Abstract
Purpose
Radiostereometric Analysis (RSA) is a well developed imaging technique used to estimate implant fixation of orthopaedic implants in randomized clinical trials. The precision of RSA depends on a number of factors including image quality related to the individual modality properties. This study assesses the precision of RSA with a novel Digital Radiography (DR) system compared to a CR imaging system using different imaging techniques. Additionally, the study assesses the precision of locating beads embedded in a modified spine pedicle screw.
Method
A modified titanium spinal pedicle screw 4.5 mm diameter, 35 mm length, marked with two 1.0 mm tantalum beads, one inside the head and one near the screw tip was inserted into a bovine tibia segment. Six additional 1.0 mm tantalum beads were inserted into the bone segment – superiorly, distally and adjacent to the pedicle screw. The phantom was placed on a standard clinical diagnostic imaging bed above a custom RSA carbon fiber calibration cage (Halifax Biomedical Inc.). A pair of DR or CR imaging plates were placed below the calibration cage and irradiated 15 times at 100, 125 kV at 2.5 mAs. To determine precision, the standard deviation of 3D vector distances between beads was determined using RSA for each of the different imaging parameters.
Results
The precision error (PE), defined as the standard deviation of the 3D Bone Marker marker locations for CR is 35.5 m for 100kV at 0.5 mAs setting and 42.2, 39.4, and 26.7 m for the 2.5 mAs at 100, 125, and 150 kV settings respectively. However, for DR, the PE is 27.5 m for 100kV at 0.5 mAs setting and 25.7, 25.1, and 20.1 m for the 2.5 mAs at 100, 125, and 150 kV settings.
The PE for Screw Marker 3D locations, for CR is 38.2 m for the 100kV at 0.5 mAs setting and 55.2, 47.3, and 37.1 m for the 2.5 mAs at 100, 125, and 150 kV settings respectively. However for DR, the PE is 40.3 m for 100kV at 0.5 mAs setting and 33.2, 24.9, and 17.0 m for the 2.5 mAs at 100, 125, and 150 kV settings respectively. The PE for all Bone Marker and Screw Marker 3D locations were significantly lower (P<0.05) for the DR technology than the CR technology except at the 100kV at 0.5 mAs exposure of the Screw Marker, P = 0.589.
Conclusion
The PE decreases for increasing kV, especially in the case of screw markers where the error goes from 33 micron (100kV) to 17 micron (150 kV). Increasing the mAs reduces the error for the DR, but increases the error for CR. Increasing the kV did not significantly influence the precision in measuring bead locations in bone. For embedded tantalum beads within a titanium pedicle screw, imaging at higher kV values with the described DR imaging system did allow more precise localization. The current phantom design is basic in nature and does not account for any soft tissue scatter. However, initial results indicate a gain in precision when using DR compared to CR imaging equipment for RSA analysis.
