MDCT and radiography of wrist fractures: radiographic sensitivity and fracture patterns

The purpose of our study was to determine which wrist fractures are not prospectively diagnosed at radiography using CT as a gold standard and to identify specific fracture patterns.

Through a search of radiology records from January 1 to December 31, 2005, 103 consecutive patients were identified as having radiographic and CT examinations of the wrist. After excluding incomplete or nondiagnostic examinations and those with a greater than 6-week interval between imaging studies, the final study group consisted of 61 wrist examinations in 60 patients. Two musculoskeletal radiologists and one emergency radiologist blindly reviewed CT examinations, and each bone (scaphoid, lunate, triquetrum, pisiform, trapezium, trapezoid, capitate, hamate, metacarpals, distal radius, distal ulna) was categorized as normal or fractured, with agreement reached by consensus. Each prospective radiographic report was categorized as either normal or fracture/equivocal for each osseous structure. Results were compared using the chi-square and Fisher's exact tests.

In the proximal carpal row, lunate and triquetrum fractures were often radiographically occult (0% and 20%, respectively, detected at radiography); whereas in the distal carpal row, trapezoid, capitate, and hamate fractures were often occult (0%, 0%, and 40% detected at radiography, respectively). Hamate fractures were significantly associated with metacarpal fractures, and distal radius fractures were associated with scaphoid and ulna fractures.

Thirty percent of wrist fractures were not prospectively diagnosed on radiography, suggesting that CT should be considered after a negative radiographic finding if clinically warranted. The location of a dorsal scaphoid avulsion fracture emphasizes the need for specific radiographic views or cross-sectional imaging for diagnosis.