Accuracy of the visual assessment of patellar tracking is poor and is influenced by trochlear dysplasia

The J-sign is commonly used to assess patellar tracking in patellofemoral instability, though reliability is limited. This study analysed the diagnostic performance and reliability of digitally aided visual J-sign assessments by patellofemoral specialists compared to dynamic kinematic computed tomography scans (DKCT).

A diagnostic study was conducted with 20 experts assessing maltracking (≥2 quadrants lateral translation) in 17 standardised single-knee videos of patients with and without patellofemoral instability. A digital reference line was added to guide visualisation. Experts performed qualitative (binary) assessments and quantitative gradings (0-3) of patellar tracking. Inter- and intra-rater reliability was evaluated using kappa statistics. Diagnostic performance was compared to DKCT-based patellar lateralisation (bisect offset [BO]). Logistic regression analysed the influence of tibial tuberosity-trochlear groove (TTTG) distance, lateral trochlear inclination (LTI) and Caton-Deschamps Index (CDI) on accuracy.

Moderate inter-rater (κ = 0.52) and substantial intra-rater (κ = 0.71) reliability were found for binary assessments, while quantitative grading showed substantial inter-rater (κ = 0.62) and intra-rater (κ = 0.69) reliability. Mean accuracy for binary assessments was 70.6%, whereas grading accuracy was lower at 36.8%, with variability across grades. LTI significantly influenced accuracy (β = 0.19, p = 0.048), with a 20.9% increase in correct diagnoses per degree of LTI. TTTG and CDI had no significant effect.

Despite substantial intra-rater and moderate inter-rater reliability, accuracy in grading patellar tracking was low, particularly in quantitative assessments. Trochlear dysplasia affected diagnostic accuracy, suggesting the J-sign is less reliable in dysplastic patients. More advanced imaging methods may be necessary to improve patellar tracking evaluations.

Level III, diagnostic study.