Khalil W. Abu-Zahra MD, Presenter: Nothing to Disclose

Kenjirou Ohashi MD, Abstract Co-Author: Nothing to Disclose

Georges Yousef El-Khoury MD, Abstract Co-Author: Nothing to Disclose

Kaoru Sasaka MD, Abstract Co-Author: Nothing to Disclose

Kevin Staley Berbaum PhD, Abstract Co-Author: Nothing to Disclose

To measure the reproducibility of Letournel acetabular fracture classification by radiography and multi-detector-row computed tomography (MD-CT) using multiplanar reconstruction and three-dimensional (3D) imaging, and to assess whether or not multiple radiographs are still necessary for classification.

Over a period of 25 months, 82 imaging studies from 77 patients (55 men and 22 women, average age of 44 ranging from 15 - 86) were available for review. Two musculoskeletal radiologists independently reviewed and classified the fractures by radiography and MD-CT separately. At the end of each MD-CT reading, the readers were shown the radiographs to see if they would change their classification based on MD-CT. Radiography was reviewed on a workstation. CT studies were performed with a four-detector row helical CT scanner with a 3 mm beam collimation and a pitch of 4.5. The reconstructed images (3 mm slice thickness at 1 mm interval) were sent to a workstation, where multiplanar reformations and 3D images were reviewed. Interobserver agreement was tested by calculating kappa values. Medical records were reviewed for surgical findings. The readers’ classifications based on radiography and MD-CT were separately compared with the surgeons’ classifications.

Kappa values between the two readers were 0.38 (fair reliability) by radiography and 0.59 (moderate reliability) by MD-CT. There were discrepancies between radiographic and MD-CT classifications in transverse, T-shaped, transverse plus posterior wall, and both column types. After including the standard Judet views, MD-CT classification changed in only 2 cases (one case from each reader). Fifty-eight cases (70%) underwent surgery. A McNemar test showed that agreement with the surgeons’ classification was significantly higher (p < 0.05) for MD-CT (72% and 74% for the two readers) than for radiography (53% and 60% for the two readers).

Letournel acetabular fracture classification using multiplanar reconstruction and 3D imaging from MD-CT data is moderately reliable. The standard Judet pelvic radiographs add little to improve the MD-CT classification.

Abu-Zahra, K, Ohashi, K, El-Khoury, G, Sasaka, K, Berbaum, K, Assessing Reliability of the Letournel Acetabular Fracture Classification Using Multi-Detector Row CT: Are the Standard Judet Radiographic Views Still Necessary?.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4404881.html