Radiographic Hip Osteoarthritis Research Articles

Kellgren & Lawrence grading in the check cohort: methods matter

Kellgren & Lawrence grading in the check cohort: methods matter

Is lesser trochanter size a novel risk factor for hip osteoarthritis? findings from the MROS study

Purpose: Previously, we used statistical shape modelling (SSM) of hip dual x-ray absorptiometry (DXA) scans to examine associations between two-dimensional (2D) whole hip shape and radiographic hip osteoarthritis (rHOA). A limitation in this approach is that it is difficult to know precisely which aspects of shape underlie the relationships we observed. We aimed to define which shape characteristics are related to rHOA in more detail, by examining individual subregions of the hip, and utilising three-dimensional (3D) hip shape derived from quantitative computed tomography (QCT) scans to validate our findings.

The population-based prevalence of hip symptoms, radiographic, and symptomatic hip OA over 26 years: the Johnston county osteoarthritis project

The population-based prevalence of hip symptoms, radiographic, and symptomatic hip OA over 26 years: the Johnston county osteoarthritis project

American Academy of Orthopedic Surgeons Appropriate Use Criteria for Hip Preservation Surgery: Variables That Drive Appropriateness for Surgery

Determining appropriate candidates for hip preservation surgery is challenging because criteria for judging appropriateness are not defined. The American Academy of Orthopaedic Surgeons (AAOS) recently published a hip preservation surgery appropriateness classification system using the RAND/University of California, Los Angeles approach. This study was undertaken to determin the extent and pattern of contribution of each of the variables used to predict the appropriateness of hip preservation surgery. An AAOS-appointed multidisciplinary expert panel wrote 270 clinical vignettes incorporating all permutations of 5 indication variables derived from an evidence synthesis. A second independent panel of experts rated the appropriateness of each vignette during multiple Delphi surveys. We used logistic regression to determine the relative contribution of each variable to classification. We also used a classification-tree approach to determine which indication variables, in combination, contributed to the final classification. Odds ratios from the regression indicated that patient age and radiographic hip osteoarthritis (OA) evaluation were the main indications of appropriateness classification (e.g., the odds ratio for age <40 years was >999.99, with age >65 years as the referent group, for "appropriate/may be appropriate" as compared to "rarely appropriate" vignettes). Hip range of motion, risk for negative outcome, and function-limiting pain did not meaningfully contribute to the final classification. The AAOS appropriateness classification system for hip preservation surgery is driven almost exclusively by age and radiographic hip OA evaluation. Additional research on appropriateness classification for hip preservation surgery is needed to identify important indication variables beyond age and radiographic hip findings.

Development and Validation of a Multitask Deep Learning Model for Severity Grading of Hip Osteoarthritis Features on Radiographs.

Background A multitask deep learning model might be useful in large epidemiologic studies wherein detailed structural assessment of osteoarthritis still relies on expert radiologists' readings. The potential of such a model in clinical routine should be investigated. Purpose To develop a multitask deep learning model for grading radiographic hip osteoarthritis features on radiographs and compare its performance to that of attending-level radiologists. Materials and Methods This retrospective study analyzed hip joints seen on weight-bearing anterior-posterior pelvic radiographs from participants in the Osteoarthritis Initiative (OAI). Participants were recruited from February 2004 to May 2006 for baseline measurements, and follow-up was performed 48 months later. Femoral osteophytes (FOs), acetabular osteophytes (AOs), and joint-space narrowing (JSN) were graded as absent, mild, moderate, or severe according to the Osteoarthritis Research Society International atlas. Subchondral sclerosis and subchondral cysts were graded as present or absent. The participants were split at 80% (n = 3494), 10% (n = 437), and 10% (n = 437) by using split-sample validation into training, validation, and testing sets, respectively. The multitask neural network was based on DenseNet-161, a shared convolutional features extractor trained with multitask loss function. Model performance was evaluated in the internal test set from the OAI and in an external test set by using temporal and geographic validation consisting of routine clinical radiographs. Results A total of 4368 participants (mean age, 61.0 years ± 9.2 [standard deviation]; 2538 women) were evaluated (15 364 hip joints on 7738 weight-bearing anterior-posterior pelvic radiographs). The accuracy of the model for assessing these five features was 86.7% (1333 of 1538) for FOs, 69.9% (1075 of 1538) for AOs, 81.7% (1257 of 1538) for JSN, 95.8% (1473 of 1538) for subchondral sclerosis, and 97.6% (1501 of 1538) for subchondral cysts in the internal test set, and 82.7% (86 of 104) for FOS, 65.4% (68 of 104) for AOs, 80.8% (84 of 104) for JSN, 88.5% (92 of 104) for subchondral sclerosis, and 91.3% (95 of 104) for subchondral cysts in the external test set. Conclusion A multitask deep learning model is a feasible approach to reliably assess radiographic features of hip osteoarthritis. © RSNA, 2020 Online supplemental material is available for this article.

