The posterior cruciate ligament–posterior femoral cortex angle: a reliable and accurate MRI method to quantify the buckling phenomenon of the PCL in ACL‐deficient knees


Renaud Siboni 1, 2 · Charles Pioger 1, 3 · Caroline Mouton 1, 4 ·

Romain Seil 1, 4, 5

Link to the article

1 - Department of Orthopaedic Surgery, Centre Hospitalier Luxembourg-Clinique d’Eich, 78 Rue d’Eich,1460 Luxembourg, Luxembourg

2 - Medicine and Science, Luxembourg, Luxembourg Department of Orthopaedic Surgery, Reims Teaching Hospital, Hôpital Maison Blanche, 45 Rue Cognacq-Jay, 51092 Reims, France

3 - Department of Orthopaedic Surgery, Ambroise Paré Hospital, Paris Saclay University, 9, avenue Charles de Gaulle, 92100 Boulogne-Billancourt, France

4 - Luxembourg Institute of Research in Orthopaedics, Sports Medicine and Science, Luxembourg, Luxembourg

5 - Human Motion, Orthopaedics, Sports Medicine and Digital Luxembourg Institute of Research in Orthopaedics, Sports Methods, Luxembourg, Luxembourg 



The aim was to validate a new MRI method to measure the buckling phenomenon of the PCL, representative of anterior tibial translation, by comparing its reliability and accuracy to identify anterior cruciate ligament (ACL)-deficient knees with existing methods.


Patients were selected retrospectively and separated into a group of primary ACL injuries and ACL-intact knees. Exclusion criteria were: skeletal immaturity, PCL or a concomitant collateral ligament injury, signs of osteoarthritis (> 1 Kellgren and Lawrence score), tibial plateau fracture, previous ACL reconstruction or displaced meniscal bucket handle tear. The assessment of the curvature of the anterolateral bundle of the PCL was performed on T2 sagittal MRI slices according to 3 methods: (1) the PCL angle (PCLA), (2) the PCL inclination angle (PCLIA) and (3) a new method: the PCL-posterior cortex angle (PCL–PCA), representing the angle between the vertical part of the PCL-ALB and the posterior diaphyseal cortex of the femur. For each method, the inter- and intra-observer reliability was measured. The ability to discriminate both ACL-deficient and ACL-intact knees was evaluated using ROC curves.


Twenty-four patients were included in each group. Intra-observer reliability was excellent for all 3 methods (ICCs > 0.90). Inter-observer reliability was excellent for the PCL–PCA (ICC > 0.90) and good for the PCLA and the PCLIA (ICCs between 0.75 and 0.90). The PCL–PCA had the highest precision (lowest standard error of measurement: 2.7°). It yielded an excellent discrimination between the ACL and CTL groups (AUC 0.80 [0.67–0.93]) with the highest sensitivity (71% [52.8–89.2]) and specificity (88% [75–100]) for a positive threshold when the angle was ≤ 22.7°. The PCLA and PCLIA methods led to acceptable discrimination and lower sensitivities and specificities (PCLA: AUC 0.71, sensitivity 63%, specificity 79%, threshold ≤ 117.9°; PCLIA: AUC 0.62, sensitivity 50%, specificity 83%, threshold ≤ 21.4°).


In comparison with previously described methods, the PCL–PCA was the most reliable and accurate method to measure the PCL buckling phenomenon on MRI in anterior cruciate ligament (ACL)-deficient knees. It offers an easy and objective method for the follow-up of ACL-injured patients and can therefore be recommended for routine use.