Prostate Cancer Detection through MR Images based on Black Hole Optimization Algorithm

Abstract

Introduction: Prostate cancer is the most common type of malignant cancer among men and is known as one of the leading causes of cancer mortality in men. The complexity of diagnostic procedures such as mass biopsy has made new diagnostic strategies for prostate cancer, such as MRI imaging, a research priority in recent years. Methods: In this applied-descriptive study, a four steps method for diagnosing prostate cancer through MR image is presented. In the first step, the destructive effect of noise on the input images by using two-dimensional wavelet transform and histogram equalization is reduced. In the second step, the black hole optimization algorithm is used for segmentation of the input image based on the multilevel threshold technique. By doing this, the suspicious areas are identified in the image and in the third step, the features of each target area are extracted. In the fourth step, a combination of three learning algorithms, including: artificial neural network, decision and support vector machine is used to diagnose prostate cancer. Results: The effectiveness of the proposed method in diagnosing prostate cancer has been evaluated from various aspects and its performance has been compared with other learning models. Based on the results, the proposed method can diagnose prostate cancer through MRI images with an average accuracy of 99%. Discussion & Conclusion: The proposed method uses a combination of image processing, optimization and machine learning techniques to achieve this goal. Compared with other models, this proposed method was of the highest accuracy.