Comorbid conditions and the transition among states of hip osteoarthritis and symptoms in a community-based study: a multi-state time-to-event model approach

BackgroundWe examined the association of three common chronic conditions (obesity, diabetes mellitus [DM], and cardiovascular disease [CVD]) with transitions among states of hip osteoarthritis (HOA).MethodsThis longitudinal analysis used data from the Johnston County Osteoarthritis Project (JoCo OA, n = 3857), a community-based study in North Carolina, USA, with 18.4 ± 1.5 years of follow-up. Transitions across the following states were modeled: development of radiographic HOA (rHOA; Kellgren-Lawrence grade [KLG] of< 2); development of hip symptoms (self-reported hip pain, aching, or stiffness on most days) or symptomatic HOA (sxHOA; rHOA and symptoms in the same hip), and resolution of symptoms. Obesity (body mass index ≥ 30 kg/m2) and self-reported DM and CVD were the time-dependent comorbid conditions of interest. Markov multi-state models were used to estimate adjusted hazard ratios and 95% confidence intervals to describe the associations between the conditions and HOA states.ResultsThe sample included 33% African Americans, 39% men, with a mean (SD) age of 62.2 (9.8) years; the frequencies of the comorbidities increased substantially over time. When considered individually, obesity was associated with incident hip symptoms, while CVD and DM were associated with reduced symptom resolution. For those with > 1 comorbidity, the likelihood of incident sxHOA increased, while that of symptom resolution significantly decreased. When stratified by sex, the association between obesity and incident symptoms was only seen in women; among men with DM versus men without, there was a significant (~ 75%) reduction in symptom resolution in those with rHOA. When stratified by race, African Americans with DM, versus those without, were much more likely to develop sxHOA.ConclusionsComorbid chronic conditions are common in individuals with OA, and these conditions have a significant impact on the persistence and progression of HOA. OA management decisions, both pharmacologic and non-pharmacologic, should include considerations of the inter-relationships between OA and common comorbidities such as DM and CVD.

Differences between race and sex in measures of hip morphology: a population-based comparative study

This paper aims to (i) identify differences in measures of hip morphology between four racial groups using anteroposterior (AP) hip x-rays, and (ii) examine whether these differences vary by sex. 912 hip x-rays (456 individuals) from four racial groups (European Caucasians, American Caucasians, African Americans and Chinese) were obtained. Males and females (45-75 years) with no radiographic hip OA (Kellgren and Lawrence<Grade 2 or Croft<Grade 1) were included. Eleven features of hip joint morphology were analysed. Linear regression with generalised estimating equations (GEE) was used to determine race and sex differences in hip morphology. Post-hoc Bonferroni procedure was used to adjust for multiple comparisons. The final analysis included 875 hips. Chinese hips showed significant differences for the majority of measures to other racial groups. Chinese were characterised by more shallow and narrow acetabular sockets, reduced femoral head coverage, smaller femoral head diameter, and a lesser angle of alignment between the femoral neck and shaft. Variation was found between other racial groups, but with few statistically significant differences. The average of lateral centre edge angle, minimum neck width and neck length differed between race and sex (p-value for interaction<0.05). Significant differences were found in measures of morphology between Chinese hips compared to African Americans or Caucasian groups; these may explain variation in hip OA prevalence rates between these groups and the lower rate of hip OA in Chinese. Sex differences were also identified, which may further explain male-female prevalence differences for OA.

