Lateral Patellar Inclination Angle Measured via a Two-Image Technique on Axial Magnetic Resonance Imaging.

Abstract

Lateral patellar inclination (LPI) measures patellar tilt and is historically described on axial X-ray or a single magnetic resonance image (MRI). Given the variability in patellar height, LPI may be better represented by performing this measurement on two separate axial MRI images. We hypothesized that a two-image LPI measurement would be different from the current single-image LPI and have similar, if not superior reliability. Sixty-five patients treated for patellar instability (PI) between 2014 and 2017 were identified. Single image and two-image LPI were measured on axial MRI images. All measurements were performed by two independent observers. Reliability analysis was based on three observers' measurements of 30 randomly selected patients. Both the one image and two image LPI showed good inter-rater reliability (intraclass correlation coefficient [ICC] = 0.71 and 0.89, respectively), although the two image LPI had less variability. Both single image and two image LPI had near perfect intra-rater reliability (ICC = 0.98 and 0.98, respectively). Average single image LPI (14.6 ± 9.9 degrees) was 6.1 ± 3.4 degrees less than the average two image LPI (19.6 ± 9.4 degrees) (p = 0.037). Referencing a previously described 13.5 degrees maximum threshold, 54% of the patients had excessive patellar tilt based on single image LPI, while 73% had pathologic patellar tilt based on two image LPI. Two image LPI has similar reliability with less inter-rater variability compared with the historical single image LPI measurement. Significantly greater patellar tilt was identified with two image LPI that was found with single image LPI. A larger percentage of patients were classified as having pathologic patellar tilt based on two image LPI than single image LPI. The two image LPI provides more consistent and representative measurements of patellar tilt. Previously described threshold values for patellar tilt should be re-examined using this new measurement technique to appropriately risk stratify patients with PI and patellofemoral pain.