MRI criteria for diagnosis and predicting severity of carpal tunnel syndrome.

Abstract

To study MRI criteria for diagnosing and predicting severity of carpal tunnel syndrome (CTS). Sixty-nine wrists in 41 symptomatic CTS patients and 32 wrists in 28 asymptomatic subjects were evaluated by MRI. Circumferential surface area (CSA), flattening ratio, relative median nerve signal intensity, and retinacular bowing were measured. CTS severity was classified as mild, moderate, or severe. Parameters for patients with and without CTS and for the three severity groups were compared. ROC curves were plotted to assess accuracy for CTS diagnosis and severity prediction. Significant differences were found between CTS and control wrists for median nerve CSA, flattening ratio at inlet, relative median nerve signal intensity, and retinacular bowing. ROC curve analysis revealed a sensitivity, specificity, and accuracy of median nerve CSA >15mm2 proximal to the tunnel (CSAp) of 85.5, 100, and 90.1%. Using either CSAp or CSAd > 15mm2 as a diagnostic criterion, MRI could achieve a sensitivity of 100% and specificity of 94% for diagnosis of CTS while overall accuracy was 98%. Significant differences were found among the three severity groups. Sensitivity, specificity, and accuracy of prediction of severe CTS using for CSAp > 19mm2 were 75.0, 65.9, and 69.6%, respectively. MRI is highly accurate at diagnosing CTS and moderately accurate at determining CTS severity. We recommend using CSA > 15mm2 either proximal to or distal to the tunnel as a diagnostic criterion for CTS and CSA > 19mm2 proximal to the tunnel as a marker for severe CTS.