A Pilot Randomized Controlled Trial for Aerobic and Strengthening Exercises on Physical Function and Pain for Hip Osteoarthritis

Hip osteoarthritis (OA) limits lower extremity physical function. Exercise therapy may improve physical function and reduce pain in patients with hip OA. To evaluate the feasibility of a randomized controlled trial (RCT) to measure the effect of a structured hip-specific resistance and aerobic exercise program on physical function and self-reported pain in adults with hip OA. Pilot RCT. Academic medical center. Thirty one adults with radiographic hip OA. Participants were randomly allocated in a 2:1 ratio to a 3-month structured exercise intervention (n = 21) or a 3-month waitlist control (n = 10). The 6-minute walk test was the primary outcome measure. Self-reported physical function was assessed with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) physical function subscale and pain visual analog scale (VAS). Measures of feasibility were assessed as secondary outcomes. From baseline to 3 months, the median change from baseline in distance covered during the 6-minute walk test in the intervention group (49 m) was double that of the control group (22 m), but the difference was not statistically significant (P = .13). Likewise, the WOMAC physical function subscale score median changes in the intervention group (5 points) were double that of the controls (2 points), although the difference was not statistically significant (P = .06). Median change in pain scores was slightly reduced in each group but not significantly different between groups (P = .53). Although no statistically significant between-group differences were found in this pilot RCT, improvements were seen in measures of physical function compared to controls. A larger RCT of this structured exercise program may be warranted.

Incidence rate and risk factors for radiographic hip osteoarthritis in Japanese men and women: a 10-year follow-up of the ROAD study

To investigate the incidence and progression rate of radiographic hip osteoarthritis (OA) and its risk factors in Japanese men and women using a large-scale population of a nationwide cohort study, Research on Osteoarthritis/osteoporosis Against Disability (ROAD). From the baseline survey of the ROAD study, 2,975 participants (1,043 men and 1,932 women) aged 23-94 years (mean, 70.2 years) living in urban, mountainous, and coastal communities were followed up with hip radiography at 3, 7, and 10 years (mean follow-up, 7.1 years). Radiographs were scored using the Kellgren/Lawrence (K/L) grading system, and radiographic hip OA was defined as K/L≥2. The incidence and progression rate of hip OA were examined. Acetabular dysplasia was defined as a central-edge angle <20°. Cox's proportional hazard model was used to assess risk factors for incident and progressive radiographic hip OA. The incidence rate of radiographic hip OA was 5.6/1,000 person-years and 8.4/1,000 person-years in men and women, respectively. The progression rate of hip OA was 2.2/1,000 person-years and 6.0/1,000 person-years in men and women, respectively. The significant risk factors for incident hip OA were age, obesity, and acetabular dysplasia at baseline (hazard risk [HR] 1.05, 95% confidence interval [CI] 1.03-1.08; 1.78, 1.10-2.75; 2.06, 1.30-3.17, respectively). The significant risk factors for progressive hip OA were baseline hip pain and acetabular dysplasia (HR 5.68, 95%CI 1.07-22.61; 14.78, 3.66-56.06, respectively). Continued longitudinal surveys of the ROAD study will contribute to knowledge about and potential prevention of incident and progressive hip OA.

Extracting Voxel-Based Cartilage Relaxometry Features in Hip Osteoarthritis Subjects Using Principal Component Analysis.

MRI-based relaxation time measurements provide quantitative assessment of cartilage biochemistry. Identifying distinctive relaxometry features in hip osteoarthritis (OA) might provide important information on regional disease variability. First, to incorporate fully automatic voxel-based relaxometry (VBR) with principal component analysis (PCA) to extract distinctive relaxometry features in subjects with radiographic hip OA and nondiseased controls. Second, to use the identified features to further distinguish subjects with cartilage compositional abnormalities. Cross-sectional. Thirty-three subjects with radiographic hip OA (20 males; age, 50.2 ± 13.3 years) and 55 controls participated (28 males; 41.3 ± 12.0 years). A 3.0T scanner using 3D SPGR, combined T1ρ /T2 , and fast spin echo sequences. Pelvic radiographs, patients' self-reported symptoms, physical function, and cartilage morphology were analyzed. Cartilage relaxation times were quantified using traditional regions of interest and VBR approaches. PCA was performed on VBR data to identify distinctive relaxometry features, and were subsequently used to identify a subgroup of subjects from the controls that exhibited compositional abnormalities. Chi-square and independent t-tests were used to compare group characteristics. Logistic regression models were used to identify the possible principal components (PCs) that were able to predict OA vs. control classification. In T1ρ assessment, OA subjects demonstrated higher T1ρ values in the posterior hip region and deep cartilage layer when compared with controls (P = 0.012 and 0.001, respectively). In T2 assessment, OA subjects exhibited higher T2 values in the posterior hip region (P < 0.001). Based on the PC score classification, 16 subjects without radiographic evidence of OA demonstrated relaxometry patterns similar to OA subjects, and exhibited worse physical function (P = 0.003) and cartilage lesions (P = 0.009-0.032) when compared with the remaining controls. The study identified distinctive cartilage relaxometry features that were able to discriminate subjects with and without radiographic hip OA effectively. 1 Technical Efficacy Stage: 2 J. Magn. Reson. Imaging 2020;51:1708-1719.

