It is generally accepted that definitive diagnosis of partial anterior cruciate ligament (ACL) tears is difficult without arthroscopic examination. Compared with patients with complete ACL tears, patients with partial ACL tears require fewer ACL reconstructions and more often return to sports when conservatively managed. The distinction between partial and complete ACL tears influences patient management and prognosis, so it is clinically very important. The reported accuracy of magnetic

more »

... ce (MR) imaging performed with conventional twodimensional spin-echo sequences in the diagnosis of complete ACL tears is as high as 95%. MR imaging using conventional sequences, however, has not proved sensitive in the detection of partial ACL tears. The currently used three-dimensional (3D) gradient-echo (GE) MR imaging is known to provide high resolution with no loss of information caused by intersection gaps. Although several papers have suggested that 3D GE MR imaging is at least comparable to conventional sequences, little has been published concerning the distinction between partial and complete ACL tears in a large population. Moreover, few MR studies on ACL injuries have considered acute and chronic ACL tears separately. We designed this study to clarify the ability of 3D Field Echo (FE) MR imaging, which is one of the 3D GE sequences, to identify chronic partial ACL tears and to allow the distinction of partial from complete ACL tears. We compared the findings of chronic ACL tears through 3D FE MR imaging with those of an arthroscopic standard in a relatively large group of patients. The selection criteria for this study were as follows: 1) a history of 3D FE MR imaging prior to diagnostic arthroscopic examination; 2) a median interval since knee injury at the time of 3D FE MR imaging of more than four weeks; and 3) a partial or complete tear of ACL proved by diagnostic arthroscopic examination. Seventy patients fulfilled the selection criteria. There were 47 male and 23 female patients with an average age of 25.1 years (range, 14 to 54). On diagnostic arthroscopic examination, the damage to ACL was evaluated as a partial tear if a significant portion of at least one bundle was in continuity and was potentially functional as judged by palpation with a probe and arthroscopic anterior drawer testing. There were 16 partial ACL tears and 54 complete ACL tears. All of the MR images were obtained with a commercially available unit (MRT, Toshiba, Japan) operating at 0.5 tesla, and only the 3D Fourier transform FE sequence was performed. The repetition time was 51 msec, and the echo time 14 msec, with a flip angle of 45 degrees. All patients were imaged in the sagittal plane, and section thickness was 1.5 mm with section gaps of 0 mm. For this 3D FE MR imaging, eight or more sagittal sections are required to span both anterior and posterior cruciate ligaments, and normal ACL is consistently well depicted as a linear continuous low-signal band on three or more sections (Fig. 1a) . The 3D FE MR images were retrospectively interpreted by an experienced orthopaedist who was unaware of the arthroscopic findings; they were classified as either partial or complete tears. The MR images of the ACL tears were designated as partial on the basis of the appearance of a linear continuous low-signal band on one or two sagittal sections (Fig. 1b) , and as complete when there was no linear continuous low-signal band on any sagittal sections.