AbstractTumor aggressiveness and subsequent metastasis still remain as major limitations to curative treatments for cancer patients. The epithelial to mesenchymal transition (EMT) is a process that occurs naturally during embryogenesis and has been linked to metastasis in many cancer types. EMT is characterized by phenotypic changes that allow for tissue extravasation and migration of cancer cells into the bloodstream. These changes may be linked to specific patterns in the gene expression signature of cells undergoing EMT. Recent theoretical efforts have predicted the existence of a stable, hybrid (E/M) phenotype which has also been observed experimentally in single cells. However, the effects of this hybrid phenotype on cancer patient survival have not been well characterized. The ability to quantify a patients EMT status via a simple test involving a small collection of prognostic genes would provide an important tool for clinical risk stratification and treatment in the context of many cancer types. Here, we apply iterative statistical methods to generate an EMT score based on gene expression samples. We use this score to characterize the degree of the hybrid phenotype signature present in test samples. Predictions from our model are verified against cell lines with known EMT status. Lastly, we apply our model to clinical samples in order to assess survival differences in various EMT groups. We demonstrate that in many cases, EMT status successfully classifies patients into groups with statistically significant differences in survival, which is of immediate clinical relevance.Citation Format: Jason T. George, Mohit K. Jolly, Herbert Levine. Quantitative EMT expression score for predicting survival outcome [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 563. doi:10.1158/1538-7445.AM2017-563
