Multidetector computed tomography arthrography of the knee: diagnostic accuracy and indications.

Purpose: To evaluate the diagnostic accuracy and indications of arthrography with multidetector computed tomography (arthro-MDCT) of the knee, in patients with absolute or relative contraindications to MRI and in patients with periarticular metal implants using diagnostic arthroscopy as the gold standard. Materials and methods: After intra-articular injection of iodixanol and volumetric acquisition, 68 knees in patients of both sexes (30 females, 38 males, age range 32-60 years) were examined with a 16-detector-row CT scanner. The patients had arthralgia but no radiologically detected fractures. They could not be studied by MRI either because of absolute contraindications (subcutaneous electronic implants), surgical metal implants or claustrophobia. In 37 of 68 patients who had had previous knee surgery, the arthro-CT examination was preceded by an MRI on the same day. All examinations were interpreted by two experienced musculoskeletal radiologists. The findings were compared with arthroscopic findings carried out within 28 days of the CT study. Results: In non-operated patients the comparison between arthro-MDCT and arthroscopy showed sensitivity and specificity ranging between 86% and 100%. In the 37 operated knees, arthro-MDCT had an accuracy of 95% compared with 53% of the MRI. Inter-observer agreement was almost perfect (K = 0.97) in the evaluation of all types lesions, both on MDCT and MRI. When arthro-MDCT was compared with MRI in post-operative patients by a McNemar test, a significant difference (p < 0.05) was found between these two techniques. Conclusions: Arthro-MDCT of the knee is a safe technique that provides accurate diagnosis in identifying chondral, fibrocartilaginous and intra-articular ligamentous lesions, in patients that cannot be evaluated by MRI, and in patients after surgical.