Abstract
It is proposed that the necessary balance between received wisdom and innovation in the hormonal manipulation of breast cancer has been lost through the dominance of the randomized, controlled trial. This method of investigation has displaced, rather than complemented, other means of progress. It has encouraged the indiscriminate use of hormones, especially tamoxifen, without reference to selective receptor tests, both old and new. The history of hormone manipulations and the knowledge accumulated thereby are reviewed. Data collected in the Breast Study Centre are presented to support the contention that treatment progress can be evaluated without recourse to randomization. The results confirm received wisdom that about one-third of breast cancers are hormone dependent. At a time when oestrogen receptor testing was not readily available, a trial of tamoxifen in an unselected series of breast cancer patients resulted in measurable shrinkage in less than half the cases and in 17% produced measurable growth. The latter were poorly differentiated grade III tumours. A satisfactory response to hormones was found to be associated with highly differentiated tumours. Recent advances in immunohistochemistry enable detailed profiles of histological grade, hormone receptors, proliferation indices and other markers of tumour behaviour in formalin-fixed sections only a few microns apart. Treatment may be selected for the particular tumour and the response monitored. A database representative of the wide range of tumour types, growth rates and hormone dependence may then be accumulated. This approach leads to refinements in therapy in contrast to trial results, representing an average, that are not applicable to the next individual patient.