Novel multidisciplinary and multilevel approaches are required to link biologic and social mechanisms with cancer. We propose a new biosocial concept, "neighborhood cancerization," which postulates that residents of the same geographically-defined regions can be exposed to common unfavorable circumstances. These common neighborhood-level exposures can in turn have biological consequences that may result in an increased risk of cancer. Just as common "molecular signatures" can differentiate tumor types, "neighborhood signatures" can identify neighborhoods at increased risk for cancers of similar etiologic origins. Under a shared chronic stress hypothesis, we test the neighborhood cancerization theory by first determining the effect of neighborhood circumstances on telomere length (TL), a cellular marker of oxidative stress often implicated in cancer development at the population level. After addressing common methodologic concerns often cited in TL studies, we tested neighborhood and TL associations in a multi-racial, multi-center setting and in the context of individual-level stressors using quantile regression. We then developed and conducted a neighborhood-wide association study (NWAS) using all available U.S Census variables and the Pennsylvania State Cancer Registry in order to empirically identify common neighborhood factors related to prostate cancer. Our novel NWAS approach demonstrates how agnostic, high-dimensional data analyses can be used to identify neighborhoods and people at risk for high grade/high stage, aggressive prostate cancer. Our work has implications for health disparities research, and provides evidence to support the neighborhood cancerization hypothesis.