Associations and Diagnostic Accuracy of Ultrasound Features in Knee Osteoarthritis: Cross‐sectional Results from a Large Community‐based Cohort

To determine 1) associations among ultrasound (US) features of knee osteoarthritis (KOA), radiographic KOA (rKOA), and patient-reported symptoms, and 2) diagnostic accuracy of US definitions for rKOA, in a community-based cohort. Participants enrolled in the Johnston County Health Study (JoCoHS, 2019-24, n=902) provided demographics, comorbidities, clinical features and symptoms, along with imaging with standardized acquisition and scoring protocols. Logistic regression models provided odds ratios adjusted for age, sex, race, ethnicity, body mass index (BMI), education level, comorbidities, and knee injury for associations among US features and KOA outcomes. Diagnostic accuracy was assessed using standard metrics with rKOA as the gold standard. Complete imaging data were available for 861 participants (1711 knees), 34% men, 25% Black, 10% Hispanic, mean age 55 years and mean BMI 33 kg/m2. Half of knees were symptomatic, 1/3 had rKOA, and 1 in 5 had symptomatic rKOA. US-identified osteophytes, effusion, meniscal extrusion, cartilage damage, calcium crystals, and popliteal cysts were associated with KOA outcomes. A US definition including both mild osteophytes and mild cartilage damage gave an area under the receiver operating characteristic curve (AUC) of 0.76 for diagnosing rKOA (validated in an external cohort). We identified common US features in participants with and without KOA, along with significant associations between US features and rKOA, symptomatic rKOA, and symptoms. US-based diagnosis of rKOA shows promise for general use. US is a valuable and accessible modality for assessment of knee OA features in clinical and research settings, including those with limited resources.