Detection of increased level of chromosome breakage in peripheral blood of Iranian women with sporadic breast cancer using neural networks

Elevation of radiosensitivity may be a marker for breast cancer predisposition. This paper shows that the sensitivity to the induction of chromosomal damage by ionizing Gama Exposure is, on average higher in breast cancer patients than of normal healthy controls [1]. In order to investigation of gamma effect in each person´s lymphocytes and comparison between two groups, seventy two hours after blood sample culturing, the samples are exposed by gamma rays and they are harvested. Gamma exposure causes aberrations in chromosomes. Our database includes chromosome breakage in seven chromosome groups and age of patients. Principle Component Analysis (PCA) is used for feature selection stage. Then we used an Artificial Neural Networks (ANN) for classification of normal cases from abnormal cases. Satisfactory result obtained with an accuracy rate of 93.09% for Neural Networks (NN) classifier.