Reexamination of pelvic inlet and outlet images using 3-dimensional computed tomography reconstructions

OBJECTIVES Inlet and outlet views are essential in the evaluation of patients with pelvic injuries. The optimal angles that should be used to obtain these views are still debated. METHODS Computed tomography scans of 70 patients without pelvic ring injuries were analyzed. Using the raw data from the computed tomography images, virtual pelvic x-rays were generated. The images were rotated 1 degree at a time and a total of 360 images were generated. The same procedure was repeated to create solid 3-dimensional (3D) reconstructions. Two trauma-trained orthopaedic surgeons then reviewed all 360 images to select inlet and outlet views as described in previous literature. Intraclass correlation coefficients were calculated for each inlet and outlet value. RESULTS The average inlet angle (caudal tilt) that was required to achieve a view where the promontory overlaps the S1 body was found to be 26.7 degrees (95% confidence interval, 25-29) with virtual x-rays and 24.3 degrees (95% confidence interval, 22-26) with 3D reconstructions. The average outlet angle (cephalad tilt) that was required to achieve a view where the superior border of the pubic symphysis overlaps the S2 body was 43.7 degrees (95% confidence interval, 42-45) with virtual x-rays and 43.8 degrees (95% confidence interval, 42-45) with 3D reconstructions. There was no difference in these angles based on gender, whereas sacral dysmorphism increased the angle needed to obtain the outlet view by an average of 5 degrees. CONCLUSIONS This study demonstrated that the ideal angle to obtain inlet views should be approximately 25 degrees and to obtain outlet views should be approximately 45 degrees during screening x-rays to evaluate patients who were diagnosed with pelvic injuries. Application of these new values in the clinical practice would provide higher rates of adequate studies and further prospective clinical studies should be performed to validate these parameters.