OBJECTIVE:Knee-adjacent subcutaneous fat (kaSCF) has emerged as a potential biomarker and risk factor for osteoarthritis (OA) progression. This study aims to develop an artificial intelligence-based tool for the automatic segmentation of kaSCF thickness and evaluate the cross-sectional associations between kaSCF, cartilage thickness, magnetic resonance imaging-based cartilage T2 relaxation time, knee pain, and muscle strength independent of body mass index (BMI). DESIGN:Baseline 3.0T MR images of the right knee from the entire Osteoarthritis Initiative cohort (n=4796) were used to quantify average values of kaSCF, cartilage thickness, and T2 using deep learning algorithms. Regression models (adjusted for age, gender, BMI, and race) were used to evaluate the associations between standardized kaSCF and outcomes of cartilage thickness, T2, pain, and knee extension strength. RESULTS:Model prediction CVs for kaSCF thickness ranged from 3.57% to 9.87% across femoral and tibial regions. Greater average kaSCF was associated with thinner cartilage in men (std. β= -0.029, 95% CI: -0.050 to -0.007, p=0.010) and higher T2 in women (std. β=0.169, 95% CI: 0.072 to 0.265, p=0.001). Greater kaSCF was also associated with lower knee extension force (std. β= -15.36, 95% CI: -20.39 to -10.33, p<0.001) and higher odds of frequent knee pain (std. odds ratio=1.156, 95% CI: 1.046 to 1.278, p=0.005) across all participants. CONCLUSIONS:Greater kaSCF was associated with thinner cartilage in men, higher T2 in women, reduced knee strength, and greater knee pain, independent of BMI. These findings suggest a potential role of kaSCF as a predictor for knee osteoarthrits-related structural, functional, and clinical outcomes independent of the effects of BMI.

OBJECTIVE:Using WBCT arthrography (WBCTa) as the referent standard, this study evaluated the diagnostic performance of MRI for visualizing cartilage lesions and determined the rates at which WBCTa detects cartilage lesions that were missed or scored as more severe compared to MRI. MATERIALS AND METHODS:This cross-sectional, level II evidence, prospective observational study included 65 participants undergoing MRI for evaluation of suspected cartilage damage. WBCTa images were acquired following MRI. Knee cartilage morphology was semi-quantitatively assessed using the MRI Osteoarthritis Knee Score (MOAKS) on both imaging modalities. Sensitivity, specificity, accuracy, PPV, and NPV were calculated. The frequencies for cartilage severity detected on WBCT versus MRI in patellar, femoral, and tibial subregions were summarized. RESULTS:MRI visualization of cartilage lesions had a sensitivity ranging from 0-79% for lesions detected on WBCT, specificity of 84-100%, and accuracy of 71-92%. Excluding the posterolateral tibia, cartilage lesion severity was greater on WBCTa than on MRI in all other subregions (3.1-33.8% of knees) where lesions were detected by both modalities. CONCLUSION:MRI did not visualize a significant proportion of cartilage lesions that were visualized with WBCTa. In all subregions, excluding the posterolateral tibia, cartilage lesion severity was found to be greater on WBCTa than on MRI. Reasons for lower performance of MRI compared to WBCTa are likely multifactorial, potentially relating to scanning configurations, inferior conspicuity of the articular surfaces due to a lack of intraarticular contrast or fluid, and technical reasons such as inferior resolution and type of MRI sequence used.

Corticosteroid injections were associated with higher progression of knee osteoarthritis compared with hyaluronic acid injections and controls, as determined by Whole-Organ MRI Score.

