In vivo, only a small fraction of cells are successfully able to form the complete metastatic cascade, which begins with leaving the primary tumor and invading into the original tissue, travelling through the blood stream, extravasating into a distal tissue, and colonizing a secondary site. It has been hypothesized that this is a Darwinian process in which cells which metastasize have the highest “fitness,” and cells which succeed are “selected” by morphological or other characteristics which are reflective of the combined genetic, transcriptomic, and proteomic state. During metastasis, cells undergo numerous mechanical stresses, including compressive stress in the primary tumor, migration through the dense origin and distal tissues, shear stress in the circulatory system, and migration through confined spaces as small as 1-2 µm in the basement membrane lining organs and blood vessels. It has become clear that this confined migration (CM) requires severe nuclear deformation, and can result in nuclear rupture, replication stress, and DNA-damage. Cells likely respond with specific but context-dependent signal transduction, and phenotypic, transcriptomic, and epigenetic alterations, which become more temporally persistent or extreme with increasing exposure to confinement. To date, the immediate and long-term consequences of confined migration are still becoming clear, but have been limited by the availability of existing methods to study individual cell fates during migration. Existing studies utilizing bulk methods have demonstrated differences in migration, persistence, and chemotactic ability in cells which have been “selected” through multiple rounds of confined migration, but lack the ability to study individual cell fates. Other methods have been able to characterize individual cells before, during, and after CM and note heterogeneity in phenotype, migration mode, but are not compatible with harvest of successful cells for other downstream characterization without expansion. In this thesis, I describe the development of a microfluidic device and protocol to bridge this gap and overcome these limitations by tracking individual cells before, during, and after migration through multiple constrictions posing varied degrees of confinement, quantifying the fraction which apoptose during transit, and then collecting the remaining fraction of the population to perform functional or sequencing-based studies, thus allowing the study of subpopulation dynamics. As proof of principle, we subjected MDA-MB-231 breast cancer cells to migration through devices with varying degrees of confinement. Simultaneous imaging revealed a portion of cells were “selected” by migration through confined conditions and unable to tolerate the associated stresses and apoptose during transit, while the remaining portion of cells only transiently altered transcriptional signature. We validated the most robust phenotypes predicted of the RNA-sequencing and observed increased autophagic flux and p62 accumulation, and observed that the remaining population exhibits indicators of decreased cell health and survival. Taken altogether, we validated a new device and protocol which provide a new degree of resolution over other available methods, and demonstrated divergent fates of cells within a population for confined migration with respect to cell health and survival.