Postmenopausal patients with oestrogen receptor-positive locally advanced T4b, N0-1, M0 and large operable breast cancers T2>3 cm, T3, T4, N0-1 and M0 have been treated with 2.5 mg letrozole (12 patients), 10 mg letrozole (12 patients), 1 or 10 mg anastrozole (24 patients) and 20 mg tamoxifen (65 patients). There was no apparent difference in response rate between 2.5 and 10 mg letrozole. Only 17 patients with anastrozole have so far completed the 3-month treatment period. Median clinical, mammographic and ultrasound reductions in tumour volumes for patients treated with letrozole were 81% (95% confidence interval (CI) 66-88), 77% (95% CI 64-82) and 81% (95% CI 69-86) respectively and for anastrozole, values were 87% (95% CI 59-97), 73% (95% CI 58-82) and 64% (95% CI 52-76) respectively. This compares with a median reduction in tumour volume for tamoxifen-treated patients as assessed by ultrasound of 48% (95% CI 27-48). There were seven complete clinical responses (CR), sixteen patients who achieved 50% or greater reduction in tumour volume (PR) and one no change (NC) for letrozole and four CRs, twelve PRs and one progressive disease for anastrozole. Best radiological responses were one CR, twenty PRs and three NCs for letrozole and one CR, fifteen PRs and one NC for anastrozole. This study has shown that the new aromatase inhibitors, letrozole and anastrozole, are highly effective agents in the neoadjuvant setting and they should now be compared with tamoxifen as first-line treatment in a randomised study.