Improving the efficiency of CT image analysis using new texture radiomics features

This article discusses the development of feature extraction techniques from medical images to improve diagnosis and data analysis in oncology. Three new radiomic features for analyzing lung CT images are presented: adaptive texture contrast (ATC), directional texture uniformity (DTU), and co-occurrence of texture transitions (CTT). These features are specifically designed to improve the analysis of lung CT images, which can have a significant impact on the diagnostic accuracy and recognition of EGFR mutations. This article details the methods and algorithms used to create and test these features, and presents results demonstrating a 4% improvement in Accuracy and Precision for the task of detecting EGFR mutations compared to traditional methods. This study highlights the potential of integrating novel radiomic signatures into clinical practice for more accurate and efficient diagnosis of lung cancer.