Gamekeeper thumb: comparison of MR arthrography with conventional arthrography and MR imaging in cadavers.

Abstract

To compare magnetic resonance (MR) arthrography with conventional arthrography and standard MR imaging in the evaluation of gamekeeper thumb. Eighteen cadaveric thumbs with (n = 12) or without (n = 6) experimental abduction stress were examined with conventional arthrography, standard MR imaging, and MR arthrography. Standard MR imaging and MR arthrography were performed with 0.2-T (low-field-strength) extremity-only and 1.5-T (high-field-strength) magnets. Results of blinded imaging analyses were correlated with pathologic findings. Paired MR images and MR arthrograms were rated with a three-point scale. Fourteen ulnar collateral ligaments were torn, including eight nondisplaced (57%) and six displaced (43%) tears. For the presence of tear, diagnostic accuracy of conventional arthrography, low-field-strength MR imaging, high-field-strength MR imaging, low-field-strength MR arthrography, and high-field-strength MR arthrography was 83%, 89%, 90%, 94%, and 100%, respectively. With regard to displacement of the torn ligament, diagnostic accuracy was 61%, 89%, 90%, 94%, and 100%, respectively. MR arthrograms were rated superior to standard MR images in 72% and 90% of specimens with low-field-strength and high-field-strength magnets, respectively. Of the tested methods, MR arthrography proved to be the most sensitive to the diagnosis of a torn ulnar collateral ligament of the first metacarpophalangeal joint and of displacement of the torn ligament.