MR arthrography of the hip with and without leg traction: Assessing the diagnostic performance in detection of ligamentum teres lesions with arthroscopic correlation

Abstract

ObjectivesTo retrospectively assess the diagnostic performance of hip MR arthrography with and without traction in detecting ligamentum teres (LT) lesions with arthroscopic correlation and to evaluate the effect of traction on the imaging appearance of the LT. Methods73 MR arthrograms (73 consecutive patients, mean age, 34.5 years; range, 14–55 years) obtained without and with leg traction (application of 15–23kg, use of a supporting plate for the contralateral leg) were included. Two blinded readers independently evaluated LT lesions on MR arthrograms on separate occasions: coronal images without traction; coronal images with traction; a multiplanar traction protocol. MR findings were correlated with arthroscopic records. Sensitivity/specificity of traction and non-traction imaging was compared on coronal images with the exact McNemar test. Imaging appearance of the LT with and without traction was assessed in consensus and compared on coronal images using McNemar and McNemar–Bowker tests. (p<0.05, * corrected for type I error). ResultsWith arthroscopy 29 (40%) LT lesions were identified in 73 patients. Sensitivity was 72%/90% (without traction/with traction; p=0.25*), specificity was 89%/77% (p=0.25*) for reader 1 in assessing coronal images and for reader 2 sensitivity was 59%/86% (p=0.044*) and specificity was 93%/82% (p=0.25*). Alterations in fiber orientation, signal intensity, surface, dimension, fiber continuity after application of traction were observed in 33/73 (45%, p=0.002*), 6/73 (8%, p=0.223), 9/73 (12%, p=0.36*), 6/73 (8%, p=0.031) respectively 9/73 (12%, p=0.003) cases. Traction-related alterations in at least one criterion were observed in 41/73 (56%) cases. ConclusionApplication of traction can considerably alter the imaging appearance of the LT and resulted in higher rates of true-positive and false-positive findings compared to conventional MR arthrography.