Preliminary evidence has associated testosterone and prostate-specific antigen (PSA) with breast cancer. Our objective was to determine whether a) testosterone levels in nipple aspirate fluid (NAF), serum, or breast tissue are associated with breast cancer; b) testosterone levels in serum are associated with levels in NAF; c) PSA in NAF, serum, or breast tissue is associated with breast cancer; and d) serum PSA is associated with NAF PSA levels. We obtained 342 NAF specimens from 171 women by means of a modified breast pump. Additionally, we collected 201 blood samples from 99 women and 51 tissue samples from 41 subjects who underwent surgical resection for suspected disease. Women currently using birth control pills or hormone replacement therapy were excluded from the study. Controlling for age and menopausal status, serum testosterone was significantly increased in women with breast cancer (p = 0.002). NAF and serum testosterone levels were not associated. Neither NAF nor tissue testosterone was associated with breast cancer. Controlling for menopausal status and age, NAF PSA was significantly decreased in women with breast cancer (p < 0.001). We did not find serum PSA to be associated with breast cancer, although we found an indication that, in postmenopausal women, its levels were lower in women with cancer. Serum PSA was associated with NAF PSA in postmenopausal women (p < 0.001). PSA levels in cancerous tissue were significantly lower than in benign breast specimens from subjects without cancer (p = 0.011), whereas levels of PSA in histologically benign specimens from subjects with cancer were intermediate. Our results suggest that serum testosterone is increased and NAF PSA is decreased in women with breast cancer, with PSA expression being higher in normal than in cancerous breast tissues. NAF and serum PSA levels in postmenopausal women are correlated, suggesting that as laboratory assessment of PSA becomes more sensitive, serum PSA may become useful in identifying women with breast cancer.