Accuracy of Magnetic Resonance Imaging in the Diagnosis of Multiple Ligament Knee Injuries: A Multicenter Study of 178 Patients

Abstract

Background: Magnetic resonance imaging (MRI) has shown limited diagnostic accuracy for multiple ligament knee injuries (MLKIs), especially posterolateral corner (PLC) injuries. Hypothesis: The diagnostic accuracy of MRI for MLKIs will only be moderate for some knee structures. Patient-related factors and injury patterns could modify the diagnostic accuracy of MRI. Study Design: Cohort study; Level of evidence, 3. Methods: All patients with MLKIs surgically treated between January 2014 and December 2020 in the centers participating in the study were reviewed. We recorded sex, age, mechanism of injury, time from injury to MRI, and vascular and neurological associated lesions. Lesions to the anterior cruciate ligament (ACL), posterior cruciate ligament, medial collateral ligament, lateral collateral ligament (LCL), popliteus tendon, popliteofibular ligament, iliotibial band, biceps tendon, medial and lateral meniscus, and articular cartilage from MRI reports and surgical records were also collected. The sensitivity, specificity, positive predictive value, negative predictive value, diagnostic accuracy, diagnostic odds ratio, positive and negative likelihood ratio, and intraclass correlation coefficient of MRI were calculated for each knee structure. With logistic regression, associations between patient and injury characteristics and MRI accuracy were assessed. Results: A total of 178 patients (127 male; mean age, 33.1 years) were included. High-energy trauma was the most common mechanism of injury (50.6%), followed by sports trauma (38.8%) and low-energy trauma (8.4%). The ACL was the structure with the best diagnostic accuracy, diagnostic odds ratio, and positive predictive value (94.4%, 113.2, and 96.8%, respectively). PLC structures displayed the worst diagnostic accuracy among knee ligaments (popliteus tendon: 76.2%; LCL: 80.3%) and diagnostic odds ratio (popliteus tendon: 9.9; LCL: 17.0; popliteofibular ligament: 17.5). MRI was more reliable in detecting the absence of meniscal and chondral lesions than in identifying them. Logistic regression found that the diagnostic accuracy was affected by the Schenck classification, with higher Schenck grades having worse diagnostic accuracy for peripheral structures (iliotibial band, popliteus tendon, and biceps tendon) and improved diagnostic accuracy for the ACL and posterior cruciate ligament. Conclusion: The diagnostic accuracy of MRI for MLKIs largely varied among knee structures, with many of them at risk of a misdiagnosis, especially PLC, meniscal, and chondral lesions. The severity of MLKIs lowered the diagnostic accuracy of MRI for peripheral structures.