Classification of malignant tumors using multiple sonographic features

Abstract

Breast cancer is the most common form of cancer among women and the second one in the world trailing behind lung cancer. In this paper, we present a diagnostic algorithm that uses multiple features of ultra-sonography for identifying breast nodule malignancy to provide better chance of a proper treatment. An artificial neural network has been put into operation in the form of multilayer perceptron to generate the predictive model. MATLAB has been used for the simulation of this algorithm and the results obtained are presented in this paper.