PURPOSE: To evaluate and compare conventional magnetic resonance (MR) imaging and MR arthrography in the diagnosis of the most common traumatic metacarpophalangeal (MCP) joint injuries, which were created surgically in cadavers. MATERIALS AND METHODS: Injuries to various MCP joint structures were surgically created randomly in 28 fingers of seven human cadaveric hands. Injuries to the main collateral ligaments (CLs) (n = 12), accessory CL (n = 15), sagittal band (n = 14), transverse fibers of the extensor hood (n = 5), first annular pulley (n = 16), deep transverse metacarpal ligament (DTML) (n = 5), and palmar plate (n = 10) were analyzed. Conventional MR images and MR arthrograms were evaluated, with differences in interpretation resolved in consensus. The sensitivities, specificities, and accuracies of both MR imaging methods were determined, and the differences were tested for significance by using the McNemar test. RESULTS: Sensitivity was 28.6%-93.8% with conventional MR imaging versus 50.0%-93.3% with MR arthrography. Specificity was 66.7%-100% with conventional MR imaging versus 83.3%-100% with MR arthrography. Although the MR arthrographic results usually were higher, the differences were not significant. The kappa values for interobserver agreement were 0.314-0.638 for conventional MR imaging versus 0.364-1.00 for MR arthrography. Sensitivity for the detection of lesions of the main and accessory CLs and the first annular pulley was slightly higher than that for the detection of lesions of the extensor hood, DTML, and palmar plate structures. CONCLUSION: MR imaging and MR arthrography enable the diagnosis of simulated MCP joint injuries. MR arthrography does not have a significant advantage over conventional MR imaging.