Applying Bayesian Network Approach to Determine the Association Between Morphological Features Extracted from Prostate Cancer Images

Abstract

Cancer is a major public health problem across the globe due to which millions of deaths occur every year. In the United States, prostate cancer is the second leading cause of cancer-related deaths in men. The major causes of prostate cancer include increasing age, family history, diet, sexual behavior, and geographic location. Early detection of prostate cancer can effectively reduce the mortality rate. In the past, researchers have adopted various multimodal feature extracting strategies to extract diverse and comprehensive quantitative imaging features and employed machine learning methods to detect prostate cancer. However, existing techniques lack detailed analysis of the magnitude of relationship among different individual discriminatory features, which is very important to understand the dynamics of the disease. In this study, we extracted diverse morphological features to summarize the imaging profile of patients of prostate cancer imaging database and employed Bayesian network analysis approach to quantify the association between different features and the strength of the association. The features and the association between the features were, respectively, modeled as the nodes and the edges of the network. The strength of association between the nodes was computed using Pearson's correlation, mutual Information and Kullback-Liebler methods. The strongest associations were found between multiple features: (Area → Equidiameter), (Area → Circulatory 2), (Circulatory 1→ (Elongatedness), (Circulatory 1→ Entropy), (Circulatory 1→ Max. Radius), and (Min. Radius → Eccentricity). Moreover, interaction impact among nodes and node force was also computed. This analysis will help in finding the features that are more dominant to establish the relationship and can further increase the detection performance.

A Bayesian Dynamic Inference Approach Based on Extracted Gray Level Co-Occurrence (GLCM) Features for the Dynamical Analysis of Congestive Heart Failure

“…The researchers [26] computed texture features to predict breast cancer. The researchers [27][28][29][30][31][32] computed various features based on texture, morphology, scale invariant feature transform (SIFT), and elliptic Fourier descriptors (EFDs) to predict brain tumors, lung cancer, breast cancer, and prostate cancer. In this study, we extracted Gray-level co-occurrence matrix (GLCM) features and then ranked the features based on empirical receiver operating characteristic (EROC) and a random classifier slope, as utilized in [33][34][35] to rank the features' importance.…”

Analyzing the Dynamics of Lung Cancer Imaging Data Using Refined Fuzzy Entropy Methods by Extracting Different Features

“…In the past, researchers applied different complexity measures to quantify the dynamics of highly complex and nonlinear dynamical measures by applying Multiscale sample entropy (MSE), MPE, symbolic entropy and refined Fuzzy entropy measures, time-frequency representation measures for detecting epileptic seizures, arrhythmia, heart rate failure, gait dynamics etc. [39], [40], [45]- [48], [56], [87]. The entropy-based complexity measures have been used in diverse fields to quantify the dynamics of highly nonlinear physiological and neurophysiological signals and systems.…”

“…Researchers computed morphological features for automated prostate cancer localization [50], quantifying phenotypes for image analysis [51]- [53], detecting lung cancer [54], detection of seed of wild castor oil plants [55], colon cancer detection [18]. Recently Hussain et al computed morphological features from prostate cancer images and computed the associations among these features to determine the strength between these features [56]. Morphological feature extraction module (FEM) takes input inside the shape of binary cluster and finds associated components in the clusters [57].…”

Lung Cancer Prediction Using Robust Machine Learning and Image Enhancement Methods on Extracted Gray-Level Co-Occurrence Matrix Features

“…Researchers are paying much attention to compute the most relevant features [17]. In the field of medical imaging problems, the medical data are collected without sacrificing the result quality [25][26][27][28][29][30][31]. To capture the most relevant properties, we computed GLCM features.…”