Patient-detected breast cancer is not well described, and its association with survival is not known.Initial breast cancer detection methods were extracted from patient chart notes in a registry cohort of women aged 40-93 with primary invasive breast cancer seen at the community-based regional cancer center between 1990 and 1999 (N = 2228). Patients were followed for vital status and recurrence (current as of 2003/2004) with a mean follow-up of 7.6 years. Physician-detected breast cancers were excluded from the analysis (n = 231).Forty-six percent of cases were patient detected (n = 1019), and 54% were mammography detected (n = 1209). Those with patient-detected disease received more aggressive treatment (combined radiation and chemotherapy; 49% vs. 23%; P < 0.001) and were twice as likely to have a modified radical mastectomy (41% vs. 20%; P < 0.001). Patient-detected cases were more likely to have a disease recurrence (5-year recurrence rate: 13% vs. 6%; log-rank test, 30.51; P < 0.001) and were more likely to die of disease (5-year disease-specific mortality rate: 8% vs. 3%; log-rank test, 34.7; P < 0.001). In a Cox proportional hazards model, detection method was not associated with risk of breast cancer-related death.Patient-detected breast cancer appears to be a more virulent form of breast cancer than mammography-detected breast cancer with higher recurrence and mortality risk despite more aggressive treatment.