Quantifying digital volume correlation strain uncertainty at the bone-pedicle screw interface: an ex-vivo zero-strain analysis.
Authors
Affiliations (3)
Affiliations (3)
- Department of Systems Design Engineering, University of Waterloo, Waterloo, Ontario, Canada.
- Department of Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, Ontario, Canada.
- Department of Systems Design Engineering, University of Waterloo, Waterloo, Ontario, Canada. Electronic address: [email protected].
Abstract
Pedicle screws are essential for postoperative construct stability in spinal fusion surgeries. However, screw loosening remains a common clinical complication, often linked to excessive loading that induces damaging trabecular bone strains surrounding the pedicle screw. Toward improving pre-clinical assessment of pedicle screw fixation, this study combines micro-CT imaging, U-Net deep learning segmentation, and digital volume correlation (DVC) techniques to quantify the errors associated with measuring full-field strains at the screw-bone interface under nominal zero-strain conditions. Pedicle screws were surgically implanted into six human lumbar vertebrae specimens. Each specimen underwent micro-CT imaging. For three specimens, paired unloaded scans were acquired and analyzed using a standard DVC uncertainty workflow. To mitigate local bone-metal artifacts, bone segmentation was performed using a U-Net model, and segmentation accuracy was quantified using the Sørensen-Dice similarity coefficient. The U-Net model reproduced manual segmentations with high fidelity (average Dice = 0.94 ± 0.01). DVC accuracy and precision depended on the spatial resolution of strain measurement; at a DVC spatial resolution of 1.8 mm, strain estimates around the screw were reliable with low error (accuracy = 387 ± 110 µε; precision = 222 ± 82 µε). This study establishes the feasibility of a micro-CT-based DVC workflow and characterizes the strain measurement errors within trabecular bone surrounding pedicle screws using a zero-strain analysis. This workflow provides a foundation for future studies involving loaded tests, further optimization of segmentation to suppress artifacts, and validation across larger cohorts and implant designs.