[Computerized tomography false positives in knee medial meniscus injuries. Retrospective study of 400 cases]

We retrospectively evaluated the diagnostic reliability of high resolution CT (HRCT) in the study of knee joint conditions.

We reviewed the results of 400 HRCT examinations of the knee performed in 1993 and compared them with arthroscopic findings. The patients were affected with knee pain and the physical findings were insufficient for diagnosis but a sufficient indication for further instrumental tests. CT slices were 1 mm, with .5 mm feed; the images were reconstructed on the coronal and sagittal planes (2 mm thickness and 2 mm gap) with a CT Sytec 3000 unit (GE Systems, Milwaukee, WI, USA).

The patients positive for meniscal fracture were submitted to arthroscopy and operated on if positive arthroscopically. We had 164 surgical cases even though the meniscus was damaged in 236 cases. CT was negative in 16 of 164 surgical cases. The meniscal injury was demonstrated with both CT and arthroscopy in 236 cases; there were 8 false positives and 32 questionable cases. In the latter, arthroscopy showed 24 cases of meniscal degeneration and 8 horizontal fractures. Diagnostic accuracy was 96%. CT showed the 8 false positives as radial injuries of the free margin, but in fact they were false images caused by a normal fold which increases in knee valgus.

HRCT is more reliable than conventional CT (diagnostic accuracy: 96% vs 85-90%) and MRI (90% reliable). The use of thin slices also permitted to show that a meniscal tissue interruption on the radial margin is not referrable to fracture, as indicated also in the recent literature, but to a normal fold.