Background: The purpose of this study was to determine the accuracy of computed tomography and fluoroscopy in assessing joint penetration by periacetabular screws.
Methods: A 3.5-mm acetabular periarticular screw was inserted in each of thirty-nine cadaveric hemipelves. Twenty screws were intentionally directed to violate the articular surface, whereas nineteen screws were positioned to avoid the articular surface. Using two fluoroscopic views (tangential and axial) in a manner simulating the clinical setting, an examiner blinded to the actual screw location determined whether each screw was violating the articular surface. In addition, each hemipelvis was examined with computed tomography with use of two different techniques: (1) a 1-mm slice thickness at 1-mm intervals, and (2) a 4-mm slice thickness at 3-mm intervals. Each scan was evaluated by another examiner who was blinded to the actual screw location. Sensitivity, specificity, and percent correct interpretations were then calculated for each method.
Results: The sensitivity, specificity, and percent correct interpretations were 95%, 84%, and 90%, respectively, for axial fluoroscopy; 85%, 89%, and 87% for tangential fluoroscopy; 100%, 84%, and 92% for the computed tomography scans with a 1-mm slice thickness at 1-mm intervals; and 100%, 58%, and 79% for the computed tomography scans with a 4-mm slice thickness at 3-mm intervals. Tangential fluoroscopy was found to be more specific than the computed tomography scans with a 4-mm slice thickness at 3-mm intervals (p = 0.02). No other significant differences were found.
Conclusions: Fluoroscopy and computed tomography are equally accurate for determining intra-articular screw penetration. Computed tomography scans with thick slices (4 mm at 3-mm intervals) have a low specificity. Their use postoperatively may lead to a false-positive interpretation of the scan and unnecessary exploration of a hip for screw penetration.