To determine whether measuring pedicle size on CT is accurate and reproducible using the WEBPACS ruler tool

Radiological analysis.

A human cadaveric spine along with 5 geometrical shapes were scanned using a multislice spiral CT scanner with 1mm cuts. The objects and the pedicle diameters for lumbar and thoracic vertebrae in the axial plane were measured independently using the WEBPACS ruler tool by 2 observers (to the nearest 0.1mm). The geometrical shapes and pedicle size on the skeleton were then measured using Vernier callipers by an independent third observer. All measurements were repeated a week later.

Reproducibility of the measurements was assessed using Bland and Altman plots. Accuracy was assessed using the Vernier calliper measurements as the gold standard and comparing the plots.

Perfect reproducibility was achieved when measuring the geometric objects with the Vernier callipers. The error of the measurement associated when measuring the pedicles was 0.5mm. The error of the measurement for the geometric objects for observers 1 and 2 was 0.5 and 0.6mm respectively, and for the pedicles it was 1.0 and 0.6mm respectively.

The WEBPACS ruler on a CT scan is accurate to within 0.5-0.6mm of the true size of an object. The error for pedicle measurements is marginally higher (0.6-1.0mm) and this may reflect the fact that they are ill defined geometric shapes. Measuring pedicle size on CT for surgical planning may have implications for small pedicles when sizing them up for a good screw.

Ethics approval None Interest Statement None