We examined the impact of metropolitan racial residential segregation on stage at diagnosis and all-cause and breast cancer-specific survival between and within black and white women diagnosed with breast cancer in California between 1996 and 2004. We merged data from the California Cancer Registry with Census indices of five dimensions of racial residential segregation, quantifying segregation among Blacks relative to Whites; block group ("neighborhood") measures of the percentage of Blacks and a composite measure of socioeconomic status. We also examined simultaneous segregation on at least two measures ("hypersegregation"). Using logistic regression we examined effects of these measures on stage at diagnosis and Cox proportional hazards regression for survival. For all-cause and breast-cancer specific mortality, living in neighborhoods with more Blacks was associated with lower mortality among black women, but higher mortality among Whites. However, neighborhood racial composition and metropolitan segregation did not explain differences in stage or survival between Black and White women. Future research should identify mechanisms by which these measures impact breast cancer diagnosis and outcomes among Black women.