Abstract
The aim of this report was to present a reproducible workflow for integrating three-dimensional (3D) facial scans into computed tomography-based virtual surgical planning of orthognathic surgery, thereby enriching the virtual twin with clinically relevant soft-tissue information. This report describes the sequential stages of the workflow: 1) standardised preoperative imaging including computed tomography, intraoral scans (IOS), and three facial scans (rest, smile, lip-retracted); 2) preprocessing of facial scans in 3-Matic to remove dental structures and generate soft-tissue stereolithography (STL) models; 3) virtual twin construction in Materialise Enlight v7.0 by integrating digital imaging and communications in medicine skeletal data, IOS, and surface-based facial models; 4) application of repositioning tools and STL import functions to integrate multiple facial scans; 5) visualisation and analysis of natural head position, digital smile design, and comparison of pre- and postoperative soft-tissue predictions; and 6) finalisation of the case with the generation of intermediate and final splints and export of all planning data. The integration of facial scanning into virtual surgical planning is feasible and may be clinically relevant. While promising, further validation is required to confirm accuracy and reproducibility in routine clinical practice, particularly with respect to natural head position, digital smile design, and soft-tissue prediction. This workflow highlights the potential of recent software advances and facial scanning technologies in orthognathic planning.
Introduction
Orthognathic surgery refers to surgical procedures aimed at repositioning the jaws. This repositioning seeks to restore a new equilibrium among several functional units of the face, including the jaws, teeth, temporomandibular joints, and soft tissues such as the tongue and lips. Several concepts for attaining facial harmony through orthognathic skeletal adjustments have been put forward, with influential contributions by Arnett et al, , Reyneke, and Mommaerts.
Prior to any repositioning, a comprehensive facial analysis is mandatory. Over the past decades, this analysis has evolved from two-dimensional (2D) to three-dimensional (3D) imaging. Traditional planar radiographs such as panoramic radiographs and lateral cephalograms have gradually been replaced by cone-beam computed tomography (CBCT) and computed tomography (CT). With the advent of virtual surgical planning (VSP), as described by Swennen et al, the era of manual tracings on acetate has come to an end. , Conventional facebows have been substituted by the pursuit of the natural head position (NHP) as a more reproducible and clinically relevant reference. ,,
Today, VSP relies on the creation of a virtual twin, which is a digital replica of the patient that allows for simulated surgical movements and prediction of the postoperative outcome. These virtual twins are constructed using both skeletal and soft tissue data derived from digital imaging and communications in medicine (DICOM) files and are now increasingly augmented with intraoral scans (IOS). This results in a composite model, which merges DICOM-based imaging with the high-resolution detail of the dentition from IOS, thereby improving the accuracy of planning. ,,,,
More recently, facial scanning technologies, which generate lifelike 3D surface renderings of the patient’s face, have been integrated into this workflow. These facial scans produce true-colour, surface-based 3D photographs of the facial soft tissues. Most current systems operate on the principle of optical light reflection. ,,,
With the development of new software platforms such as Materialise Enlight v7.0, it is now possible to import various stereolithographic (STL) files, including facial scans, into the standard clinical VSP workflow. These innovations allow the virtual twin to be enriched with detailed soft tissue data, enhancing the accuracy of digital smile design, NHP alignment, and prediction of postoperative outcomes. ,,
The aim of this study is to explore and demonstrate the potential of integrating 3D facial scans into next-generation software platforms for orthognathic surgical planning.
Material and methods
Preoperative workup and imaging protocol
Each orthognathic case is subjected to a standardised preoperative diagnostic workup. In our institution, this includes the collection of both 2D and 3D data. , As part of the 2D documentation, panoramic radiographs, and lateral cephalograms are obtained, complemented by standardised clinical photographs taken under controlled lighting conditions.
