Deep Learning–Based Facial and Skeletal Transformations for Surgical Planning

ObjectiveThe use of virtual surgical planning (VSP) in orthognathic surgery is increasing. The aim of this study was to investigate which type of dentofacial deformity makes the surgeon choose VSP and if VSP influences the choice of surgical techniques. Furthermore, we sought to evaluate if use of VSP affects surgical time and blood loss during surgery.Material and methodsPatients having undergone orthognathic surgery and registered in the Swedish quality registry for orthognathic surgery (NROK) between 2018 and 2019 were eligible for inclusion in this study. An evaluation of the usage of VSP was performed together with an analysis of operation time and bleeding during surgery. A total of 862 patients were evaluated.ResultsOut of the 862 patients who had orthognathic surgery, 224 were VSP cases. Patients diagnosed with maxillary retrusion (p = 0.0024), maxillary vertical hyperplasia (p < 0.001), mandibular protrusion (p = 0.0031), laterogenia (p < 0.001) or craniofacial deformity (p = 0.0033) were more often digitally planned. Patients who underwent VSP and bimaxillary surgery showed a tendency of reduced operating time andless bleeding during surgery; however, the results were inconclusive.ConclusionThe present study shows that VSP was used more frequently in patients with maxillary retrusion, maxillary vertical hyperplasia, mandibular protrusion, laterogenia, and craniofacial deformity in comparison with other diagnoses. Alongside the findings that VSP cases had more diagnoses per patient and that a larger number of surgical procedures were performed in this group, this indicates a preference for using digital planning in more complex orthognathic cases.

Background: Digital workflows have transformed orthognathic surgery by enabling three-dimensional virtual surgical planning (VSP) and, more recently, artificial-intelligence-based predictive modeling. While skeletal accuracy has improved substantially, the clinical value of these tools increasingly depends on their ability to predict soft-tissue outcomes and support complex treatment decisions. Methods: This narrative review synthesizes evidence from 20 peer-reviewed studies, including systematic reviews, clinical cohorts, and machine-learning validations, to evaluate (i) the accuracy of digital patient modeling and VSP transfer, (ii) the performance of predictive models for postoperative outcomes and diagnostic classification, and (iii) associated ethical and governance considerations. Results: Automated craniofacial and dental modeling achieves submillimetric to low-millimetric accuracy, with landmark localization errors typically ≤2 mm and dental landmark errors near 0.4 mm. Clinical cohorts demonstrate that VSP transfers to surgery with mean translational deviations well below 2 mm in all planes. Predictive models estimate soft-tissue and aesthetic outcomes within clinically meaningful error margins and classify the need for orthognathic surgery with AUC values up to ~0.9 in high-risk malocclusion groups. Conclusions: Digital planning in orthognathic surgery has evolved into a precision-oriented, data-driven framework. When combined with predictive modeling and appropriate ethical governance, these technologies can enhance accuracy, personalize treatment strategies, and reduce diagnostic uncertainty, particularly in borderline and asymmetric cases.

PURPOSE: Virtual surgical planning (VSP) is the gold standard for preoperative treatment planning for orthognathic surgery.1 Prior to the implementation of VSP, traditional orthognathic preoperative preparation required hours of labor and utilization of a dental laboratory. This often limited orthognathic surgery to providers with dental training. The combination of VSP and 3-dimensional printing has allowed orthognathic surgeons the ability to more accurately and efficiently create treatment plans for these cases.2 This technology has also allowed practitioners such as plastic surgeons to begin to perform orthognathic procedures. This review serves to illustrate how plastic surgeons use VSP and highlight its impact on surgical outcomes and research output within orthognathic surgery. METHODS: A literature search was conducted in the PubMed database to identify all articles published before January 2020 that report perioperative use of VSP technology in orthognathic surgeries. Only articles that were written in English, involved operations on live human patients, and used VSP to plan or assist with orthognathic surgery were included in the analysis. All articles that fit inclusion criteria were reviewed to determine the year of publication, type of orthognathic surgery performed, how VSP was used, and the primary outcomes measured. RESULTS: A total of 419 publications regarding VSP in orthognathic surgery resulted from initial search. Of them, 244 publications were authored by oral and maxillofacial surgeons, 28 publications by plastic surgeons, and 34 by surgeons with training in both fields. TEMPORAL TREND OF PUBLICATIONS: Prior to the late 1990s, VSP in orthognathic surgery publications was primarily authored by oral and maxillofacial surgeons. Plastic surgeons first began publishing on VSP technology for orthognathic surgery in 1998. Since then, there has been a rising trend in plastic surgery usage of VSP paralleled by increasing VSP publications in the field, with 2 publications before the year 2010 (7.2%), 9 from 2010 to 2015 (32%), and 18 between 2015 and 2019 (64%). CATEGORIES OF PUBLICATIONS: Most publications from oral and maxillofacial trained surgeons highlight optimal positioning of craniofacial bones, anticipated stability and functionality of occlusion, and the interplay between orthodontic treatments and orthognathic surgery. In contrast, plastic surgeon authored papers often discuss postoperative facial aesthetics by ensuring facial symmetry and treatment especially in the cases of congenital craniofacial abnormalities like cleft lip/palate, Treacher-Collins, and hemifacial microsomia. CONCLUSIONS: VSP is an increasingly utilized tool that reduces the workflow without compromising accuracy and outcomes in orthognathic surgery. This is the first paper to create a temporal relationship between a new technology and the emergence of plastic surgeons into a field typically predominated by oral and maxillofacial surgeons. As plastic surgeons become more familiar with this technology, they have produced literature that provides alternative perspectives on orthognathic surgery that will push the field forward. REFERENCES: 1. Langdon J, Patel M, Brennan P. Operative Oral and Maxillofacial Surgery. 2nd ed. CRC Press; 2010. 2. Baker SB, Goldstein JA, Seruya M. Outcomes in computer-assisted surgical simulation for orthognathic surgery. J Craniofac Surg. 2012;23:509–513.

