A Novel Radiologic Assessment of Screw Loosening Focusing on Spatial Position Change of Screws Using an Iterative Closest Point Algorithm with Stereolithography Data: Technical Note

Objective: In thoracolumbar spinal instrumentation surgery, pedicle screw fixation is widely used, whereas screw loosening occurs only occasionally over time. It is common to evaluate screw loosening by the radiographic lucent zone around screws, which can neither evaluate loosening quantitatively nor detect slight screw loosening. In the present technical note, we describe a novel assessment technique of screw loosening by generating 3-dimensional screw images from computed tomography data and superposing them in time series.

Methods: Computed tomography data were exported in digital imaging and communications in medicine dataset and imported to the 3-dimensional computer-aided designing software, by which screws and rods were segmented and outputted in stereolithography (STL) format. The STL files were imported to the software, and registration based on iterative closest point algorithm was performed to assess screw position changes.

Results: Positional changes on STL in time series were classified into 3 types: 1) no position changes existed in the entire system of screws and rods; 2) position changes existed in the entire system of screws and rods, but no position changes were shown when the left-sided and right-sided screws and rod were separately evaluated; and 3) position changes existed in the left and/or right-sided screws and rods even when evaluated separately.

Conclusions: This technique enables the quantitative evaluation of screw loosening and loosening between screws and rods. In conjunction with conventional methods of assessing radiographic lucent zone, we are able to obtain more accurate information regarding screw loosening after spinal instrumentation surgery.