Ki-67 expression in ductal cells obtained by random periareolar fine needle aspiration (RPFNA) is currently being used as a response biomarker in phase II breast cancer chemoprevention trials; however, Ki-67 in RPFNA has not been well studied as a risk predictor for cancer, which would support its use as a response indicator. We examined the expression of Ki-67 in RPFNA specimens with hyperplasia +/- atypia obtained from 147 women at high risk for development of breast cancer. Median Ki-67 was 1.4% (range 0-24%). Ki-67 was higher in specimens from women < 50 versus those > or = 50 (median 2% versus 0.6%; P = 0.006) and from premenopausal women versus postmenopausal women (P = 0.037); however, hormone replacement therapy (predominately low-dose estrogen without progestins) had no effect. By univariate analysis, Ki-67 was positively correlated with ductal cell number (P = 0.001) and hyperplasia with atypia (P = 0.007). By multivariable analysis, the proportion of ductal cells expressing Ki-67 was again predicted by cell number, which, in turn, was predicted by cytologic atypia. The association of Ki-67 expression with cytologic atypia, a known risk factor for development of breast cancer, provides preliminary justification for its use as a response biomarker in phase II chemoprevention trials.