A controlled, prospective study of 30 patients with suspected acute internal derangement of the knee was undertaken to evaluate the sensitivity, specificity, predictive value, and accuracy of nonorthogonal (oblique) sagittal magnetic resonance imaging (MRI) in the assessment of anterior cruciate ligament (ACL) injuries. Thirty patients with acute hemarthroses underwent MRI within 12 days of injury, followed by arthroscopy within 24 h of the MRI. A control population of 30 chondromalacia patients underwent similar evaluation. In the acute hemarthrosis patient population, the incidence at arthroscopy of acute complete ACL tears was 60% (18/30); of partial ACL tears, 13.3% (4/30); and of chronic tears, 10% (3/30). A normal ACL was found in 16.6% (5/30) of patients. In the MRI evaluation of patients with suspected ACL injury the following results were obtained for both acute and chronic complete disruption on orthogonal (sagittal) and nonorthogonal (oblique sagittal) imaging, respectively: sensitivity, 61 (16/26) versus 100%; specificity, 70 (21/34) versus 100%; positive predictive value, 61 (16/26) versus 100%; negative predictive value 70 (24/34) versus 100%; and accuracy, 66 (40/60) versus 100%. In the evaluation of partial ACL injury, four partial tears were correctly diagnosed on nonorthogonal MRI, with one false-positive diagnosis. Orthogonal imaging failed to correctly identify any of the partial ACL injuries. Two patients in the control population demonstrated evidence of chronic ACL tears. We believe that we have demonstrated the superiority of T2-weighted nonorthogonal sagittal over conventional orthogonal sagittal ACL MRI in the evaluation of ACL injury.