Using multidimensional topological data analysis to identify traits of hip osteoarthritis

Area of Science:

• Orthopedics and Sports Medicine
• Radiology
• Data Science

Background:

• Osteoarthritis (OA) is a complex condition with numerous diagnostic and progression factors.
• Topological Data Analysis (TDA) offers a powerful big data approach to integrate diverse variables into a multidimensional framework.
• TDA enables simultaneous analysis of imaging and gait data for OA research.

Purpose of the Study:

• To identify biochemical and biomechanical biomarkers for classifying hip OA progression phenotypes.
• To differentiate between subjects with and without radiographic signs of hip OA based on identified biomarkers.

Main Methods:

• A longitudinal study comparing progressive and non-progressive subjects (n=102) with and without hip OA.
• Utilized 3T MRI (SPGR 3D MAPSS T1ρ/T2, FSE) for cartilage composition and bone shape assessment.
• Applied TDA, Kolmogorov-Smirnov testing, and Benjamini-Hochberg for multidimensional data analysis.

Main Results:

• Later-stage OA subjects showed higher SHOMRI, KL scores, and older age (P<0.0001).
• Healthier subjects had intact cartilage and less pain.
• Knee biomechanics emerged as an early marker (P<0.0001) preceding morphological changes.
• Anterior labral tears were significant in femoroacetabular impingement (FAI) subjects with OA symptoms (P=0.0017).

Conclusions:

• TDA-driven analysis proposes novel OA phenotypes, partially aligning with radiographic classifications.
• Identified knee biomechanics as a potential early indicator for intervention.
• Highlights TDA's utility in uncovering complex disease patterns and guiding early treatment strategies.