Introduction
Increased posterior tibial slope (PTS) has been linked to higher rates of anterior cruciate ligament (ACL) injury and ACL graft failure post-reconstruction. In recent years, more specific investigations have delineated distinct roles of the medial tibial plateau (MTP) and lateral tibial plateau (LTP) geometries on anteroposterior and rotatory stability of the knee. While the diagnostic workup of patients with ACL insufficiency typically consists of radiographs and magnetic resonance imaging (MRI), computed tomography (CT), which is the gold standard for evaluating the osseous geometry of the proximal tibia, is not routinely obtained. With increased understanding of the roles of the MTP and LTP geometries on tibiofemoral kinematics, accurate sagittal slope measurements of both the MTP and LTP will be crucial for further defining the treatment algorithm for ACL insufficiency. It is currently unknown whether traditional MRI can be used to accurately measure PTS and how well these measurements correlate to those on CT. Therefore, the purpose of this study was to compare PTS measurements of the MTP and LTP on MRI versus CT.
Methods
After Institutional Review Board approval was obtained, an institutional picture archiving and communication system imaging database was retrospectively queried for patients that received concurrent MRI and CT imaging of the same knee within a one-year interval. Knees with significant arthrosis (Kellgren-Lawrence grade >2), deformity, proximal tibia fracture, or artifact that obscured visualization of proximal tibia landmarks were excluded. On paired MRI and CT studies, PTS of the MTP and LTP were measured using a previously described, validated method (Hudek et al, Clin Orthop Relat Res, 2009) by two independent raters. Interrater reliability of PTS measurements was assessed using the intraclass correlation coefficient (ICC). Intermethod agreement between MRI and CT measurements was assessed using the ICC and Bland-Altman analysis.
Results
A total of 46 knees in 45 patients (mean age, 37.3 ± 12.9 years, 73.9% male) met final inclusion criteria. Interrater reliability was high for MRI measurements (ICC = 0.78-0.83) and moderate for CT measurements (ICC = 0.64-0.80). Mean measurements of PTS at the MTP were 3.5 ± 2.6° (range, -4.7-8.4°) and 3.6 ± 3.2° (range, -3.8-10.4°) on MRI and CT, respectively. Mean measurements of PTS at the LTP were 4.7 ± 3.4° (range, -2.9-10.6°) and 4.9 ± 3.6° (range, -3.6-15.7°) on MRI and CT, respectively. The mean absolute differences in PTS measurements between MRI and CT were 2.5 ± 2.0° and 2.3 ± 1.7° at the MTP and LTP, respectively. Poor intermethod agreement between MRI and CT was observed at the MTP (ICC = 0.34-0.42), while moderate intermethod agreement was observed at the LTP (ICC = 0.59-0.70). Bland-Altman plots demonstrated high variability and minimal bias of PTS measurements on MRI compared to CT (0.16° [95% Limit of Agreement (LOA) -6.1-6.4°] for MTP; 0.22° [95% LOA -5.1-5.5°] for LTP).
Conclusion
In this study, PTS measurements on MRI demonstrated lower interrater reliability compared to PTS measurements on CT. Poor-to-moderate agreement was observed for measurements of PTS of the MTP and LTP between paired MRI and CT studies. Therefore, measurements of medial and lateral PTS may not be reliable on traditional MRI.