Many Radiographic and Magnetic Resonance Imaging Assessments for Surgical Decision Making in Pediatric Patellofemoral Instability Patients Demonstrate Poor Interrater Reliability.

ABSTRACT

PURPOSE: To evaluate the interrater reliability of several common radiologic parameters used for patellofemoral instability and to attempt to improve reliability for measurements demonstrating unacceptable interrater reliability through consensus training. METHODS: Fifty patients with patellar instability between the ages of 10 and 19 years were selected from a prospectively enrolled cohort. For measurements demonstrating unacceptable interrater reliability (intraclass correlation coefficient [ICC] <0.6), raters discussed consensus methods to improve reliability and re-examined a subset of 20 images from the previous set of images. If reliability was still low after the second round of assessment, the measure was considered unreliable. RESULTS: Of the 50 included subjects, 22 (44%) were male and the mean age at the time of imaging was 14 ± 2 years. With 1 or fewer consensus training sessions, the interrater reliability of the following radiograph indices were found to be reliable trochlea crossing sign (ICC 0.625), congruence angle (ICC 0.768), Caton-Deshamps index (ICC 0.644), lateral patellofemoral angle (ICC 0.768), and mechanical axis deviation on hip-to-ankle alignment radiographs (ICC 0.665-0.777). Reliable magnetic resonance imaging (MRI) indices were trochlear depth (ICC 0.743), trochlear bump (ICC 0.861), sulcus angle (ICC 0.684), patellar tilt (ICC 0.841), tibial tubercle to trochlear groove distance (ICC 0.706), effusion (ICC 0.866), and bone marrow edema (ICC 0.961). CONCLUSIONS: With 1 or fewer consensus training sessions, the interrater reliability of the following patellofemoral indices were found to be reliable for trochlear morphology trochlea crossing sign and congruence angle on radiograph and trochlear depth, trochlear bump, and sulcus angle on MRI. Reliable patellar position measurements included Caton-Deshamps index and lateral patellofemoral angle on radiograph and patellar tilt and tibial tubercle to trochlear groove distance on MRI. Additional global measurements (e.g., mechanical axis deviation on standing radiographs) and MRI assessments demonstrated acceptable reliability. LEVEL OF EVIDENCE II, prospective diagnostic study.