Introduction

Knee osteoarthritis (KOA) is one of the leading causes of chronic pain, functional limitations and disability worldwide, affecting millions of individuals and placing a significant burden on healthcare systems [[1], [2], [3]]. KOA impacts patients’ quality of life and daily activities, often resulting in reduced mobility, difficulty with weight-bearing tasks, and reliance on healthcare resources [4]. Pain and functional impairment are hallmark features of KOA, often worsening due to progressive structural damage in the knee joint [5]. Structural changes, including cartilage degeneration, changes in subchondral bone and meniscal damage are central to the pathophysiology of KOA and are closely associated with worsening disease severity [6]. Advanced imaging techniques, such as MRI, are critical for identifying structural changes and quantifying disease progression. Scoring systems such as the Whole-Organ Magnetic Resonance Imaging Score (WORMS) provide detailed assessments of structural KOA severity by evaluating various joint compartments, including cartilage, bone marrow, ligaments, tendons and menisci [7]. These imaging-based assessments are critical for understanding KOA progression, as structural damage often correlates with increased pain and functional decline.

Muscle plays a crucial role in maintaining joint stability, mobility, and biomechanical balance, serving as a key component of the musculoskeletal system [8,9]. The integrity of muscle function is essential for preserving joint health, as it ensures optimal load distribution and stabilization during movement. Sarcopenia, characterized by a decrease in muscle mass, strength, and function, has been identified as a major risk factor for functional decline and disease progression in KOA patients [10]. Muscle loss can lead to altered joint biomechanics and increased cartilage stress, resulting in accelerated structural degeneration [11]. Recent evidence further emphasizes the important role of muscle impairment in KOA. A recent systematic review showed consistent neuromuscular deficits associated with pain and functional limitations [12], while reduced quadriceps strength was associated with worse WOMAC scores and impaired functional performance in patients with hip or knee osteoarthritis [13]. Previous studies have established the link between muscle weakness and joint instability [11]. For example, Mohajer et al. reported that longitudinal changes in quadriceps cross-sectional area and intramuscular adipose tissue, derived from single-slice MRI measurements, were associated with symptom progression and risk of knee replacement [14]. Most MRI studies investigating muscle in KOA have relied on single-slice cross-sectional measurements. However, single-slice measurements may not fully capture the complex three-dimensional distribution of muscle tissue and may be limited in representing global muscle morphology [15]. Therefore, despite existing evidence linking muscle characteristics to KOA outcomes, the association between longitudinal changes in volumetric muscle measures and structural joint damage over a long term and in a large cohort, as well as their impact on functional outcomes, remains less explored. In particular, the potential value of fully automated, deep learning-based assessments of thigh muscle volume (TMV) for providing additional or more robust information with structural disease progression has not yet been established. Volumetric assessment may capture muscle morphology more comprehensively than previously used single-slice or area-based approaches [14]. However, a direct comparison is beyond the scope of our study.

Functional limitations in KOA are frequently assessed using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), a validated, widely used tool designed to quantify pain, stiffness, and physical function in individuals with knee or hip OA [16]. Higher WOMAC scores reflect greater levels of pain, stiffness, and functional limitations, which are typically associated with more severe disease progression in KOA [17]. However, the association between changes in volumetric muscle measures and functional outcomes, as measured by WOMAC, remains unclear [18].

Given the critical role of muscle in KOA structural and functional outcomes, this study aimed to investigate the longitudinal association between changes in deep learning-based thigh muscle volume, structural progression of KOA, and functional impairment over 48 months. We hypothesized that muscle loss would be associated with greater progression of degenerative changes (WORMS score increase) and worse functional outcomes relative to stable muscle, and that muscle gain would be associated with slowed progression of degenerative changes (stable WORMS score).