To determine if patient outcome variables differ between conventional and virtual surgical planning of orthognathic surgery for class III asymmetry. This retrospective case-control study examined 95 patients with class III asymmetry who had been consecutively treated with at least a Le Fort I osteotomy and a bilateral sagittal split osteotomy with a surgery-first approach. Two groups were examined: 51 patients treated with conventional surgical planning and 44 with virtual surgical planning. After treatment, quantitative assessment was determined with measurements of midline symmetry, contour symmetry, and overall facial symmetry using standardized frontal photographs. Subject assessments were analyzed with questionnaires regarding self-perception of overall appearance, satisfaction with appearance, and quality of life. Conventional and virtual surgical planning resulted in significant improvements in outcomes for all patients. However, facial midline and overall facial symmetry were significantly greater for the virtual compared with the conventional group. There were no significant differences in subjective measures of appearance, satisfaction with appearance, and quality of life for patients treated with conventional or virtual surgical planning; measures were high for both groups. Conventional and virtual surgical planning of surgery-first bimaxillary orthognathic surgery resulted in quantitative and qualitative improvements in facial symmetry. Although patient satisfaction was similar for both approaches, virtual surgical planning was superior to conventional surgical planning for the improvement of midline and overall asymmetry. Improvements with virtual surgical planning in facial midline, facial contour, and overall facial symmetry are as good as or better than conventional surgical planning.

Three-dimensional accuracy of virtual planning in orthognathic surgery

This study aimed to compare the clinical accuracy of maxillary positioning in bimaxillary orthognathic surgery based on the method of virtual surgical planning (VSP) implementation, specifically contrasting the use of the wafer-only versus the customized guide-plate system. Thirty-four patients who underwent bimaxillary orthognathic surgery were retrospectively analyzed and divided into 2 groups: a control group (wafer-only, n=17) and a study group (wafer+guide+plate, n=17). Surgical accuracy was quantitatively evaluated by comparing the 3-dimensional coordinates of 5 maxillary landmarks between the virtual surgical plan (Tv) and postoperative cone-beam computed tomography (T1) using dedicated 3D analysis software. In addition, postoperative skeletal stability was evaluated by comparing T1 and T2. The study group showed significantly better accuracy in transverse (0.36mm versus 0.86mm) and anterior vertical dimensions (0.51mm versus 1.45mm). Both groups experienced challenges with posterior vertical positioning and sagittal accuracy. Six-month stability analysis revealed no significant differences between groups, with mean absolute changes of 0.39-1.41mm (study) and 0.44-1.39mm (control). The guide-plate system offers a clinically meaningful improvement in maxillary positioning accuracy over the wafer-only approach, particularly in complex cases requiring high precision. Both methods demonstrated comparable 6-month stability, indicating that the enhanced initial accuracy does not compromise postoperative stability.

Virtual planning in orthognathic surgery

Orthognathic surgery is a procedure used to correct facial deformities and jaw bone misalignment. The use of technology, specifically virtual surgical planning (VSP), has become increasingly prevalent in preoperative planning for orthognathic surgery. High-resolution computed tomography (CT) imaging has enabled computer-aided modelling. Artificial intelligence (AI) implementation has transformed orthognathic surgery methodology. This article highlights the latest market trends and modern-day advancements in the field, including the conventional and surgery first approach for orthognathic surgery. The use of computer-aided surgical simulation (CASS) in VSP for orthognathic surgery was studied. The different software used for orthognathic surgical planning and the detailed protocol followed during the surgery, including the preoperative procedure were discussed along with utilisation of 3-dimension cone-beam computed tomography (3D CBCT) images for surgical planning. The implementation of VSP with CASS had significantly enhanced the accuracy and efficiency of orthognathic surgery for dentofacial deformity correction. The use of technology allowed improved preoperative planning, resulting in better outcomes for patients. The study of different software for orthognathic surgical planning and the protocol followed during surgery has provided valuable insight into the surgery. The continued advancement of technology in orthognathic surgery is promising for the field and for the patients.