We investigated whether cartilage composition and thickness and its change over time were associated with future intermittent and constant knee pain. Osteoarthritis Initiative participants with 3T MRI scans from baseline to 36-month visits were selected. Outcomes were Intermittent and Constant Osteoarthritis Pain (ICOAP) scores in the right knee at the 48-month visit (0 to 100 = highest pain). We measured T2 values and cartilage thickness in 5 regions in the right knee from baseline to 36-months using deep-learning-based segmentation. Associations between baseline and change in cartilage biomarkers with pain scores were tested using adjusted logistic and linear regression models. Of 3780 included participants, 1042(28

Objective:Weight loss can modify the progression of osteoarthritis (OA), and this may, in part, be achieved by decreased synovitis, a known accelerator of cartilage degeneration. The purpose of this study was to investigate whether change in synovitis mediates the beneficial effect of weight loss on longitudinal cartilage outcomes. Method:We analyzed right knees with baseline Kellgren & Lawrence grades 1-3 of 1153 obese and overweight participants of the Osteoarthritis Initiative (OAI) cohort with Whole Organ MRI Scores (WORMS) and semi-quantitative assessment of effusion synovitis and synovial proliferation scores form 3T MRIs at baseline and 48 months. There were 295 participants with weight-loss >5 ​% and 858 stable weight controls. Ordered logistic regression was used to assess the association of weight-loss status with concurrent changes in synovitis as well as cartilage WORMS scores; models were adjusted for age, gender, race, presence of radiographic OA, and abdominal circumference at baseline. A mediation analysis was used to determine whether change in overall cartilage degeneration was mediated by change in synovitis scores. Results:Individuals who lost weight had significantly lower odds for a higher grade on the scale assessing change in overall synovitis (OR 0.72; 95%CI 0.54, 0.95; p ​= ​0.018). Mediation analysis showed that slowing synovitis during weight loss had a small mediating effect on longitudinal cartilage outcomes. Conclusion:Decreased cartilage degeneration during weight loss was partially mediated by concurrent deceleration in synovitis, showing that decreasing systemic inflammation during weight-loss may not be mirrored in imaging markers of joint inflammation.

The degree of thigh intramuscular fat in individuals without OA is fundamental for distinguishing natural variations in intramuscular fat from pathological changes. The goals of this study were to estimate the degree of thigh intramuscular fat in individuals without radiographic OA or frequent pain and assess the associations of age, sex, and BMI with the degree of intramuscular fat. Individuals without knee or hip radiographic OA, without total knee/hip arthroplasty, and without frequent knee/hip pain were selected from the OAI database (n = 710). Goutallier Grades (GGs) of the quadriceps and hamstring muscles were assessed based on 3 T MR images on a scale from 0 (normal muscle) to 4 (more fat than muscle). The associations between demographic variables and GG outcomes were evaluated using mixed effects models. The most prevalent GGs among the muscles were Grades 1 and 2; Grade 4 was infrequent (< 1%). Greater BMI (p < 0.001) and age (p < 0.001) were each associated with greater GG. Women had greater GG than men (greatest difference in the vastus medialis: coeff. = 0.214, p < 0.001). At lower BMI, women had greater intramuscular fat than men; at higher BMI, men had greater intramuscular fat than women (p = 0.029 for BMI-sex interaction). While individuals without radiographic OA or frequent pain generally had low thigh intramuscular fat, higher BMI and age were associated with greater intramuscular fat, and GGs were greater in women than men. The relationship between BMI and intramuscular fat was sex-dependent. Thus, demographic variables must be considered when evaluating intramuscular fat.

In osteoarthritis (OA) research it is increasingly recognized that stratification according to disease phenotypes is essential for optimizing treatment regimens. Gender-specific differences in clinical OA manifestations have been identified, and this analysis aimed to assess whether these differences extend to imaging phenotypes. From the Osteoarthritis Initiative database 2523 participants (1409 women and 1114 men) with completed 3T MRI and whole-organ magnetic resonance imaging scores (WORMS) of the right knee at baseline were included. Imaging phenotypes were assigned based on modified Rapid OsteoArthritis MRI Eligibility Score: the inflammatory, meniscus-cartilage, and bone phenotype. Logistic regression was performed to investigate the effect of gender on phenotype, independent of BMI, race, Kellgren & Lawrence (KL) grade and level of physical activity. Female gender was independently associated with lower odds for the meniscus-cartilage (OR 0.61, 95%CI 0.47-0.80, p < 0.001) but not for the inflammatory (OR 1.04, 95%CI 0.89-1.24, p = 0.697) or the subchondral bone phenotype (OR 1.13, 95%CI 0.95-1.36, p = 0.166). This difference highlights an opportunity for future refinements to better accommodate gender/sex differences in disease trajectories while investigating different treatment regimes in knee OA.

