STATISTICAL STUDIES OF PROGNOSIS IN ADVANCED BREAST CANCER

From statistical studies of a group of patients with advanced carcinoma of the breast who underwent ablative surgery (either hypophysectomy or bilateral adrenalectomy with oophorectomy) certain variables describing their clinical, surgical and natural histories are shown to have prognostic values. Urinary steroid assays stood out as the most useful guide to predicting the future course of the disease. The discriminant function, determined from previous studies for the prediction of a successful or intermediate response to ablation, would appear to be improved if androsterone and the 17-hydroxycorticosteroids were used rather than the latter with aetiocholanolone. There is also some evidence to suggest an interaction between the previous history of mastectomy and ablation, but this is not conclusive because the type of mastectomy, if the operation was performed at all, was not allocated randomly, a decision being made on the basis of the condition of the patient. The reliability of a prognosis based on the steroid levels is better for patients submitted for hypophysectomy than it is for patients submitted for adrenalectomy. The time interval between first seeing the patient and ablation, reflecting the speed of progression of the disease, is also useful for prognosis, but is again more reliable for patients submitted for hypophysectomy than it is for adrenalectomy patients. The age of the patient at ablation and the site of the predominant lesion do not appear to have prognostic value. The only reliable test of the value of particular variables in prognosis is the use of these variables to predict results for patients who were not used in the retrospective analysis. Evidence gained [4] from the use of the original discriminant function in predicting the course of the disease in patients whose response to ablation was not used in the original calculation of the discriminant gives cause for optimism. © 1969.