gms | German Medical Science

Objective: Prophylactic vertebroplasty is expected to reinforce weak vertebral bodies (VB) and possibly to minimize fracture risks. Cement leakage can cause severe complications, whereas smaller volumes of cement may reduce its occurrence. Significant mechanical benefit is only obtained when the cement filling connects both endplates inevitably increasing the leakage risk. A better understanding of the cement flow into bone is thus required. The aim of the current study is to develop methods for cement flow monitoring and assessment of the mechanical impact during stepwise injection.

Methods: Eight fresh frozen human cadaveric vertebrae were used for vertebroplasty. Firstly, they were XtremeCT-scanned (Scanco, Switzerland). The VB were then fixed on a custom carbon holder in the center of a CT bore (Siemens Emotion6). A syringe driver (Harvard Apparatus, USA) was connected to a PEEK cannula for stepwise cement injection (Vertecem V+, Synthes GmbH, Switzerland). Injection forces were recorded using a load-cell (Type 9211, Kistler Instrumente AG, Switzerland) connected to the syringe plunger. Each CT image was then registered and resampled to the original XtremeCT image, and the cement cloud was segmented. Image data for each incremental step were then prepared for micro finite element modeling (microFE). The models were solved for stiffness and Von Mises Stress (VMS) distribution. Finally, the VB were tested biomechanically in axial compression (Mini Bionix II 858, MTS Systems Corp., USA).

Results and conclusion: Peak injection forces increased successively with the injected cement volume (16.5±12.7N at 1ml to 70.82±21.14N at 6ml). Bone mineral density (93.1±31.6 mg HA/cm3) did not correlate significantly with the injection forces. In 4 specimens the cement reached endplate-to-endplate connection after an average of 5.0±1.2ml of injected volume. Cement leakage occurred in all VB before endplate-to-endplate connection was established. Axial stiffness (3897±1331 N/mm) correlated with the FE models (R2 = 0.778). Finally, as long as no endplate-to-endplate connection was established by the cement, the mean of the top 100 VMS increased to a higher level in the VB. On the contrary, if endplate-to-endplate connection existed, the mean of the top 100 VMS decreased around the cement cloud which links the two endplates.

Our time-lapsed CT approach visualises precisely the cement flow distribution within the bone structure at each incremental injection step. Combined with microFE modeling, this approach allows assessment of the mechanical properties of the augmented bone during injection. In addition, partial cement filling would result in an increased risk of failure of the trabecular bone adjacent to the cement cloud. This is in agreement with previous results.