OBJECTIVE:To explore the relationship between ultra-processed food (UPF) intake and knee osteoarthritis (KOA)-related imaging and clinical outcomes in men and women. DESIGN:Osteoarthritis Initiative participants with sufficient dietary and sociodemographic data (n = 4403) were included in this cross-sectional study. UPF was assessed by food frequency questionnaire-based NOVA Classification, categorizing diet according to processing level, with NOVA-4 indicating UPF. The exposure variable was standardized UPF proportion/day (%)-the proportion of NOVA-4 servings in the daily diet. The outcomes were Western Ontario and McMaster University OA Index (WOMAC) pain, activities of daily living (ADL), stiffness, total scores, average cartilage thickness (quantified using 3D-dual echo steady-state sequences on 3T MRI), Chair Stand Test (CST) and gait speed. Mixed effects and linear regression models were used for knee-level and participant-level outcomes, respectively. Models were adjusted for age, race, BMI, daily calories, physical activity, and medical insurance availability. Sex differences were tested by interactions between UPF and sex. RESULTS:Sex interactions were significant for WOMAC-pain, ADL, total, gait speed, cartilage thickness (p-interaction-range < 0.001-0.006). Greater UPF was associated with significantly worse pain (β = 0.17 [0.093, 0.242], p < 0.001), ADL (β = 0.59 [0.365, 0.832], p < 0.001), total scores (β = 0.81 [0.483, 1.13], p < 0.001), thinner cartilage (β = -0.013 [-0.02, -0.006], p < 0.001), slower gait (β = -0.035 [-0.042, -0.027], p < 0.001) in women. Sex interactions were non-significant for CST and WOMAC-stiffness (p-interaction = 0.319, 0.573, respectively). With greater UPF, CST and WOMAC-stiffness showed significant poor outcomes (β = -0.008 [-0.013, -0.004], p < 0.001, β = 0.04 [0.008, 0.064], p = 0.011, respectively). WOMAC-stiffness results were not significant after Bonferroni corrections. CONCLUSIONS:UPF-rich diet is linked to worse KOA outcomes disproportionately more in women.

OBJECTIVE:Mechanical loading is an important, modifiable risk factor for knee osteoarthritis. Identifying walking loads associated with disease worsening presents intervention opportunities. Our purpose was to evaluate the longitudinal relation of the baseline vertical ground reaction force (GRF) during walking to worsening bone marrow lesions (BMLs) and cartilage damage using cohort data from the Multicenter Osteoarthritis Study (MOST). METHODS:MOST participants with GRF data at baseline and magnetic resonance imaging examinations at baseline and 2-year follow-up were included. Peak impact force (PIF) and average loading rate (ALR) from the vertical GRF were analyzed with respect to four joint regions (i.e., the medial and lateral portions of the tibiofemoral and patellofemoral joints). Analyses used logistic regression with generalized estimating equations and adjusted for relevant covariates. RESULTS:Higher PIF was associated with increased odds of worsening BMLs in the lateral patellofemoral joint (odds ratio (95% confidence interval [CI]): 1.33 (1.11, 1.60)) and worsening cartilage damage in the lateral patellofemoral joint (1.48 (1.24, 1.77)), lateral tibiofemoral joint (1.24 (1.03, 1.50)), and medial tibiofemoral joint (1.25 (1.06, 1.48)). Higher ALR was associated with reduced odds of BML worsening in the lateral tibiofemoral joint (0.60 (0.41,0.87)). CONCLUSIONS:Higher peak forces when walking were associated with worsening BMLs in the lateral patellofemoral joint and with worsening cartilage damage in regions of the knee associated with higher contact forces during walking. Higher ALRs were not associated with increased odds of structural worsening (BMLs or cartilage).

