PURPOSE: To evaluate the usefulness of thin-section proton density oblique sagittal MR imaging in the diagnosis of tear involving the anterior cruciate ligament (ACL). MATERIALS AND METHODS: In 61 arthroscopically confirmed cases (29 patients with ACL injury and 32 normal subjects), thin section proton-density images (TSPDI) were obtained and compared with conventional oblique sagittal PDI and T2-weighted images (T2WI). In TSPD imaging, the scan plane was parallel to the course of the ACL, based on a coronal scanogram; the parameters used were TR/TE 2000 msec/20 -33 msec, 2-mm slice thickness, 16 x16 cm FOV, 256 x192 matrix, two excitations, and no intersection gap. We evaluated the sensitivity and specificity of MR images for diagnosing ACL tear, and their quality, on the basis of whether or not they successfully visualised the anterior/posterior margin of the ACL and linear signal intensities within the ACL fascicles. We also investigated the effects of partial volume averaging between the proximal portion of the ACL and the lateral femoral condyle. RESULT: The sensitivity/specificity of TSPD imaging for diagnosing ACL tear were not significantly different from those of conventional oblique sagittal PDI and T2WI. In the ACL injury group, TSPDI was better in detecting increased signal intensity, ACL thickening, and visualization of torn ACL than conventional oblique sagittal PDI and T2WI. In normal subjects, image quality was constantly better on TSPDI than on conventional oblique sagittal PDI and T2WI. TSPDI clearly revealed the anterior margin in 31/32 cases (97%) and linear signal intensities within the ACL fascicles in all 32 (100%), and also markedly reduced the partial volume effect of the proximal ACL and lateral femoral condyle. CONCLUSION: In evaluating the ACL, the use of TSPD imaging is likely to lead to improved image quality. In addition, where routine MR imaging reveals indeterminate ACL injury, TSPDI can provide additional clues to diagnosis.