Hip joint moments in symptomatic vs. asymptomatic people with mild radiographic hip osteoarthritis

Our primary objective was to examine external hip joint moments during walking in people with mild radiographic hip osteoarthritis (OA) with and without symptoms and disease-free controls. Three groups were compared (symptomatic with mild radiographic hip OA, n = 12; asymptomatic with mild radiographic hip OA, n = 13; OA-free controls, n = 20). Measures of the external moment (peak and impulse) in the sagittal, frontal and transverse plane during walking were determined. Variables were compared according to group allocation using mixed linear regression models that included individual gait trials, with group allocation as fixed effect and walking speed as a random effect. Participants with evidence of radiographic disease irrespective of symptoms walked 14–16% slower compared to disease-free controls (p = 0.002). Radiographic disease without symptoms was not associated with any altered measures of hip joint moment compared to asymptomatic OA-free controls once speed was taken into account (p ≥ 0.099). People with both mild radiographic disease and symptoms had lower external peak hip adduction moment (p = 0.005) and lower external peak internal rotation moment (p < 0.001) accounting for walking speed. Among angular impulses, only the presence of symptoms was associated with a reduced hip internal rotation impulse (p = 0.002) in the symptomatic group. Collectively, our observations suggest that symptoms have additional mechanical associations from radiographic disease alone, and provide insight into potential early markers of hip OA. Future research is required to understand the implications of modifying walking speed and/or the external hip adduction and internal rotation moment in people with mild hip OA.

An automated workflow based on hip shape improves personalized risk prediction for hip osteoarthritis in the CHECK study

To design an automated workflow for hip radiographs focused on joint shape and tests its prognostic value for future hip osteoarthritis. We used baseline and 8-year follow-up data from 1,002 participants of the CHECK-study. The primary outcome was definite radiographic hip osteoarthritis (rHOA) (Kellgren-Lawrence grade ≥2 or joint replacement) at 8-year follow-up. We designed a method to automatically segment the hip joint from radiographs. Subsequently, we applied machine learning algorithms (elastic net with automated parameter optimization) to provide the Shape-Score, a single value describing the risk for future rHOA based solely on joint shape. We built and internally validated prediction models using baseline demographics, physical examination, and radiologists scores and tested the added prognostic value of the Shape-Score using Area-Under-the-Curve (AUC). Missing data was imputed by multiple imputation by chained equations. Only hips with pain in the corresponding leg were included. 84% were female, mean age was 56 (±5.1) years, mean BMI 26.3 (±4.2). Of 1,044 hips with pain at baseline and complete follow-up, 143 showed radiographic osteoarthritis and 42 were replaced. 91.5% of the hips had follow-up data available. The Shape-Score was a significant predictor of rHOA (odds ratio per decimal increase 5.21, 95%-CI (3.74-7.24)). The prediction model using demographics, physical examination, and radiologists scores demonstrated an AUC of 0.795, 95%-CI (0.757-0.834). After addition of the Shape-Score the AUC rose to 0.864, 95%-CI (0.833-0.895). Our Shape-Score, automatically derived from radiographs using a novel machine learning workflow, may strongly improve risk prediction in hip osteoarthritis.

Hip Radiograph Findings in Patients Aged 40 Years and Under with Posterior Pelvic Pain.