Weight loss is considered an essential treatment component for knee osteoarthritis (OA), but data on weight-loss-associated changes in skeletal muscle is sparse. This study aimed to assess volumetric changes in major thigh muscle groups during weight loss. Included in this longitudinal analysis of the Osteoarthritis Initiative were participants who had (a) stable weight (SW) (body mass index (BMI) ± 3

Abstract Objective While risk factors for osteoarthritis (OA) are well known, it is not well understood why certain individuals maintain high mobility and joint health throughout their life while others demonstrate OA at older ages. The purpose of this study was to assess which demographic, clinical and MRI quantitative and semi-quantitative factors are associated with preserving healthy knees in older individuals. Methods This study analyzed data from the OA Initiative (OAI) cohort of individuals at the age of 65 years or above. Participants without OA at baseline (BL) (Kellgren-Lawrence (KL) ≤ 1) were followed and classified as incident cases (KL ≥ 2 during follow-up; n = 115) and as non-incident (KL ≤ 1 over 96-month; n = 391). Associations between the predictor-variables sex, age, BMI, race, clinical scoring systems, T2 relaxation times and Whole-Organ Magnetic Resonance Imaging-Score (WORMS) readings at BL and the preservation of healthy knees (KL ≤ 1) during a 96-month follow-up period were assessed using logistic regression models. Results Obesity and presence of pain showed a significant inverse association with maintaining radiographically normal joints in patients aged 65 and above. T2 relaxation times of the lateral femur and tibia as well as the medial femur were also significantly associated with maintaining radiographically normal knee joints. Additionally, absence of lesions of the lateral meniscus and absence of cartilage lesions in the medial and patellofemoral compartments were significantly associated with maintaining healthy knee joints. Conclusion Overall, this study provides protective clinical parameters as well as quantitative and semi-quantitative MR-imaging parameters associated with maintaining radiographically normal knee joints in an older population over 8 years.

Purpose (the aim of the study): The association of thigh intramuscular fat and knee OA has been investigated in previous studies. However, quantifying intramuscular fat in individuals without OA remains unexplored. This study aimed to (1) estimate the degree of thigh intramuscular fat in individuals without radiographic OA or frequent pain in the knee or hip, and to (2) assess the associations of age, sex, and BMI with the degree of intramuscular fat.

BACKGROUND:Preventing worsening osteoarthritis (OA) in persons with early OA is a major treatment goal. We evaluated if different early OA definitions yielded enough cases of worsening OA within 2-5 years to make trial testing treatments feasible. METHODS:We assessed different definitions of early OA using data from Multicenter Osteoarthritis (MOST) Study participants who were followed up longitudinally. We defined early OA as having at least minimal knee pain (WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index) pain ≥3/20) with different levels of pre-radiographic OA. For MRI, we required knee pain and used MRI definitions with combinations of cartilage damage, osteophytes, bone marrow lesions and meniscus damage.The primary outcome, worsening OA at 2 or 5 years, combined structural (Kellgren and Lawrence grade ≥2 with joint space narrowing ≥1) and symptom (WOMAC pain ≥6 with increase ≥2 from baseline) outcomes. We also examined structural and symptom outcomes separately. RESULTS:For worsening OA at 2 years, we included 750 participants (mean age 65 years, 60% female, 90% white, mean body mass index 29.2 kg/m2). Fewer than 10% of early OA knees had the combined outcome at 2 or 5 years. At 2 years, for several early OA definitions, roughly 20% of knees had either structural or symptom worsening outcomes. Two-year trials of either, but not both, outcomes would need to recruit over 1200 patients. CONCLUSION:Most knees with early OA are stable and do not progress. Some painful knees experience worse pain but not structural progression and vice versa. Trial testing treatments to prevent OA illness or disease will be challenging.

