PREDICTION OF BREAST-CANCER MALIGNANCY USING AN ARTIFICIAL NEURAL-NETWORK

Background. An artificial neural network (ANN) was developed to predic t breast cancer from mammographic findings. This network was evaluated in a retrospective study. Methods. For a set of patients who were sch eduled for biopsy, radiologists interpreted the mammograms and provide d data on eight mammographic findings as part of the standard mammogra phic workup. These findings were encoded as features for an ANN. Resul ts of biopsies were taken as truth in the diagnosis of malignancy. The ANN was trained and evaluated using a jackknife sampling on a set of 260 patient records. Performance of the network was evaluated in terms of sensitivity and specificity over a range of decision thresholds an d was expressed as a receiver operating characteristic curve. Results. The ANN performed more accurately than the radiologists (P < 0.08) wi th a relative sensitivity of 1.0 and specificity of 0.59. Conclusions. An ANN can be trained to predict malignancy from mammographic finding s with a high degree of accuracy.