Most radiologists recommend follow-up mammography in 3 to 6 months for asymptomatic patients with mammographic lesions that do not appear suspicious enough to warrant immediate biopsy. In this retrospective cohort study, the medical records of 776 patients were audited 24 to 35 months after mammography to estimate the frequency of indeterminate lesions, the probability of malignancy, and the rate of patients' compliance with recommendations. Almost 2% of patients retested after an indeterminate mammogram were found to have breast cancer (95% confidence limits 0.2% and 5.9%). Only 26% of the patients retested during the audit period had the test within the suggested upper limit of 6 months. We conclude that timely follow-up after an indeterminate mammogram is advisable and that for the population we studied, compliance with this recommendation is low.