The success of virtual surgical planning (VSP) in bimaxillary orthognathic surgery relies on the precise translation of digital plans into operative outcomes. Although computer-assisted design and 3D-printed surgical guides are increasingly adopted to improve surgical accuracy, clinical data on the performance of systems employing Mimics software-particularly concerning long-term postoperative stability-remain scarce. Twelve consecutive patients undergoing bimaxillary surgery between July 2023 and September 2024 were prospectively enrolled. Preoperative cone-beam computed tomography (CBCT) and dental models were integrated into Mimics software for multimodal registration, 3D reconstruction, and VSP. Surgical procedures included Le Fort I osteotomy, bilateral sagittal split osteotomy (BSSO), and genioplasty. A fully customized 3D-printed surgical guide system, including osteotomy guides, repositioning guides, and occlusal splints was employed. Postoperative CBCT scans were acquired at 10days (T1), 1month (T2), 3months (T3), and 6months (T4). CBCT scans from T1 and T4 were analyzed in 3-Matic to quantify 3D deviation. Clinical success was defined as linear deviations ≤ 1.5mm and angular deviations ≤ 2°. Longitudinal comparisons of CBCT datasets across all postoperative intervals (T1-T4) were performed to assess stability. Statistical analysis was conducted using paired t-tests in SPSS 26.0 (α= 0.05). Mean linear deviation was 1.34 ± 0.28mm at T1 and 1.17 ± 0.22mm at T4, while angular deviation was 1.95° ± 0.41°and 1.81° ± 0.35°, respectively. All values were within the predefined clinical thresholds. Maxillary landmarks (ANS, A-point, dental cusps) remained stable throughout the observation period (P > 0.05), whereas mandibular landmarks (B-point, pogonion, menton, dental cusps) demonstrated statistically significant progressive stabilization (P < 0.05). Directional analysis revealed the highest accuracy in the horizontal plane (0.58 ± 0.16mm) and the greatest deviations in the sagittal direction (1.71 ± 0.61mm at T4). The 3D-printed digital guide system achieved clinically acceptable accuracy and demonstrated favorable postoperative stability. Maxillary skeletal landmarks exhibited superior precision compared to dental references. Notably, mandibular structures showed significant positional improvement over time, suggesting distinct healing dynamics between the jaws. All observed deviations fell within published tolerances, validating the system's reliability for clinical application in orthognathic surgery. This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

The primary objective of this study was to evaluate how the utilization of virtual surgical planning (VSP) and the epidemiological profile of patients undergoing orthognathic surgery (OGS) have changed in the past decade. The records of patients who had undergone orthognathic surgery at a national dental hospital were reviewed. Trends in VSP, epidemiological data, presentation of dentofacial deformity, and management details were recorded. A total of 1184 patients were included in this study. The majority of the patients seeking treatment in this dental hospital were young Chinese adults with dentofacial deformities requiring bimaxillary surgeries. Most patients presented with a skeletal Class III pattern (79.0%), and asymmetry was diagnosed in 80.8% of all cases. There was an initial slow pick-up rate for VSP, but this rapidly increased to a high adoption rate of 98.7-100% between 2019 and 2021. Together with an increasing body of evidence suggesting greater accuracy in VSP, utilization in this technology can be enhanced with greater familiarity with the technology and improvements in the VSP services.

Virtual Surgical Planning in Orthognathic Surgery in Cleft Patients

Orthognathic surgery is applied to restore esthetical facial profile and functional occlusion for patients with dentofacial deformity. Virtual surgical planning (VSP) is indispensable for precise and individualized treatment. Manually designing osteotomy planes is time-consuming and highly experience-dependent. This study aimed to develop and validate an automatic osteotomy plane design method based on deep learning. Methods: A deep learning model, Ortho-OPD (orthognathic osteotomy planes de signer), was proposed, consisting of a segmentation network and the random sample consensus (RANSAC) algorithm. The segmentation network was based on a convolutional neural network (CNN), orthognathic segmenting the craniomaxillofacial (CMF) CT data. Osteotomy planes were then defined by the RANSAC algorithm. Ortho-OPD was trained on 71 samples and tested on 31 cases. The performance was evaluated quantitatively and qualitatively. Results: Ortho-OPD functioned smoothly, and all cases were successfully performed. The 3D boundary-sensitive loss was employed to optimize precision. Evaluation metrics included accuracy and clinical efficiency. The mean dice similarity coefficient (DSC) was 0.920.032 in CMF seg mentation. Ortho-OPD showcased excellent productivity, taking an average of about 9 seconds to complete virtual bimaxillary osteotomy compared to manual work. The angular errors between the predicted planes and ground truth planes, plus the shortest distance from the neural tube or the adjacent apical points to predicted planes, were examined, indicating no significant difference and reliability for preserving vital anatomical structures. Overall, the automatic osteotomy plane design from raw CT data was realized using Ortho-OPD, composed of CNN and RANSAC, providing an efficient and ideal alternative in orthognathic osteotomy planning.

