The existence of aromatase activity in human breast carcinomas has been established for about 20 years but the clinical and biological importance of this remains unclear. A number of studies in clinical material suggest that aromatase activity may be a prerequisite of response to aromatase inhibitors and that aromatase activity may be enhanced in those tumours relapsing during treatment with one such inhibitor, aminoglutethimide. These results would carry more significance, however, if it was demonstrable that the growth of breast carcinomas is affected by the conversion of androgens to oestrogens by intratumoural aromatase. We have tried to address this by establishing model systems with aromatase-transfected MCF7 breast cancer cells. We have demonstrated that these cells can be stimulated mitogenically with androgen and that this proliferation is suppressible with aromatase inhibitors. Similarly the growth of aromatase transfected cells but not wild type cells as xenografts is supported by androstenedione and inhibitable by both the steroidal aromatase inhibitor, 4-hydroxyandrostenedione and non-steroidal inhibitor, CGS 20267. Work with the former of these, which is a suicide inhibitor allowed us to demonstrate that growth can proceed with aromatase activity approximating to the highest level seen in breast carcinomas indicating that at least at this extreme level the intratumoural conversion of androgens to oestrogens may indeed be able to support tumour growth. Further work with this mode system should allow us to define the minimal amount of intratumoural activity which can support tumour growth.