OBJECTIVE:To study the association between various radiographic definitions of acetabular dysplasia (AD) and incident radiographic hip osteoarthritis (RHOA), and to analyze in subgroups. METHODS:Hips free of RHOA at baseline and with follow-up within 4-8 years were drawn from the World COACH consortium. The Wiberg center edge angle (WCEA), acetabular depth width ratio (ADR), and the modified acetabular index (mAI) were calculated. AD was defined as WCEA≤25°, and for secondary analyses as WCEA≤20°, ADR ≤250, mAI ≥ 13°, and a combination. A logistic regression model with generalized mixed effects with 3 levels adjusted for age, biological sex, and body mass index (BMI) was used. Descriptive statistics stratified by age, biological sex and BMI were reported. RESULTS:A total of 18,807 hips from 9 studies were included. Baseline characteristics: age 61.84 (± 8.32) years, BMI 27.40 (± 4.49) kg/m², 70.1% women. 4766 hips (25.3%) had WCEA≤25°. Within 4-8 years (mean 5.8 ±1.6) follow-up, 378 hips (2.0%) developed incident RHOA. We found an association between AD and RHOA (adjusted OR [aOR] 1.80 95% confidence interval [CI] 1.40-2.34). In secondary analyses, all other definitions of AD were also associated with incident RHOA (aOR ranging from 1.52 95% CI 1.19-1.94 to 1.96 95% CI 1.26-3.02). Descriptive statistics showed that the relative risk (RR) in AD hips to develop RHOA was higher compared to non-AD hips in age group 61-70 (RR 1.70), BMI<25 (RR 1.66), and in female hips (RR 1.73). CONCLUSION:AD was consistently associated with incident RHOA. Explorative analyses show that AD hips in women and age group 61-70 years seem to be more at risk of developing RHOA compared to non-AD hips.

Purpose (the aim of the study): While studies have shown that increased thigh intermuscular fat is associated with reduced muscle strength in knee osteoarthritis (OA), the impact of intramuscular on knee strength is not well understood. This study aims to assess thigh intramuscular fat based on MRI and its associations with knee strength (flexion and extension) and function, focusing specifically on individuals without radiographic OA or frequent pain in the knee and hip.

INTRODUCTION Different hip morphologies have been shown to be risk factors for hip OA. However, the associations found differ in size and are sometimes even contradictory. Different follow-up times, radiographic hip OA (RHOA) definitions, and quantification of hip morphology might explain these conflicting results. Additionally, previous studies have shown that hip pain in combination with a different hip morphology might alter the association with incident RHOA. OBJECTIVE We aimed to investigate whether hips with both hip pain and different morphologies, i.e., acetabular dysplasia, pincer and cam morphology, had a different association with the development of incident RHOA compared to asymptomatic hips with the same morphology. METHODS Individuals from three prospective cohort studies, Cohort Hip and Cohort Knee (CHECK), Johnston County Osteoarthritis Project (JoCoOA) and the Multicenter Osteoarthritis Study (MOST), were included in the current study. Standardized anteroposterior (AP) pelvic or long-limb radiographs were taken at baseline and follow-up (CHECK 8 years, JoCoOA 6 years, MOST 5 years). Incident RHOA was defined at follow-up as KLG ≥ 2 or total hip replacement. The presence of hip pain was self-reported through survey questions and dichotomized. Hip morphology was automatically quantified on the baseline radiographs using an in-house developed, validated pipeline. Acetabular dysplasia was defined by a Wiberg center edge angle (WCEA) ≤ 25°, pincer morphology was defined by a lateral center edge angle (LCEA) ≥ 40°, and cam morphology was defined by an alpha angle ≥ 60°.Hips free of RHOA (KLG 0) at baseline, with available baseline age, body mass index (BMI), hip morphology measurements, and follow-up KLG were included. When assessing acetabular dysplasia, all hips with pincer morphology were excluded from the dataset, and vice-versa, so that the reference group only consisted of hips with a normal femoral head coverage by the acetabulum. This resulted in the inclusion of 1,768 participants for acetabular dysplasia, 1,526 participants for pincer morphology, and 1,985 participants for cam morphology, see Table 1.Three generalized mixed-effects logistic regression models with the interaction between hip pain and morphology were employed. These models included three random effect levels (cohort, individual, hip side) and were adjusted for age, biological sex, and body mass index (BMI). Associations were expressed as adjusted odds ratios (aORs) or ratios of aORs with 95% confidence intervals (CIs). RESULTS No significant interactions between hip morphology and pain and the development of incident RHOA were found, see Table 2. Additionally, only cam morphology was significantly associated with incident RHOA in asymptomatic hips, aOR 4.7 (95% CI 2.0 – 11.2). CONCLUSION The presence of hip pain was not an effect modifier in the association between acetabular dysplasia, pincer or cam morphology and the development of incident RHOA within 5-8 years. Future research should investigate whether including the duration and severity of symptoms also does not alter this relationship.

