Computer-aided assessment of receptor status in human breast cancer

Computer-aided analysis of simulated cytosolic estradiol-17 beta receptor binding data using Scatchard and nonlinear Mass Action models showed the latter to be superior. Simulation studies of binding data, incorporating the magnitude and uncertainties associated with inefficiencies of separating free from bound hormone, nonspecific and specific binding, allowed the prediction of uncertainties associated with estimates of receptor content. Realistic values for the minimum detectable receptor concentration that can be distinguished from zero with a given probability were also obtained and may be used as a cutoff value for selection of patients for hormone therapy. This improved reliability in quantification should more clearly define the prognostic value of these receptor assays with respect to adjuvant therapy for primary cancer following surgery or anti-estrogen treatment for recurrent disease, thereby improving the management of breast cancer. Computer simulations and regression analyses were performed on a PDP11/34 using programs written in FORTRAN IV.