Several sacroiliac joint (SIJ) provocative tests used to assess posterior pelvic pain involve moving and stressing the hip. It is unknown if there is a subgroup of patients with posterior pelvic pain who have underlying hip deformity that could potentially influence performance and interpretation of these tests. To describe the prevalence of radiographic hip deformity and hip osteoarthritis in a group of adults 40 years old and under who met the clinical diagnostic criteria for treatment of posterior pelvic pain with an image guided intra-articular SIJ injection. Retrospective cohort study. Tertiary university orthopedic department PATIENTS (OR PARTICIPANTS): One hundred and forty-eight patients were evaluated (83% (123/148) female; mean age 31.3 ± 6.2 years). All had completed a trial of comprehensive noninvasive treatment for posterior pelvic pain and had a minimum of three positive SIJ provocative tests on physical examination. Retrospective review identified patients undergoing SIJ injection for pain recommended and performed by seven physiatrists between 2011 and 2017. Hip radiographs were read by a physician with expertise in hip measurements with previously demonstrated excellent intrarater reliability. Percentage of patients with hip deformity findings. No patients meeting the inclusion criteria had significant radiographic hip osteoarthritis (Tonnis ≥2 indicating moderate or greater radiographic hip osteoarthritis) and 4/148 (3%) were found to have mild radiographic hip osteoarthritis. Prearthritic hip disorders were identified in 123 (83%, 95% CI: 76, 89%) patients. For those patients with prearthritic hip disorders, measurements consistent with femoroacetabular impingement (FAI) were seen in 61 (41%) patients, acetabular dysplasia in 49 (33%) patients, and acetabular retroversion in 85 (57%) patients. Acetabular retroversion was identified in 43% (crossover sign) and 39% (prominent ischial spine) of patients. Approximately 57% of adult patients under the age of 40 years with the clinical symptom complex of SIJ pain were found to have radiographic acetabular retroversion. This is a higher percentage than the 5%-15% found in asymptomatic people in the current literature. Further study is needed to assess links between hip structure, hip motion, and links to pelvic pain including peri and intra-articular SIJ pain. III.

A 12-month prospective exploratory study of muscle and fat characteristics in individuals with mild-to-moderate hip osteoarthritis

BackgroundReductions in lower extremity muscle strength, size and quality and increased fat content have been reported in advanced hip osteoarthritis (OA). Whether these differences are also evident at earlier stages of the disease and the extent to which they might develop over time is unclear. The main purpose of this 12-month exploratory prospective study was to compare changes in muscle and fat characteristics in individuals with mild-to-moderate hip OA and healthy controls.MethodsFourteen individuals with mild-to-moderate symptomatic and radiographic hip OA (n = 9 unilateral; n = 5 bilateral), and 15 healthy controls similar in age and sex without symptoms or radiographic hip OA were assessed at baseline and at 12-month follow-up. Maximal voluntary isometric strength of the hip and knee muscle groups was assessed using an isokinetic dynamometer. Lower extremity lean and fat mass were assessed using dual-energy x-ray absorptiometry, and thigh muscle and fat areas and thigh muscle density were assessed using peripheral quantitative computed tomography.ResultsKnee extension (p = 0.01), hip extension (p < 0.01), hip flexion (p = 0.03), and hip abduction (p < 0.01) strength, lower extremity lean mass (p < 0.01), thigh muscle area (p = 0.03), and thigh muscle density (p < 0.01) were significantly lower in hip OA compared to controls. Hip extension (p < 0.05), hip flexion (p = 0.03), and hip abduction (p = 0.03) strength significantly declined over the follow-up period in the hip OA group.ConclusionsPre-existing deficits in hip muscle strength in individuals with mild-to-moderate hip OA were accentuated over 12-months, though no changes in symptoms or joint structure were observed. A longer follow-up period is required to establish whether strength deficits drive clinical and structural decline in these patients. Interventions to prevent or slow declines in strength may be relevant in the management of mild-to-moderate hip OA.

Individuals with mild-to-moderate hip osteoarthritis exhibit altered pelvis and hip kinematics during sit-to-stand

BackgroundPerformance of the sit-to-stand (STS) task is compromised in individuals with advanced hip osteoarthritis (OA). Understanding how STS performance is altered in individuals with mild-to-moderate hip OA may inform interventions to improve function and slow disease progression. Research questionDo trunk, pelvis, and hip biomechanics differ during a STS task between individuals with mild-to-moderate hip OA and a healthy, age-matched control group? MethodsThirteen individuals with mild-to-moderate symptomatic and radiographic hip OA and seventeen healthy, age-matched controls performed a standardized STS task. Data were acquired using a three-dimensional motion capture system. The primary outcome measures were task duration, sagittal and frontal plane trunk, pelvis, and hip joint angles, and sagittal and frontal plane trunk and hip joint moments. Comparisons of lower-limb measures were between the most affected side in the hip OA group and a randomly chosen limb for the control group, termed the index limb, prior to and following lift-off from the chair. ResultsParticipants with mild-to-moderate hip OA took longer to perform the STS task compared to controls. Prior to lift-off, the hip OA group exhibited greater posterior pelvic tilt, greater pelvic rise on the index side and less hip joint flexion relative to controls. Following lift-off, the hip OA group exhibited greater pelvic rise on the index side compared to controls. SignificanceIndividuals with mild-to-moderate hip OA exhibit subtle alterations in movement strategy compared to healthy controls when completing a STS task similar, to a small extent, to adaptations reported in advanced stages of the disease. Interventions to target these features and prevent further decline in physical function may be warranted in the management of mild-to-moderate hip OA while the opportunity remains.

