Assessing intra-rater reliability of peripheral quantitative computed tomography in knee joint bone evaluation on individuals with and without obesity: A GRRAS study

Common imaging approaches for knee osteoarthritis (KOA) do not reliably assess early bone changes. Obesity is a major KOA risk factor and can complicate imaging due to positioning and image quality. This pilot study primarily aimed to assess the intra-rater reliability of peripheral quantitative computed tomography (pQCT) for evaluating bone at the distal femur and proximal tibia, with a secondary objective of comparing reliability metrics between individuals with and without obesity. Fifty-two participants (81% female, mean age 50 [range 23-79] years) were matched by body mass index (BMI) (obese: BMI ≥ 30 kg/m 2 , non-obese: BMI < 30 kg/m 2 ), within ± 5 years of age, sex, and >5 kg/m 2 BMI difference. A single technician acquired two independent pQCT scans (Stratec XCT 3000) of the right tibiofemoral joint during one visit. Reliability was assessed using coefficient of variation root mean squared (CV rms ), reliability coefficient (R), and Bland-Altman analyses for repeated bone density (total, trabecular, cortical) and cross-sectional area (CSA) measures overall and by obesity group. pQCT demonstrated no systematic bias overall, with no substantial differences in reliability between obesity groups. Density measures showed excellent precision at all tibial regions and the medial femoral condyle. Tibial CSA measures showed no systematic bias but low precision. CSA measures at the medial femoral condyle were unreliable, and all measures at the lateral femoral condyle were unreliable due to inconsistent image capture. pQCT is a potential tool to reliably detect early KOA-related bone changes, with no substantial differences in reliability between individuals with and without obesity.