Astrocytoma Brain Tumor Automated Classification Procedure Based on Ki67 Expression

Abstract

It is known to be a very challenging task for pathologists to perform astrocytoma biopsy analysis especially with the high demand for an accurate and precise analysis. An inaccurate diagnosis might affect the treatment decisions for the patient. Therefore, this study aims to develop an automated procedure for classification astrocytoma brain tumor. The procedure consists of four stages, namely image pre-processing, image segmentation, feature extraction and selection, and classification of Ki67 expression. Initially, the image’s quality was enhanced by using global stretching and unsharp masking filter. Then, the adaptive color thresholding technique was used to segment the Ki67 expression. The process was followed by extracting the features related to the positive and negative Ki67 cells. ReliefF feature selection algorithm was used to select the best features in defining the positive and negative cells. In the classification stage, two classifiers, namely Support Vector Machine (SVM) and Multilayer Perceptron (MLP) were used to differentiate between the positive and negative cells. From 81 samples images, both classifiers yielded promising results, with an accuracy of 98.5% for SVM and 99.6% for MLP respectively.