Characterization of Non-Small Cell Lung Carcinoma Gross Target Volume with 18F-FDG PET scan using Texture Analysis

Abstract

This study concern to characterize the lung area to cardiac, lung, tumor and submucosal using Gray Level Co-occurrence Matrix (GLCM) and extract classification features from PET/CT with fluorine-18 fluorodeoxyglucose images. Using the GLCM techniques to find the gray level variation in PET/CT images it complements the features extracted from PET/CT images with variation of gray level in pixels and estimate the distribution of the sub-patterns using Interactive Data Language IDL software. The results show’s that the Gray Level Co-occurrence Matrix and features extracted give a classification accuracy of cardiac 91.6%, lung 100%, tumor 99.6%, while the sub-mucosal showed accuracy 91.2%. The overall classification accuracy of lung area 96.0%. These relationships are stored in a Texture Dictionary that can be later used to automatically annotate new PET/CT images with the appropriate lung area names.