to their normal counterparts. Even though a large number of hematoxylin and eosin (H&E)-stained sections and the corresponding RNA sequencing (RNASeq) data from CRC are available from The Cancer Genome Atlas (TCGA), prediction of gene expression patterns from tissue histological features has not been attempted yet. In this manuscript, we identified histological features that are strongly associated with MUC2 expression patterns in a tumor. Specifically, we show that large nuclear area is associated with MUC2-high tumors (p < 0.001). This discovery provides insight into cancer biology and tumor histology and demonstrates that it may be possible to predict certain gene expressions from histological features.