Use of Hausdorff Distance and Computer Modelling to Evaluate Virtual Surgical Plans with Three-Dimensional Printed Guides against Freehand Techniques for Navicular Bone Repair in Equine Orthopaedics.

Abstract

The aim of this study was to evaluate the surgical execution of a virtual surgical plan (VSP) with three-dimensional (3D) guides against a freehand approach in the equine navicular bone using an automated in silico computer analysis technique. Eight pairs of cadaveric forelimb specimens of adult horses were used in an ex vivo experimental study design with in silico modelling. Limbs received either a 3.5 mm cortical screw according to a VSP or using an aiming device. Using computed tomography and computer segmentation, a comparison was made between the executed screw and the planned screw using the Hausdorff distance (HD). Navicular bone mean HD registration error was -0.06 ± 0.29 mm. The VSP with 3D printing demonstrated significantly superior accuracy with a mean deviation of 1.19 ± 0.42 mm compared with aiming device group (3.53 ± 1.24 mm, p = 0.0018). The VSP group was 5.0 times more likely to result in a mean aberration of less than 1.0 mm (95% confidence interval, 0.62-33.4). A 3.5 mm screw with an optimal entry point can have a maximum deviation angle of 3.23 ± 0.07, 2.70 ± 0.06 and 2.37 ± 0.10 degrees in a proximal, dorsal and palmar direction respectively, prior to violating one of the cortical surfaces. Procedures performed using the 3D guides have a high degree of accuracy, with minimal mean deviations (<1 mm and <1 degree) of a VSP compared with those using the conventional aiming device. The use of VSP and the HD for evaluation of orthopaedic surgeries and outcome measures shows promise for simplifying and improving surgical accuracy.