Comparison of sagittal computed tomography and plain film radiography in a scaphoid fracture model

To compare computed tomography (CT) in the sagittal plane and plain film radiography in the diagnosis of scaphoid fracture and displacement.

Three groups of scaphoids (no fracture, undisplaced fractures, fractures with displacement > 1 mm) from 11 cadaver wrists were prepared. Each wrist then was imaged by using sagittal plane CT scans in the long axis of the scaphoid and plain film imaging using 6 standard scaphoid views. Eight readers from 3 specialties read each group of images. The sensitivity and specificity for the presence of fracture and fracture displacement were calculated in addition to interobserver and intraobserver reliabilities for each.

Both x-ray and CT scans showed a high sensitivity and specificity in detecting the presence of a fracture with no interspecialty differences. The sensitivity for displacement greater than 1 mm was lower for both modalities with no inter-specialty differences. The specificities for x-ray and CT for detecting displacement greater than 1 mm were 84% and 89%, respectively. The poor sensitivity for detecting displacement was explained by the low sensitivity of CT in the diagnosis of radial/ulnar displacement compared with x-ray and the low sensitivity of x-ray in the diagnosis of volar/dorsal displacement compared with CT scans. When fellowship-trained hand surgeons reviewed CT scans and plain films together the sensitivity and specificity for fracture displacement increased significantly. Intraobserver and interobserver reliability for both x-ray and CT scans was excellent except for the reading of CT scans among emergency physicians and for the reading of plain x-rays among senior house staff, representing moderate agreement.

Based on these results both CT scans in the sagittal plane and plain films accurately detect fractures with a high degree of interobserver and intraobserver reliability, but they fall short in detecting displacement greater than 1 mm.