Session I - Pelvis


Thursday, October 12, 2000 Session I, Paper #7, 9:06 am

Reliability of Detecting Screw Penetration of the Acetabulum Using Intraoperative Fluoroscopy and Computed Tomography

David B. Carmack, MD; Berton R. Moed, MD; Kathleen McCarroll, MD; David Freccero, BS; Wayne State University, Detroit, MI

Purpose: To determine the value of computed tomography and intraoperative fluoroscopy in evaluating the location of periacetabular screws, specifically to determine if joint penetration has occurred.

Methods: A 3.5-mm acetabular periarticular screw was placed in each of 39 cadaver hemipelves. Twenty screws were directed intentionally to violate the articular surface. The remaining 19 screws were positioned so as to avoid the articular surface. Using 2 fluoroscopic views (tangential and axial) in a manner approximating the clinical setting, an examiner blinded to actual screw location made a determination as to whether or not each screw was violating the articular surface. In addition, each pelvis underwent computed tomography using two different techniques: 1) 1-mm slice thickness with 1-mm intervals and 2) 4-mm slice thickness with 3-mm intervals. A separate examiner blinded to actual screw location evaluated each scan. Subsequently, all the hips were dislocated, and actual screw location was confirmed by visual inspection of the acetabular subchondral surface. Sensitivity, specificity, and percent correct were then calculated for each method.

Results: The sensitivity, specificity, and percent correct (respectively) for each of the methods were as follows: axial fluoroscopy, 95%, 84%, 90%; tangential fluoroscopy, 85%, 89%, 87%; 1x1-mm computed tomography, 100%, 84%, 92%; and 4x3-mm computed tomography, 100%, 58%, 79%. Tangential fluoroscopy was found to be more specific than the 4x3-mm computed tomography (P = 0.02). Axial fluoroscopy and 1x1-mm computed tomography specificities approached statistical significance versus 4x3-mm computed tomography (P = 0.07). No other statistically significant differences were found.

Conclusions: Fluoroscopy and computed tomography appear to be equally effective for determining periacetabular screw location. Although computed tomography has been touted as the method of choice, its specificity is too low to be considered as such. This potential for false-positive postoperative computed tomographic results, which were unacceptably high for the 4x3-mm study, exists as an important but previously unrecognized risk for unnecessary reoperation.

Clinical Relevance: Overall, intraoperative fluoroscopy appears to be the most effective method to evaluate periacetabular screw location; its results can be used immediately, eliminating the risk of reoperation. If postoperative computed tomography is to be used for this purpose, the 1x1-mm study should be selected.