The accuracy of pedicle screw placement is essential for successful spinal reconstructive surgery. The authors of several previous studies have described the use of image-based navigational templates for pedicle screw placement. These are designed based on a pre-operative computed tomographic (CT) image that fits into a unique position on an individual's bone, and holes are carefully designed to guide the drill or the pedicle probe through a pre-planned trajectory. The current study was conducted to optimise navigational template design and establish its designing method for safe and accurate pedicle screw placement.

Thin-section CT scans were obtained from 10 spine surgery patients including 7 patients with adolescent idiopathic scoliosis (AIS) and three with thoracic ossification of the posterior longitudinal ligament (OPLL). The CT image data were transferred to the commercially available image-processing software and were used to reconstruct a three-dimensional (3D) model of the bony structures and plan pedicle screw placement. These data were transferred to the 3D-CAD software for the design of the template. Care was taken in designing the template so that the best intraoperative handling would be achieved by choosing several round contact surfaces on the visualised posterior vertebral bony structure, such as transverse process, spinous process and lamina. These contact surfaces and holes to guide the drill or the pedicle probe were then connected by a curved pipe. STL format files for the bony models with planned pedicle screw holes and individual templates were prepared for rapid prototype fabrication of the physical models. The bony models were made using gypsum-based 3D printer and individual templates were fabricated by a selective laser melting machine using commercially pure titanium powder. Pedicle screw trajectory of the bony model, adaptation and stability of the template on the bony model, and screw hole orientation of the template were evaluated using physical models. Custom-made titanium templates with adequate adaptation and stability in addition to proper orientation of the screw holes were sterilised by autoclave and evaluated during surgery.

During segmentation, reproducibility of transverse and spinous processes were inferior to the lamina and considered inadequate to select as contact surfaces. A template design with more bone contact area might enhance the stability of the template on the bone but it is susceptible to intervening soft tissue and geometric inaccuracy of the template. In the bony model evaluation, the stability and adaptation of the templates were sufficient with few small round contact surfaces on each lamina; thus, a large contact surface was not necessary. In clinical patients, proper fit for positioning the template was easily found manually during the operation and 141/142 screws were inserted accurately with 1 insignificant pedicle wall breach in AIS patient.

This study provides a useful design concept for the development and introduction of custom-fit navigational template for placing pedicle screws easily and safely.