3D-printed upper cervical models: accuracy validation and feasibility study into anterior occiput-to-axis screw fixation on them

Abstract

Listen

Translate article

Objective To validate the accuracy of 3D-printed upper cervical models and investigate the feasibility of use of the models in anterior occiput-to-axis screw fixation, in an attempt to provide a protocol of pre-operative plan for surgeons. Methods Forty-five adult atlantoaxial CT scans were obtained, imported into Mimics software for 3D reconstruction, successively imported into 3D printer to print the 3D models. Fourteen parameters were measured on both imaging system and 3D-printed models to validate the accuracy of 3D-printed models. Thirty upper cervical CT data were obtained and imported into Mimics software for 3D reconstruction. Cylinders in 1.75 mm radius were drawn to simulate the trajectory of anterior occiput-to-axis screw fixation. Anteroposterior view of the minimum lateral angle (α1) and maximum lateral angle (α2) and lateral view of the minimum posterior angle (β1) and maximum posterior angle (β2) were measured. Mean value of α1 and α2 was calculated as α3 and mean value of β1 and β2 as β3. Meanwhile, the 3D models were printed, and an angle guide device was used to introduce the anterior occiput-to-axis screws into the 3D models in reference to the angles of α3 and β3. Anteroposterior view of lateral angle (α4) and lateral view of posterior angle (β4) were measured. Differences in α3 vs. α4 and β3 vs. β4 were compared. Results All above 14 parameters did not differ significantly between radiographic images and 3D-printed models (P>0.05). Intraclass correlation coefficient (ICC) values of 13 parameters were >0.800. On the 3D digital models, the α3 was (12.6±3.7)° (left) and (12.0±4.2)° (right), and the β3 was (23.9±4.8)° (left) and (23.4±4.9)° (right). On the 3D-printed models, the α4 was (12.0±4.1)° (left) and (12.3±4.1)° (right), and β4 was (23.4±4.2)° (left) and (22.8±4.4)° (right). There were no significant differences in both comparisons of α3 vs. α4 and β3 vs. β4 (P>0.05). Conclusions The 3D printing technique enables accurate fabrication of upper cervical spine. Aided by the 3D reconstruction with screw trajectory simulation and angle guide device, the anterior occiput-to-axis screws are introduced on the 3D-printed models successfully. Key words: Imaging, three-dimensional; Spine; Computer-assisted design