Anatomic Determinants of Sacral Dysmorphism and Implications for Safe Iliosacral Screw Placement.

BACKGROUND Upper sacral segment dysplasia increases the risk of cortical perforation during iliosacral screw insertion. Dysmorphic sacra have narrow and angled upper osseous corridors. However, there is no validated definition of this anatomic variation. We hypothesized that pelves could be quantitatively grouped by anatomic measurements. METHODS One hundred and four computed tomography (CT) scans and virtual outlet views of uninjured pelves were analyzed for the presence of the five qualitative characteristics of upper sacral segment dysplasia. CT scans were reformatted to measure the cross-sectional area, angulation, and length of the osseous corridor. Principal components analysis was used to identify multivariable explanations of anatomic variability, and discriminant analysis was used to assess how well such combinations can classify dysmorphic pelves. RESULTS The prevalences of the five radiographic qualitative characteristics of upper sacral segment dysplasia, as determined by two reviewers, ranged from 28% to 53% in the cohort. The rates of agreement between the two reviewers ranged from 70% to 81%, and kappa coefficients ranged from 0.26 to 0.59. Cluster analysis revealed three pelvic phenotypes based on the maximal length of the osseous corridor in the upper two sacral segments. Forty-one percent of the pelves fell into the dysmorphic cluster. The five radiographic qualitative characteristics of dysmorphism were significantly more frequent (p < 0.007) in this cluster. A combination of upper sacral coronal and axial angulation effectively explained the variance in the data, and an inverse linear relationship between these angles and a long upper sacral segment corridor was identified. A sacral dysmorphism score was derived with the equation: (first sacral coronal angle) + 2(first sacral axial angle). An increase in the sacral dysmorphism score correlated with a lower likelihood of a safe transsacral first sacral corridor. No subjects with a sacral dysmorphism score >70 had a safe transsacral first sacral corridor. CONCLUSIONS Sacral dysmorphism was found in 41% of the pelves. The major determinants of sacral dysmorphism are upper sacral segment coronal and axial angulation. The sacral dysmorphism score quantifies dysmorphism and can be used in preoperative planning of iliosacral screw placement.