Comparisons of diagnostic performance and the reliability in identifying ACL injury between two measuring protocols of anterior tibial subluxation on MR images.

Abstract

To evaluate the diagnostic performance in identifying an anterior cruciate ligament (ACL) injury and the reliability between two measuring protocols of anterior tibial subluxation (ATS). A total of 165 patients with ACL injury and 157 ACL-intact patients were included in this study. Two different measuring protocols of ATS were performed on sagittal MR images, including the modified protocol using the longitudinal tibial axis (axis protocol) and the established protocol using a line perpendicular to the tibial plateau (plateau protocol). Receiver-operating characteristic (ROC) curves were calculated to evaluate the diagnostic performance in identifying an ACL injury, and areas under the curves (AUCs) were compared between the two protocols. Intra- and interobserver reliability tests were performed to evaluate the reliability of the measurements. Lateral ATS (P < 0.001) and medial ATS (P < 0.001) were increased in patients with ACL injury under both protocols. To identify an ACL injury, ATS measured under the axis protocol showed higher AUC values than the plateau protocol, including lateral ATS (AUC 0.828 vs. 0.688, P < 0.001), medial ATS (AUC 0.829 vs. 0.789, P = 0.013), and the combined indicator of lateral and medial ATS (AUC 0.885 vs. 0.810, P < 0.001). Reliability tests showed that both protocols were reliable. ATS measured under the modified protocol using the longitudinal tibial axis showed superior diagnostic performance in identifying an ACL injury compared to the established protocol, indicating that the modified protocol may better reflect the characteristics of an ACL-deficient knee.