(804e) Phenotypic Signatures for Metastatic Cancer Cells

Cancer is the leading cause of death in U.S with ninety percent of cancer patients dying from the metastasis. Genomic analysis has been vigorous used to study cancer samples, however no consensus on molecular signatures has been found that are specific to metastasis. Accumulating evidences has suggested that cell mechanics could be a robust marker of metastasis. Therefore, the physical principles underlying metastasis could be translated into effective diagnostic and prognosis tools, as well as drug screening platforms. Here, we developed a comprehensive morphological analysis, visual aided morpho-phenotyping recognition (VAMPIRE), to classify irregular cellular and nuclear shapes using a set of common shape modes through eigenshape decomposition and clustering analysis. We collected a cohort of more than 20 different cancer cell lines including prostate, breast and pancreatic cancer cell lines which are either derived from primary tumors or from metastatic lesions. Using an automated high-throughput microscopy assay we collected more than data from more than 100,000 cells to examine the phenotypes of metastasis using VAMPIRE assay. Our results show that the lack of cell/nuclear morphological heterogeneity is a highly predictive feature of metastatic cancer cells and suggest that metastasis is associated with a selection process for biophysical features.