Use of neural network analysis to diagnose breast cancer patients

In the diagnosis of breast cancer, several different diagnostic tests can be conducted on the same patient simultaneously so as to improve diagnostic results. There are three tumor markers currently available for breast cancer diagnosis, namely CEA, CA15.3, and MCA. The purpose of this study was to investigate the usefulness of neural networks to distinguish breast cancer patients from normal people based on the pattern of the three tumor marker measurements. The neural network was built and trained with a training data set, and then tested with a separate data set. In order to evaluate the performance of the neural network which differentiates breast carcinoma from normal conditions, an advanced statistical method, relative operating characteristic (ROC) analysis, was utilized. In addition, two multivariate analysis studies using discriminant functions and logistic regression were also performed. The results show that the neural network compared favorably with the two conventional statistical methods