CARCINOMA OF THE BREAST - MEASUREMENT AND THE MANAGEMENT OF TREATMENT .2. THE REGRESSION OF TUMORS

Regression slopes fitted in 238 new cases of primary breast cancer submitted to a variety of managements are analysed. Contrary to the original hypothesis that slopes would primarily reflect the composition of the tumours, the steepness of resorption slopes is found also to be treatment dependent. Investigation of other factors influencing slopes is confined to radiotherapy where the expectation that the prescribed dose will be delivered is high. A large cumulative radiation effect before the start of shrinkage is associated with steeper slopes. A small correlation between slope and both menopausal status and starting diameter has been found, with premenopausal patients and small tumours producing the steepest slopes. The full range of slopes has been seen in postmenopausal patients. The data has been fitted to a linear model to evaluate the relative importance of these associations. Illustrative cases support the hypothesis that resorption slopes are a product of tumour architecture and histology but we conclude that the maximum possible steepness of slope is not always achieved in the clinic, even when radiation is used. A combination of histological and regression data is therefore required to characterize a tumour for predictive purposes.