Predictors of total hip replacement in community based older adults: a cohort study

Purpose: Multiple predictors of hip replacement have been described, but whether they are independent of the current gold standard predictors (pain and radiographic hip OA) is unclear. Therefore, we aimed to describe predictors of total hip replacement (THR) independent of hip pain and radiographic measures of hip OA in a community dwelling sample of older adults. Methods: At baseline, participants (n=682) had assessment of radiographic osteoarthritis (OA) and cam morphology (from pelvic radiographs), shape mode scores (from dual emission X-ray absorptiometry (DXA)) and femoral neck bone mineral density (BMD) (also from DXA). After 2.6 and 5 years, participants reported hip pain using WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index),and had hip structural changes assessed using magnetic resonance imaging (MRI). Incidence of THR was assessed over 13 years. Risk of THR was analysed using mixed-effect Poisson regression, and modelled using Cox proportional hazards regression. Results: Incidence of THR for OA over 13 years was 5.9% (n=40). As expected, WOMAC hip pain and hip radiographic OA both predicted risk of THR (Figure 1). Additionally, both shape mode 2 score (decreasing acetabular coverage) (IRR 1.53 per SD; 95% CI 1.09-2.14), and shape mode 4 scores (non-spherical femoral head) (IRR 0.61/SD; 95% CI 0.42-0.88) predicted risk of THR. Cam impingement (alpha angle>60°) increased risk of THR (IRR 2.65; 95% CI 1.34-2.32) (Figure 2), as did BMD of the neck of femur (IRR 1.73/SD, 95% CI 1.28-2.32). Bone marrow lesions increased THR risk (IRR 4.97; 95% CI 0.97 to 25.44) but this was not statistically significant, possibly due the smaller sub-cohort (n=185, p=0.055). Age, BMI and sex, and steps per day were not independent risk factors (see Table). Conclusions: In addition to hip pain and radiographic hip OA, measures of hip shape, CAM morphology and BMD independently predict THR risk. These factors can be used to develop models for improving prediction of THR.View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Individuals with mild-to-moderate hip osteoarthritis walk with lower hip joint contact forces despite higher levels of muscle co-contraction compared to healthy controls

Purpose: Conservative non-pharmacological interventions are recommended for the clinical management of hip osteoarthritis (OA), though no existing treatments can modify structural disease progression. Evidence-based treatment targets are required to inform the development of disease-course altering interventions. Abnormal hip joint loading is potentially modifiable (e.g. gait and/or neuromuscular retraining) and may play an important role in hip OA progression. No previous study has investigated whether individuals with mild-to-moderate hip OA exhibit altered hip joint contact forces (HJCF) during gait, or identified any relevant contribution of altered hip muscle co-contraction to hip joint loading in this patient population. This study aimed to compare: (1) HJCF between individuals with mild-to-moderate hip OA and healthy controls during walking using an electromyography (EMG)-informed modelling approach; and (2) hip muscle co-contraction levels between groups. Methods: Eighteen participants (age=65±7 yrs; body mass index=27.6±4.8 kg.m-2; 72% females) with mild-to-moderate hip OA and 23 healthy controls without hip pain or radiographic hip OA (age=60±8 yrs; body mass index=25.1±4.8 kg.m-2; 74% females) performed 10 over-ground walking trials at a self-selected speed. Three-dimensional lower-body motion, ground reaction forces, and surface EMG from sixteen lower-limb muscles were synchronously acquired.Hip joint contact forces were solvedby balancing the external joint moments with internal muscle forces estimated using a calibrated EMG-informed neuromusculoskeletal model. Hip muscle contributions and co-contraction ratios (abduction/adduction and flexion/extension) were determined. Demographic, spatiotemporal, and primary outcome measures were compared between groups using independent t-tests. A secondary analysis of HJCF peaks was performed usingaone-way analysis of covariance with body mass index as a covariate (p<0.05). Results: Height, body mass, body mass index, and walking speed did not differ between groups. Participants with hip OA walked with an 11% lower first peak (p=0.01) and 28% lower second peak (p<0.01) HJCF compared to controls (Fig. 1). Differences remained significant when adjusted for body mass index. Participants in the hip OA group exhibited higher levels of hip muscle co-contraction in both flexion/extension (p<0.01) and abduction/adduction (p=0.04) during all phases of stance compared to controls. Hip muscle forces (p<0.01), peakhip extension (p<0.01) and peak hip adduction (p=0.03) joint angles, and peak hip flexion (p=0.01) and peak hip extension (p<0.01) joint moments were also lower in the hip OA compared to control group. Conclusions: Individuals with mild-to-moderate hip OA walk with lower HJCF, despite exhibiting higher levels of hip muscle co-contraction, compared to healthy controls. These observations contradict previous assumptions regarding the effect of higher levels of muscle co-contraction on joint loading in this cohort.This is the first study to demonstrate that under-loading is characteristic of individuals with early hip OA, and may suggest that interventions aimed at reducing hip joint loading are not appropriate for this patient group. Future studies are required to evaluate the effects of under-loading on hip cartilage health, with the aim of identifying modifiable treatment targets to slow disease progression in individuals with hip OA.

