2025 ISAKOS Congress in Munich, Germany

2025 ISAKOS Biennial Congress Paper


Accuracy of the Visual Assessment of Patellar Tracking Is Poor but Shows an Association with Lateral Trochlear Inclination

Alexander Bumberger , MD, Vienna AUSTRIA
Andrew J. Cosgarea, MD, Owings Mills, MD UNITED STATES
Petri J. Sillanpää, MD, PhD, Tampere FINLAND
Miho J. Tanaka, MD, PhD, Boston, MA UNITED STATES

Massachusetts General Hospital, Boston, MA, UNITED STATES

FDA Status Cleared

Summary

Evaluating patellar maltracking by visually assessing the J-sign seems to be unreliable among experts and shows variable accuracy depending on the degree of trochlear dysplasia.

Abstract

Background

The J-sign can be used to assess patellar tracking in patients with patellofemoral instability. Dynamic kinematic CT scanning (DKCT) has been used to quantify the J sign, although studies have shown limited reliability with visual observation. The purpose of this study was to analyze the reliability of visual J-sign assessments on videos with digital markers when compared to DKCT imaging, and to assess the role of anatomic risk factors on the accuracy of visually diagnosing patellar maltracking.

Methods

On videos of patients undergoing active knee extension, a digital reference line was placed in line with the femoral axis. Members of the International Patellofemoral Study Group were asked to use these marked videos to determine the presence or absence of maltracking (≥2 quadrants of lateral translation as measured on DKCT) (qualitative analysis). The expert group was then asked to grade patellar tracking in the same videos as previously described: Grade 0 (<1 patellar quadrant of lateral translation), 1 (1 to <2 quadrants), 2 (2 to <3 quadrants), or 3 (≥3 quadrants) (quantitative analysis). Inter- and intra-rater reliability was assessed using kappa statistics. The accuracy of the experts was calculated compared to measurements of patellar position as assessed by the patellar bisect offset (BO) on DKCT. We analyzed the potential influence of tibial tuberosity- trochlear groove (TTTG) distance, lateral trochlear inclination (LTI) and the Caton-Deschamps index (CDI) on the experts’ accuracy using a logistic regression model.

Results

20 patellofemoral experts participated in this study and evaluated 17 videos of knees. The study found moderate inter-rater reliability (κ = 0.52) and substantial intra-rater reliability (κ = 0.71) for qualitative (binary) assessments of patellar tracking. Quantitative (grading) assessments demonstrated substantial inter-rater (κ = 0.62) and intra-rater reliability (κ = 0.69). Mean accuracy for binary assessments was 70.6%, while accuracy for grading patellar tracking was lower at 36.8%, with variability across different grades.
LTI was significantly associated with the accuracy of experts to correctly diagnose the presence of patellar maltracking (beta = 0.19, p = 0.048), indicating an estimated 20.9% increase in accuracy by every degree of LTI. None of the other covariates (TTTG, CDI) were significantly associated with changes in accuracy.

Conclusion

Despite substantial intra-rater reliability and moderate inter-rater reliability for visual J-sign assessments, accuracy in grading patellar tracking remaied low, despite the use of digital markers. The severity of trochlear dysplasia (LTI) significantly influenced the accuracy of diagnosing patellar maltracking, suggesting that anatomical factors may play a role in the variability of visual assessments. This highlights the need for more reliable methods or additional imaging modalities to improve the consistency and accuracy of patellar tracking evaluations.