O2058: Is Magnetic Resonance Imaging Reliable for Evaluation of the Anterolateral Ligament in Children and Adolescents?

Background: Growing research suggests that the anterolateral ligament (ALL) may contribute to knee stability, especially in the setting of anterior cruciate ligament (ACL) injury. The little available data on the ALL in the pediatric population is variable, with existing magnetic resonance imaging (MRI) research weighted toward patients with closing physes. The purpose of this study was to assess inter- and intra-observer reliability when identifying the ALL on MRI in children and adolescents.

Methods: Patients were included if they were < 18 years old and had a knee MRI with no abnormalities. Five reviewers (two pediatric musculoskeletal radiologists, a pediatric sports surgeon, an orthopaedic resident, and a medical student) developed a standardized protocol for identifying the ALL and reviewed all scans twice, several weeks apart. Reviewers were asked to identify the femoral and tibial insertions and determine whether the ligament was partially or fully visible. Inter- and intra-observer reliability were calculated using Fleiss’ kappa statistic (k).

Results: A total of 132 subjects with mean age 12.8±3.6 years were included (43.2% with open physes). Overall inter-observer reliability was poor for both rounds (k=0.09 and 0.06, respectively). It was similarly poor when analyzing agreement based on patient sex, physeal status, and T1/T2-weighted scans. Reliability was poor between nearly all pairs of reviewers, regardless of clinical background or experience. Intra-observer reliability was highly variable (k= 0.2 to 0.9), with higher level of training/experience seemingly correlated with lower agreement. Most reviewers had better intra-observer reliability with T1 scans than T2. There were no clear trends when analyzing intra-observer agreement along the lines of patient sex or physeal status.

Conclusion: Contrary to previous studies, inter-observer reliability was poor when attempting to identify the ALL on MRI of children and adolescents, including a large number of patients with open physes. Clinical background or degree of experience did not result in improved agreement. Intra-observer reliability was highly variable. MRI may not be a valid modality for consistent evaluation of the ALL in the pediatric population.