Citation

Kleinman PK, O'Connor B, Nimkin K, Rayder SM, Spevak MR, Belanger PL, Getty DJ, Karellas A. Pediatr. Radiol. 2002; 32(12): 896-901.

Affiliation

Department of Radiology, UMASS Memorial Health Care, Worcester, MA 01655, USA. paul.kleinman@tch.harvard.edu

Copyright

(Copyright © 2002, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/s00247-002-0778-4

PMID

12447601

Abstract

PURPOSE: To assess the diagnostic performance of digital radiography using charge-coupled device (CCD) technology in the detection of rib fracture in infant abuse. MATERIALS AND METHODS: Four fractured posterior rib arcs and eight normal ribs removed at autopsy from a 10-month-old abused infant were radiographed using a CCD prototype, four clinical film-screen systems, and direct-exposure film. Each rib was viewed with these six systems in nine different projections. The resultant 648 images were assessed for probability of fracture (0-100%) by four pediatric radiologists. The calculated area under the resultant ROC curves (A(z)) for the CCD was compared with those obtained with direct-exposure, high-detail, medium and fast film-screen radiographic systems. RESULTS: The mean A(z) for the CCD (0.937) fell within the range of the high-detail systems (0.934-0.940) and was significantly higher (p < 0.05) than the medium and fast system (0.861 and 0.858, respectively). Despite a lower spatial resolution than direct-exposure film (7 line pairs per millimeter vs. > 20 line pairs per millimeter) the CCD performance was comparable (A(z) = 0.944 vs. 0.937). The similar performance can be explained by the higher contrast resolution of the digital technology. CONCLUSION: This study indicates that in the ex vivo setting, digital radiology can perform comparably to high-detail film-screen imaging. The findings suggest that digital radiography has the potential to replace film-screen imaging in the evaluation of inflicted skeletal injury in infants.

Language: en