Use of co-registered high-resolution computed tomography scans before and after screw insertion as a novel technique for bone mineral density determination along screw trajectory

Abstract Introduction Bone mineral density (BMD) is an important factor in the examination of the performance of bone instrumentation both in and ex vivo , and until now, there has not existed a reliable technique for determining BMD at the precise location of such hardware. This paper describes such a technique, using cadaveric human sacra as a model. Methods Nine fresh-frozen sacra had solid and hollow titanium screws placed into the S1 pedicles from a posterior approach. High-resolution micro-computed tomography (CT) was performed on each specimen before and after screw placement. All images were reconstructed with an isotropic spatial resolution of 308 μm, reoriented, and the pre-screw and post-screw scans were registered and transformed using a six-degree rigid-body transformation matrix. Once registered, two points, corresponding to the center of the screw at the cortex and at the screw tip, were determined in each scan. These points were used to generate cylindrical regions of interest (ROI) with the same trajectory and dimensions as the screw. BMD measurements were obtained within each of the ROI in the pre-screw scan. To examine the effect of artefact on BMD measurements around the titanium screws, annular ROI of 1 mm thickness were created expanding from the surface of the screws, and BMD was measured within each in both the pre- and post-screw scans. Results The registration process was accurate to 190 μm, with a precision of 189 μm and error in BMD measurement of ± 2% in repeated scans. BMD values in the cylindrical ROI corresponding to screw trajectories were not statistically different from side to side of each specimen ( p = 0.23). Metal artefact created significant differences in BMD values ( p = 0.001) and followed an exponential decay curve as distance from the screws increased, approaching a low value of approximately 20 mg HA cm − 3 , but not disappearing completely. Summary CT in the presence of metal creates artefact, making measured BMD values near implants unreliable. This technique is accurate for determination of BMD, non-destructive, and eliminates the problem of this metal artefact through the use of co-registered scans. This technique has applications both in vitro and in vivo.