In this dissertation, we developed cell stiffness as a biomarker in ovarian cancer for the purpose of grading metastatic potential. By measuring single cell stiffness with atomic force microscopy and quantifying in vitro invasiveness of healthy and cancerous ovarian cells, we demonstrated that cancerous ovarian cells have reduced stiffness compared to the healthy ones and invasive ovarian cancer cells are more deformable than noninvasive ovarian cancer cells. The difference in cell stiffness between two genetically similar cell lines was attributed to actin-mediated cytoskeletal remodeling as revealed by comparative gene expression profile analysis, and was further confirmed by fluorescent visualization of actin cytoskeletal structures. The actin cytoskeletons were innovatively quantified and correlates with cell stiffness distributions, further implicating actin-mediated cytoskeletal remodeling in stiffness alteration from the perspective of structure-property relationship. The correlation between stiffness and metastatic potential was also demonstrated in pancreatic cancer cell line AsPC-1, which shows reduced invasivess and increased stiffness upon treatment with N-acetyl-L-cysteine (NAC), a well known antioxidant, reactive oxygen species (ROS), scavenger and glutathione precursor. The correlation between cell stiffness and metastatic potential as demonstrated in ovarian and pancreatic cancer cells indicated that mechanical stiffness may be a useful biomarker to evaluate the relative metastatic potential of ovarian and perhaps other types of cancer cells, and might be useful clinically with the development of rapid biomechanical assaying techniques. We have also investigated the stiffness evolution through progression of the cell cycle for the healthy ovarian phenotype and the invasive cancer ovarian phenotype, and found that the healthy phenotype at G1 phase are significantly stiffer than other single cells except the invasive phenotype at late mitosis; other groups are not significantly different from each other. We have also investigated intracellular heterogeneity and mechanical nonlinearity in single cells. To this end, we developed a methodology to analyze the deformation-dependent mechanical nonlinearity using a pointwise Hertzian method, and tested the method on ultrathin polydimethylsiloxane (PDMS) films which underwent extremely large strains (greater than 50%). Mechanical stiffening due to large strain and geometrical confinement were observed. The onset of nonlinearity or mechanical stiffening occurs at 45% of the film thickness, the geometry induced stiffening causes an increase in stiffness which shows a strong power law dependence on film thickness. By applying the pointwise Hertzian method on stiffness measurements with AFM that were collected on living cells, we also investigated the nonlinear and heterogeneous mechanics of single cells, since attachment of cells to stiff substrate during indentation may impact their mechanical responses. Even under natural biological conditions, cells confined in narrow spaces may experience heightened mechanical stiffness. Through indentation-dependent force mapping, analysis of the local cell stiffness demonstrated spatial variation. The results indicated that the mechanical properties of single cells are highly nonlinear and are dependent upon the subcellular features under the applied force as well as the dimensions of the cellular material. We identified single cell stiffness as a potential biomarker of the metastatic potential in ovarian cancer, and quantified the effect of geometrical confinement on cell mechanics. The results presented in this dissertation not only made contributions to the development of accurate, non-invasive clinical methods to estimate metastatic potential of ovarian and perhaps other types of cancer, but also shed light on the intracellular mechanical information by developing new techniques to quantify the effect of geometry on cell mechanics.