Accuracy between virtual surgical planning and actual outcomes in orthognathic surgery by iterative closest point algorithm and color maps: A retrospective cohort study

Abstract

Background To evaluate the accuracy between actual outcomes and virtual surgical planning (VSP) in orthognathic surgery regarding the use of three-dimensional (3D) surface models for registration using iterative closest point (ICP) algorithm and generated color maps. Material and Methods Construction of planning and postoperative 3D models in STL files format (M0 and M1, respectively) from CBCT of 25 subjects who had been submitted to bimaxillary orthognathic surgery was performed. M0 and M1 were sent to Geomagic software in semi-automatic alignment surface mesh order of M0 and M1 for registration using ICP algorithm to calculate mean deviation (MD, MD+, MD-, SD) and root mean square (RMS – 3D Error). Color maps were generated to assess qualitative congruence between M0 and M1. From deviation analysis, 3D Error was defined as accuracy measurement. To assess the reproducibility, the workflow was performed by two evaluators multiple times. t-tests were used to assess whether all means of MD, MD+, MD-, SD and 3D Error values would be ≤ - 2 mm and ≥ 2 mm. Results High intra and inter evaluators correlation were found, supporting the reproducibility of the workflow. t-tests proved that all MDs and 3D Error values were > - 2 mm and < 2 mm. Conclusions 3D error mean was within the standards of clinical success lower than 2 mm. ICP algorithm provided a reproducible method of alignment between 3D models and generated color maps to evaluate 3D congruence but did not answer all methodological parameters regarding the assessment of accuracy in orthognathic surgery. Key words: Accuracy, Cone-Beam computed tomography, CAD-CAM, Orthognathic surgery, three-dimensional imaging.

Accuracy of splint vs splintless technique for virtually planned orthognathic surgery: A voxel-based three-dimensional analysis

“…Afterwards, an algorithm is used to calculate the distance between different surface regions of the 3D model. The result can be presented as a color-coded map where different colors show the distance between the surfaces, with green color usually showing the lack of distance, meaning zero error of accuracy 8,49,32 .…”

In Vivo Accuracy of a New Digital Planning System in Terms of Jaw Relation, Extent of Surgical Movements and the Hierarchy of Stability in Orthognathic Surgery

“…For comparison, the root mean square distances (RMS) commonly used in other studies were also calculated. RMS is the square root of the average of squared movement errors [ 9 ] from the transformation matrix of pVision3D. All statistical analyses were performed at a 5% significance level using IBM ® SPSS ® Statistics 27 (IBM Corp., Armonk, NY, USA).…”

“…Despite the heterogeneity of the available studies, all authors agree on one point, which is that the clinical error size should not exceed 2 mm [ 5 , 6 , 7 , 8 , 9 ]. Another common feature is that the differences in jaw positions are increasingly analyzed in terms of translation and rotation of the three spatial axes, with rotations described as yaw (superoinferior axis), roll (anteroposterior axis), and pitch (mediolateral axis).…”

Virtual Surgical Planning: Modeling from the Present to the Future

“…The reference point cloud is then iteratively rotated such that both point clouds align optimally. ICP has been found to be highly accurate for medical applications [ 48 ].…”