Carcinoma was found in 30% (119 of 400) of biopsy specimens obtained for mammographically suspicious but nonpalpable findings. The authors reviewed the mammograms of these 400 cases without knowledge of the biopsy results and placed each examination into one of four groups based on their suspicion for carcinoma: minimal (n = 82), slight (n = 91), moderate (n = 174), and high (n = 53). In 127 cases, mammograms showed either minimally suspicious calcifications (n = 33) or minimally (n = 49) or slightly (n = 45) suspicious masses. A 4.7% (six of 127) rate of carcinoma was found in these groups; five of the six cancers were in situ. Had follow-up mammography been done rather than biopsy for these 127 less suspicious lesions, it is probable that the delay in diagnosis would not have altered overall prognosis. In the remaining 273 patients, the positive predictive value of mammography for carcinoma would have risen from 30% (119 of 400) to 41% (113 of 273). The authors conclude that in the management of suspicious nonpalpable mammographic findings, the rate of carcinoma for lesions at biopsy can approximate 40%. This is almost double the rate of most published series.