Evaluation of the three-dimensional prediction method in the planning of orthognathic surgery applications

This study aims to evaluate the accuracy of three-dimensional virtual planning software in predicting soft tissue changes in orthognathic surgery applications. For this purpose, this study compares the soft tissue changes predicted by the NemoFAB (Nemotec,Spain) software with the changes observed six months after surgery. This retrospective study included 20 patients (10 female- 10 male) who underwent bimaxillary orthognathic surgery due to Class III skeletal malocclusion between 2020 and 2024. Preoperative (T0) and postoperative (T1, at 6 months) computed tomography (CT) scans were analyzed along with the predicted soft tissue images (TP) obtained from NemoFAB software in Standard Tessellation Language (STL) format. Three-dimensional images were aligned using the voxel-based superimposition method in Geomagic Control X software, and soft tissue changes were evaluated. A total of 3 angular and 12 linear measurements were performed and the data were statistically analyzed. As a result of the analysis, a statistically significant difference was found between Tp and T1 in 3 angular and 4 linear measurements (p<0.001). In the angular measurements, FNA (fronto-nasal angle) and LMA (labiomental angle) prediction values were found to be high, while NLA (nasolabial angle) prediction values were found to be low. In terms of linear measurements, statistically significant differences were observed in upper lip thickness (UL-U1), lower lip thickness (LL-L1), the distance between the hard tissue B point and the soft tissue B point (B-B*), and lower lip length (stmi-ME') (p<0.001). UL-U1 TP values were found to be high (p<0.001), whereas TP values for LL-L1, B-B*, and stmi-ME' were found to be low (p<0.001). The three-dimensional virtual planning software NemoFAB has been found to be an effective method for predicting soft tissue changes after orthognathic surgery. However, the prediction values for the linear measurements UL-U1, LL-L1, B-B*, and stmi-ME' as well as the angular measurements FNA, NLA, and LMA were found to differ from the postoperative outcomes. The findings of the study suggest that the algorithms used in NemoFAB software need further improvement to enhance prediction accuracy.