What to deploy to prevent hip OA; population attributable fractions of risk factors for incident clinical and radiographic hip OA

What to deploy to prevent hip OA; population attributable fractions of risk factors for incident clinical and radiographic hip OA

Detecting hip osteoarthritic degenerative changes in MRI using deep learning

Purpose: Osteoarthritis (OA) is a debilitating joint disease that involves the degeneration of the articular cartilage and affects more than 25% of the adult population. Previously, SHOMRI (Scoring Hip Osteoarthritis with MRI) has been proposed as an MR-based hip OA evaluation system, based on the presence of morphological abnormalities, such as bone marrow edema, cartilage lesions, labral abnormalities and cysts. However, manual identification of bone and cartilage abnormalities in MR images can be laborious and time consuming. Thus, the goal of this study was to develop a fully automated deep learning pipeline to identify morphological and degenerative changes in patients with OA of the hip joint that would act as a radiology assist in image interpretation. Methods: 133 subjects with radiographic or symptomatic hip OA (as graded by SHOMRI) were recruited for this study (age 42.88±12.75 years, BMI 22.94± 3.029 kg/m3, 75 males, 58 females). T2-weighted, fat-saturated coronal hip images (TR = 60 ms, TR = 2.4s, slice thickness=4 mm, matrix=288x224, FOV=14-20 cm) were acquired on a 3T Discovery 750 MR scanner (GE Healthcare, Waukesha, WI). An end-to-end automated pipeline was built to evaluate morphological degenerative changes (Fig. 1). It included an object detection deep convolutional neural network (DCNN) that generated cropped images of the hip joint and a classification DCNN that identified the presence of morphological bone and cartilage abnormalities. The object detection network was implemented in Python and TensorFlow Object Detection API (Google, Mountain View, CA) on single-shot detector with Resnet-50 feature pyramid network (RetinaNet) architecture. The model was pre-trained on ImageNet classification and COCO object detection datasets and trained with Nvidia Titan X GPU on a dataset of 70 hip MR images with bounding boxes around the femoral head (90% train / 10% validation, random flip augmentation, batch=8, 25000 iterations). The network’s output was the central slice and the bounding box over the femoral head, which was used to extract a set of 5 slices, cropped around the weight bearing region of the joint (Fig. 1). These cropped images were labeled by two trained radiologists on the presence of bone marrow edema and cartilage lesions. The classification DCNN was implemented using PyTorch framework with DenseNet-100 architecture, pretrained on grayscale CIFAR-32 dataset. The model used an SGD optimizer with initial learning rate of 0.005 reduced to 0.001 after 20 epochs, momentum=0.9, weight_decay=5e-4, and random horizontal flip augmentation. It was trained on datasets of 539 and 647 images (edema, cartilage lesions) with batch of 60, 170 epochs, 65-20-15% train-validation-test split. Saliency maps were generated using Grad-CAM algorithm. Results: The detection network showed excellent convergence, without overfitting at 25000 iterations (Fig. 2). The object detection showed a mean intersection over union of 0.92±0.04 on the validation dataset. Figure 3 shows cross-entropy loss and precision/recall plots of the bone marrow edema detection network with the corresponding confusion matrix shown in figure 4. It achieved sensitivity of 0.73 and specificity of 0.92 on the validation dataset. The cartilage lesion detection network had sensitivity of 0.67 and specificity of 0.68 after 13 epochs. An example of a saliency map generated during inference with the bone marrow edema classification network is shown in figure 5. Conclusions: Our results have demonstrated the feasibility of automated identification of OA degenerative changes. Automated cropping using deep detection network allowed to focus attention on the weight-bearing area of the joint and reduced the number of parameters. Transfer learning using pre-trained detection and classification checkpoints allowed to achieve good results in a few iterations, despite having an unbalanced dataset for edema classification, and a small number of images in the training set. Future work includes training and evaluating the pipeline with a significantly-larger set of clinical hip MRI exams. View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Semi-automated longitudinal assessment of quantitative joint space width at the hip in a community-based cohort

