Roentgenographic measurement of pedicle screw penetration

Potential complications due to pedicle screw penetration of the anterior cortex include injury to vascular, visceral, ureteral, sympathetic, and neural structures. This study examined the accuracy of lateral roentgenographic techniques in determining actual screw penetration in vertebral levels T12 through S1 of ten unilateral sets of pedicles in five anatomic specimens. A true lateral roentgenogram alone was inaccurate for determining the penetration of the anterior cortex by a pedicle screw. The greatest discrepancy between roentgenographically apparent and actual screw penetration was found at the L4 and L5 levels. Deviation from a true lateral roentgenographic axis resulted in the most pronounced change in roentgenographically apparent screw penetration at L4 and L5. The roentgenographic axes resulting in the closest approximation of actual screw penetrations were 5 degrees and 10 degrees above the true lateral axis for the T12-L3 and the L4-S1 levels, respectively. At 50% apparent penetration, the screw may be safely assumed to not be penetrating the anterior cortex using a true lateral roentgenogram. At 80% apparent penetration, 30% and 10% probabilities of actual screw penetration of the anterior cortex exist at L4 and L5, respectively. At 100% apparent penetration, there is an almost 100% probability that the screw is actually protruding through the anterior cortex.