The diagnostic accuracy of ultrasound in assessing anterior cruciate ligament tears: a systematic review and meta-analysis

BackgroundAdvances in technology and affordability of ultrasound (US) have expanded its use in the evaluation of musculoskeletal injuries. However, US accuracy for diagnosing anterior cruciate ligament (ACL) injuries remains inconsistently reported in the literature.Hypothesis/purposeSystematically review the literature and perform a meta-analysis to statistically evaluate the accuracy of US in diagnosing ACL tears.MethodsPubMed, Embase, and Web of Science databases were queried for studies comparing US to MRI or arthroscopy for diagnosing ACL tears. Studies evaluating US techniques including anterior tibial translation relative to the femur, hematoma at the ACL insertion ("notch sign"), or direct ACL visualization were pooled using random-effects models to generate sensitivity, specificity, positive predictive values (PPV), and negative predictive values (NPV). A p-value less than 0.05 was used to indicate statistical significance.ResultsAcross 12 studies, US correctly identified 419/518 (81%) of confirmed ACL injuries and 425/465 (91%) of uninjured ACLs. Meta-analysis showed high overall specificity (92%, I2 = 24%), but variable sensitivity (84%, I2 = 86%). PPVs were high for notch sign (96%, I2 = 0%) and anterior translation (97%, I2 = 0%) but low for visualization (77%, I2 = 70%). NPVs were heterogeneous across techniques (notch sign 73%, I2 = 90%; visualization 87%, I2 = 0%; anterior translation 83%, I2 = 88%).ConclusionUltrasound demonstrates high specificity (92%) and PPV (92%) for diagnosis of ACL tears compared to confirmed tears but is limited by fair sensitivity (84%) and NPV (83%).