Orthognathic surgery (OGS) is a transformative procedure that enhances function and aesthetics. Recent advancements in virtual surgical planning (VSP) of OGS, particularly with 3-dimensional (3D) simulation and intraoral scanning, have improved surgical accuracy and postoperative stability. However, complex cases, such as edentulous jaws, present significant challenges. These cases may lead to increased complications and difficulties in surgical planning. This report introduces an innovative VSP method that incorporates dentures into 3D simulation for challenging cases of edentulous and asymmetric jaws. The composition of dentures into VSP is achieved by performing intraoral scans while the patient is wearing the dentures. Skeletal evaluations are based on bone-derived reference points (Sella, Nasion, Point A, Point B) independent of teeth. Postoperative evaluations using cone beam computed tomography showed a close alignment between the planning of VSP and actual outcomes, validating the accuracy of the VSP method that includes dentures. The differences in displacement of point A and point B between the planned and postoperative results were each <2mm in 3 axes. ANB was markedly improved compared with preoperative. The advancement at point A was slightly insufficient due to the influence of the pharyngeal flap, leading to smaller SNA and ANB than planned. Root-mean-square deviation and color map were also <2mm. This study confirms that integrating dentures into VSP significantly enhances the precision and effectiveness of the surgical procedure, ultimately improving mastication and facial aesthetics in complex craniofacial cases.

Facial gender confirmation surgery (FGCS) is a series of procedures which seek to harmonize a patient's face with his/her self-image and gender identity. Originally described in San Francisco in the 1980s, FGCS has evolved to encompass all elements of the craniofacial skeleton and facial soft tissue. This field in plastic and reconstructive surgery has quickly gained more attention in the past decade due to the pioneering work of groups around the world along with increased social acceptance and medical care of the transgender community. This narrative review focuses on the upper third of the face. Key differences in the forehead and the hairline of cis men and women are discussed which inform pharmacologic and surgical interventions. Hairline modifying therapies including pharmacotherapy and hair transplantation are explained. Virtual surgical planning (VSP), a tool broadly used in surgical fields, has a special role in FGCS and we offer advice in using VSP when addressing the frontal sinus. Use of VSP allows the surgeon to provide reproducible and accurate results. We then discuss the history of the frontal sinus setback and offer our algorithmic approach to recontouring the forehead with detailed description of the operative steps and decision making. Finally, postoperative care and complications considered.

Background: Orthognathic surgery has undergone a significant technological evolution. Digital tools such as virtual surgical planning (VSP), three-dimensional (3D) printing, and computer-aided design/computer-aided manufacturing (CAD/CAM) have become central to the specialty, promising greater precision and predictability. Objective: To synthesize the scientific evidence on new techniques in orthognathic surgery, evaluating their methodologies, clinical outcomes, accuracy, and limitations. Methods: An integrative literature review was conducted in the PubMed/MEDLINE, ScienceDirect, and PMC databases. The search included articles published in the last 20 years using descriptors such as "orthognathic surgery," "virtual surgical planning," "3D printing," "computer-aided surgery," and "surgery-first approach." A total of 22 articles, including reviews, prospective studies, and systematic reviews, were selected. The analysis focused on the description of new techniques, clinical outcomes, and methodological quality, with a qualitative assessment of the level of evidence. Results: The main innovations identified were VSP, 3D-printed surgical guides and custom plates, the surgery-first approach, piezosurgery, and intraoperative navigation. The evidence indicates that these technologies increase surgical accuracy, reduce operative time, and allow for predictable results. VSP shows mean linear deviations of approximately 1.0-1.5 mm. The surgery-first approach reduces total treatment time. However, limitations such as the learning curve, high cost, and challenges in soft tissue prediction persist. Conclusion: New techniques in orthognathic surgery, especially those based on digital workflows, represent a significant advance over conventional methods, offering greater accuracy and efficiency. Despite the promising results, there is a need for more randomized clinical trials to establish a higher level of evidence and standardize protocols.