A Three-Dimensional Cement Quantification Method for Decision Prediction of Vertebral Recompression after Vertebroplasty

Abstract

Objective. Proposing parameters to quantify cement distribution and increasing accuracy for decision prediction of vertebroplasty postoperative complication. Methods. Finite element analysis was used to biomechanically assess vertebral mechanics ( $n=51$ ) after percutaneous vertebroplasty (PVP) or kyphoplasty (PKP). The vertebral space was divided into 27 portions. The numbers of cement occupied portions and numbers of cement-endplate contact portions were defined as overall distribution number (oDN) and overall endplate contact number (oEP), respectively. And cement distribution was parametrized by oDN and oEP. The determination coefficients of vertebral mechanics and parameters ( $R^2$ ) can validate the correlation of proposed parameters with vertebral mechanics. Results. oDN and oEP were mainly correlated with failure load ( $R^2=0.729$ ) and stiffness ( $R^2=0.684$ ), respectively. oDN, oEP, failure load, and stiffness had obvious difference between the PVP group and the PKP group ( $P<0.05$ ). The regional endplate contact number in the front column is most correlated with vertebral stiffness ( $R^2=0.59$ ) among all regional parameters. Cement volume and volume fraction are not dominant factors of vertebral augmentation, and they are not suitable for postoperative fracture risk prediction. Conclusions. Proposed parameters with high correlation on vertebral mechanics are promising for clinical utility. The oDN and oEP can strongly affect augmented vertebral mechanics thus is suitable for postoperative fracture risk prediction. The parameters are beneficial for decision-making process of revision surgery necessity. Parametrized methods are also favorable for surgeon’s preoperative planning. The methods can be inspirational for clinical image recognition development and auxiliary diagnosis.