Structural and Statistical Texture Approaches in the Radiological Assessment of Knee Osteoarthritis: A Multimodal Review

Abstract Knee osteoarthritis (KOA) represents a significant degenerative joint disorder characterized by the progressive deterioration of articular cartilage. The therapeutic approach to KOA is contingent upon disease progression. Contemporary diagnostic methodologies encompass both traditional clinical assessments and automated diagnostic systems. Digital imaging-based diagnostic platforms facilitate early detection of KOA manifestations. Recent technological advancements have yielded sophisticated analytical methods, including Histogram of Oriented Gradients (HOG), fractal analysis, and Local Binary Patterns (LBP), which have emerged as valuable diagnostic tools over the past decade. This comprehensive review examines the diverse array of texture analysis methodologies employed in KOA detection across multiple imaging modalities, including Computed Tomography (CT), X-ray radiography, and Magnetic Resonance Imaging (MRI). The primary objectives of this investigation are multifaceted; To introduce novel methodological approaches in the field; To systematically categorize significant contributions in image-based texture analysis; To delineate principal challenges and significant findings; To evaluate existing research protocols and diagnostic questionnaires, and to propose strategic recommendations for enhancing diagnostic precision in KOA assessment. The analytical framework encompasses statistical, structural, transform-based, and model-based methodologies. This review synthesizes findings from more than 60 peer-reviewed publications, categorized according to their methodological approach, imaging modality, and classification algorithms. Each methodological section presents a detailed analysis of the respective approaches, including their theoretical foundations, practical applications, and inherent limitations.