Intra-and interobserver reliability of determining the femoral footprint of the torn anterior cruciate ligament on MRI scans

Abstract

Abstract Background Re-injury rates following reconstruction of the anterior cruciate ligament (ACL) are significant; in more than 20% of patients a rupture of the graft occurs. One of the main reasons for graft failure is malposition of the femoral tunnel. The femoral origin of the torn ACL can be hard to visualize during arthroscopy, plus many individual variation in femoral origin anatomy exists, which may lead to this malpositioning. To develop a patient specific guide that may resolve this problem, a preoperative MRI is needed to identify the patient specific femoral origin of the ACL. The issue here is that there may be a difference in the reliability of identification of the femoral footprint of the ACL on MRI between different observers with different backgrounds and level of experience. The purpose of this study was to determine the intra- and interobserver reliability of identifying the femoral footprint of the torn ACL on MRI and to compare this between orthopedic surgeons, residents in orthopedic surgery and MSK radiologists. Methods MR images of the knee joint were collected retrospectively from 20 subjects with a confirmed rupture of the ACL. The 2D (coronal, sagittal, transversal) proton-density (PD) images were selected for the segmentation procedure to create 3D models of the femurs. The center of the femoral footprint of the ACL on 20 MRI scans, with visual feedback on 3D models (as reference) was determined twice by eight observers. The intra- and interobserver reliability of determining the center of the femoral footprint on MRI was evaluated. Intraclass correlation coefficients (ICCs) were calculated for the X, Y and Z coordinates separately and for a 3D coordinate. Results The mean 3D distance between the first and second assessment (intraobserver reliability) was 3.82 mm. The mean 3D distance between observers (interobserver reliability) was 8.67 mm. ICCs were excellent (> 0.95), except for those between the assessments of the two MSK radiologists of the Y and Z coordinates (0.890 and 0.800 respectively). Orthopedic surgeons outscored the residents and radiologists in terms of intra- and interobserver agreement. Conclusion Excellent intraobserver reliability was demonstrated (< 4 mm). However the results of the interobserver reliability manifested remarkably less agreement between observers (> 8 mm). An orthopedic background seems to increase both intra- and interobserver reliability. Preoperative planning of the femoral tunnel position in ACL reconstruction remains a surgical decision. Experienced orthopedic surgeons should be consulted when planning for patient specific instrumentation in ACL reconstruction.