[Wrist lesions in MRI arthrography compared with wrist arthroscopy]

Wrist arthroscopy is routinely used for diagnosis of ligamentous lesions to the wrist. Although it is very sensitive and specific, it is also more invasive than other available diagnostic techniques.

In a prospective trial, 125 patients (80 men, 45 women) 37 years old (+/- 12) with clinical evidence of wrist lesions were evaluated with direct wrist arthro MRI. Within 24 hours following direct arthro MRI wrist arthroscopy was performed. The surgeons and radiologists were not aware of the other results until completion of their investigation. Specificity, sensitivity, positive predictive value, negative predictive value and accuracy were determined.

According to the MRI findings, lesions of the TFCC were suggested in 70 patients (56 % of all patients). In 65 cases this was confirmed by wrist arthroscopy. In the remaining 55 patients no TFCC lesions were detected by MRI. However in three cases TFCC lesions were found by arthroscopy. In the remaining 52 patients MRI accurately excluded TFCC lesions. There was a correlation of MRI and arthroscopy in detecting TFCC lesions in 93.6 %. Sensitivity was 94 %, specificity 89 %, positive predictive value 91 % and negative predictive value 93 %. Complete scapholunate ligament tears were detected by MRI in 12 (9.6 %) cases. Correlation with wrist arthroscopy was 99 %, sensitivity 92 %, specificity 100 %, positive and negative predictive value 100 % and 99 %, respectively. Accuracy for incomplete scapholunate lesions (n = 17, 13.6 %) and lunotriquetral tears (n = 4, 3.2 %) was poor (sensitivity 59 % and 25 %, specificity 100 % and 99 %).

Though sensitivity of MRI arthrography approaches that of arthroscopy, it cannot replace it at the moment. However, it is a potent additional tool for wrist diagnosis if intraarticular contrast is used. It can facilitate diagnosis and indications for surgery of the wrist. It may make arthroscopic and more invasive interventions for diagnostic purposes avoidable in future.