Background: This study compares median nerve cross-sectional area (CSA) measurements at the wrist obtained with ultrasound (US) and magnetic resonance imaging (MRI) using cadaveric measurements as the gold standard.Methods: Median nerve CSA was measured using US and MRI in 9 cadaveric wrists obtained from 5 subjects at 5 locations: distal forearm, proximal to tunnel inlet, at tunnel inlet, at tunnel outlet, and distal to tunnel outlet and then on identical cadaveric transverse sections obtained with a bandsaw. All US, MRI, and cadaveric measurements were repeated to determine reliability. Median nerves of 10 patients with clinical carpal tunnel syndrome (CTS) were measured with US and MRI using an identical method US.Results: Median nerve CSA MRI measurements correlated better (Pearson correlation: 0.80-0.95, P < .05) with cadaveric measurements than with US measurements (Pearson correlation: 0.61-0.79, P < .05). Median nerve CSA US measurements (8.6-12.5 mm 2 , P < .05) were smaller at all levels than MRI (11.3-14.7 mm 2 ) or cadaveric (11.0-14.9 mm 2 ) measurements while MRI and cadaver measurements were similar at all levels. Median nerve CSA MRI measurements in CTS patients were larger than US measurements at all levels.
Conclusion: Median nerve CSA measurements by MRI are larger than US measurements and correlated better with cadaveric measurements. Median nerve CSA criteria used for diagnosing CTS on US are not likely to be applicable to MRI. K E Y W O R D S cross sectional, measurement, median nerve, MRI, ultrasound