Objective Extrusion of the meniscus is known to be a key factor in the development of knee osteoarthritis. Here, we investigate the precise relationship between meniscal extrusion and weight bearing 3-D joint space width (JSW) distribution. Design Weight-bearing computed tomography (WBCT) images were acquired at the 144-month visit in the Multicenter Osteoarthritis Study. For each knee, 3-D JSW maps were created. MRI data from the same visit were assessed for both medial and lateral meniscal extrusion grade determined by the MRI Osteoarthritis Knee Score. Statistical parametric mapping (SPM) was used to test for any significant dependence of 3-D JSW distribution on meniscal extrusion grade by location. Results 568 knees were included in the analysis. SPM demonstrated significant differences in 3-D JSW distribution according to the extent of both medial and anterior extrusion of the medial meniscus in a pattern that suggested a posteromedial shift of the femur on the tibia. Medial extrusion of the medial meniscus was also associated with significantly lower JSW across the central-to-posterior medial joint space. The low prevalence of lateral meniscus extrusion meant that results for this may be underpowered or unrepresentative. Conclusion Establishing links between 3-D JSW and meniscal extrusion is an important step in developing the clinical utility of WBCT for the evaluation of osteoarthritis. Further study is required to establish whether 3-D JSW is sensitive enough to detect meniscal extrusion before articular cartilage damage has occurred, to develop this as a biomarker for early disease.

Purpose (the aim of the study): There has been growing interest in the identification of early knee osteoarthritis (OA) for clinical trials preventing worsening disease. This definition should identify knees at high risk of established OA within 2 years, the usual length of a prevention trial. Attempts defining early OA have focused on the absence of joint space narrowing on radiographs and the presence of pain. These criteria have included physical examination measures, which may not be reliable across examiners.

Objective:To determine the reliability and agreement of manual and automated morphological measurements, and agreement in morphological diagnoses. Methods:Thirty pelvic radiographs were randomly selected from the World COACH consortium. Manual and automated measurements of acetabular depth-width ratio (ADR), modified acetabular index (mAI), alpha angle (AA), Wiberg center edge angle (WCEA), lateral center edge angle (LCEA), extrusion index (EI), neck-shaft angle (NSA), and triangular index ratio (TIR) were performed. Bland-Altman plots and intraclass correlation coefficients (ICCs) were used to test reliability. Agreement in diagnosing acetabular dysplasia, pincer and cam morphology by manual and automated measurements was assessed using percentage agreement. Visualizations of all measurements were scored by a radiologist. Results:The Bland-Altman plots showed no to small mean differences between automated and manual measurements for all measurements except for ADR. Intraobserver ICCs of manual measurements ranged from 0.26 (95%-CI 0-0.57) for TIR to 0.95 (95%-CI 0.87-0.98) for LCEA. Interobserver ICCs of manual measurements ranged from 0.43 (95%-CI 0.10-0.68) for AA to 0.95 (95%-CI 0.86-0.98) for LCEA. Intermethod ICCs ranged from 0.46 (95%-CI 0.12-0.70) for AA to 0.89 (95%-CI 0.78-0.94) for LCEA. Radiographic diagnostic agreement ranged from 47% to 100% for the manual observers and 63%-96% for the automated method as assessed by the radiologist. Conclusion:The automated algorithm performed equally well compared to manual measurement by trained observers, attesting to its reliability and efficiency in rapidly computing morphological measurements. This validated method can aid clinical practice and accelerate hip osteoarthritis research.