Purpose: Computer-assisted quantitative joint space (qJSW) measurements may provide a more accurate measurement of hip joint width and help identify other risk factors for hip OA. The aim of this study was to characterize qJSW over time in the Johnson County Osteoarthritis Project (JoCo OA), overall and by incident radiographic and symptomatic hip OA. Methods: The JoCo OA is a longitudinal study of civilian, non-institutionalized African American and white men and women greater than 45 years old at baseline who were residents of one of six designated townships in Johnson County for at least one year. For this analysis, we selected individuals with anteroposterior pelvis x-rays from baseline and 3 follow up time points, approximately 6 years apart. Per protocol, women under 50 years old did not have pelvis radiography. All hips were assigned a Kellgren-Lawrence grade (KLG) by an expert musculoskeletal radiologist (JBR). At all time points, participants were asked about the presence of hip symptoms (On most days, do you have pain, aching, or stiffness in your (right, left) hip?”). Incident radiographic hip radiographic OA (rHOA) was defined as KLG 2 or more at follow up (among those without rHOA at baseline), while incident symptomatic hip OA (sxHOA) required both rHOA and symptoms in the same hip. QJSW measurements were performed at 3 locations on each hip radiograph: 10, 30 and 50 degrees with respect to an anatomical landmark. Assessment was performed using a previously described method by an independent reader (JD) blinded to other radiographic or clinical information (Figure 1). With qJSW as the outcome over time from baseline, longitudinal mixed models (LMM) for each angle were stratified by sex, adjusted for continuous age, height, BMI, time, and categorical weight gain or loss (defined as a 5% change in BMI) to estimate model-based fixed effects (using restricted maximum likelihood); additional stratification by incident rHOA or sxHOA was also performed. LMM were assessed at the hip level with an unstructured covariance matrix, accounting for random effects at the participant level (intercept and time) and at the hip level (intercept). Results: There were 577 participants with 1154 hip x-rays meeting the above criteria; after exclusion of those missing hip films or with hip replacement, 397 participants and 784 hips were analyzed (59% women, 24% African American, average baseline age 57.5 for women and 55.2 for men). The overall incidence of rHOA and sxHOA was about 20% and 12%, respectively. In men, qJSW at the 10, 30, and 50 degree measurements did not vary significantly with time; in general, qJSW was wider with increasing height in males. In women, compared to 10 and 30 degrees, the qJSW measurements at 50 degrees were more sensitive to change over time and are discussed below and shown in Figure 2. As shown in Figure 2 A-B, taller women had greater qJSW although this declined with age. Taller women (e.g., at a representative value of 165cm) with higher BMI (here, at a representative value of 34 kg/m2) had more significant narrowing if they gained weight compared with those that maintained weight. Women with normal BMI (24 kg/m2) at the same height (165cm) had a similar decrease in qJSW regardless of gaining or maintaining weight. Shorter women (155cm) who were obese (34 kg/m2) had a decline in qJSW regardless of change in weight. However, shorter women (155cm) with average BMI (24 kg/m2) showed a significant decrease in qJSW if they maintained weight compared to those that gained weight. As shown in Figure 2 C-D, women who developed HOA according to KLG criteria had significantly narrower qJSW at baseline (this was also true for men, data not shown) and greater loss over time than those who did not develop HOA. Conclusions: In this community-based cohort, sex differences were observed in qJSW measurements of the hip, suggesting that qJSW changes over time may not be as sensitive in males, potentially due to anatomical differences in men and women. In women, the 50 degree measurements appear to be the most sensitive, and the impact of gaining or maintaining weight appeared to vary based on height. At 50 degrees in both men and women, rHOA and sxHOA were strongly associated with a narrower qJSW, suggesting that qJSW may prove to be a valid measurement tool.