Reliability of patellotrochlear index in patellar height assessment on MRI-correction for variation due to change in knee flexion.

OBJECTIVES To assess the reliability of patellotrochlear Index (PTI) in patellar height assessment on successive MRI scans in asymptomatic patients. MATERIALS AND METHODS Sixty-four patients with two successive MRI scans (128 studies) of the same knee for non-patellofemoral joint symptoms were identified retrospectively. PTI and knee flexion angle were assessed independently by three observers to assess interobserver reliability. The effect of knee flexion on PTI was assessed by comparing the change in values of PTI in each patient correlated with change in knee flexion. RESULTS Sixty-four MRIs of patients (M:F) aged between 18 and 35 years (mean 24.6) years were assessed. The mean PTI for initial scan group was 0.33% (95% CI: 0.29-0.37; SD: 0.15) and consecutive scan group was 0.30% (CI: 0.27-0.33; SD: 0.3). The difference was not significant (p = 0.097 using a paired t test) with high inter-observer correlation (0.9) in both sets. Spearman's rho for knee flexion angle and PTI was found to be positive and statistically significant (0.41; p = 0.001). A linear regression model was derived using a scatter chart of change in PTI with change in knee flexion for each patient. The gradient of the linear regression line was used to estimate a cPTI (corrected PTI) value (corrected to 0 degrees of knee flexion), defined as cPTI = PTI - 1.3a (a = knee flexion angle). CONCLUSIONS This study demonstrates high inter-observer correlation of PTI on MRI and high test-retest reliability indicating unconscious quadriceps contraction does not change the index sufficiently. Knee flexion significantly alters PTI, increased patellotrochlear engagement with flexion increases the index. We propose use of the formula cPTI = PTI -1.3a to correct the index to 0 degree knee flexion in clinical practice