A wax bite registration is taken in the retruded occlusal mandibular (ROM) position with the patient in supine position. This wax bite is retained intraorally during the acquisition of the CT scan, which ensures accurate mandibular positioning during imaging. ,, To guarantee that the resulting dataset is suitable for precise VSP, our centre applies a dedicated CT protocol that adheres to established neuronavigation standards in oral and craniomaxillofacial (OCMF) surgery.
In addition, IOS of the upper and lower dental arches are acquired using a Medit Link 7 scanner. In our workflow, detailed soft tissue morphology is captured, with facial scans obtained using the MetiSmile system (FScan, Shining3D; Facial Analyzer, Shining3D). These scans are performed in a neutral rest position, during a spontaneous full smile, and with a lip retractor in place to fully expose the maxillary anterior teeth. Patients are positioned in NHP, hair is secured away from the face, and extraneous objects such as eyeglasses are removed and hands kept out of the way to minimise artefacts and variability. These facial scans are acquired using a handheld 3D surface scanner (Shining 3D Dental). ,,,
All image data are processed using Materialise Enlight v7.0 and Materialise Mimics 3-Matic v19.0 (Materialise). The integration of DICOM files, intraoral STL data, and surface-based facial scans into a composite virtual twin is conducted within this digital environment. ( Fig. 1 ).
Print screen of the facial scans in in Materialise 3-Matic v.19.0. (A) Neutral rest position. (B) Smile. (C) Retracted smile.
Initial processing of facial scans
In the first step, the facial scans are processed using 3-Matic (Materialise). Utilising the Trim Tool, all digital information related to the dentition is selectively removed. This results in facial models displaying only the soft tissue structures of the patient, with dental anatomy fully excluded ( Fig. 2 ).
Print screen of the facial scan in in Materialise 3-Matic v.19.0. (A) Normal retracted smile. (B) Smile with trim tool in situ. (C) Smile after application of the trim tool without dentition.
The two processed facial scans, facial scan with lips in resting position excluded—corresponding to the spontaneous smile, and lip-retracted condition—are subsequently exported to a dedicated storage folder. These soft tissue models will later be re-imported into Materialise Enlight v7.0 for integration using the STL import button into the VSP workflow. ,,,
Virtual twin construction in Materialise Enlight v7.0
Upon initiating a new case in Materialise Enlight v7.0, the patient’s DICOM data, acquired with the condyles positioned in the ROM position, are imported into the software environment. Through the predefined segmentation workflow, the cranium, mandible, and facial soft tissues are segmented. Subsequently, the IOS are loaded and mounted (stitched) onto the existing segmented mandible and maxilla. This results in a composite models described by Swennen et al, combining the anatomical accuracy of the DICOM-derived bony structures with the high-resolution detail of the dentition from the IOS data. Along with the segmented soft tissue envelope, these models are saved as reference STL files within the platform. ,,
After this step, the workflow proceeds with several preparatory analyses. Bilateral tracing of the inferior alveolar nerve is performed to safeguard anatomical structures during planning. The NHP is then established directly from the facial scans, providing a reproducible and clinically relevant reference orientation. Finally, a cephalometric analysis is carried out to guide skeletal repositioning. The sequence of these steps may vary depending on the specific surgical protocol being applied ( Fig. 3 ).
Establishing natural head position within the virtual surgical planning environment (print screen Materialise Enlight v7.0).
Once the preparatory analyses are complete, the osteotomy phase is initiated. Within the virtual environment, all osteotomy surfaces are carefully reviewed and verified before execution. Standard osteotomies are then planned, including a Le Fort I osteotomy, bilateral sagittal split osteotomy, and genioplasty when indicated. ,, These osteotomies are prepared within the virtual environment to enable precise surgical simulation.
Following this, the final occlusion setup—based on the desired postoperative dental relationship—is imported into the planning system. At this stage, the virtual twin is considered nearly complete and ready for final surgical planning and simulation ( Fig. 4 ).
Print screen of Materialise Enlight v7.0.
Advanced repositioning and STL import functionalities in the clinical workflow
Within the clinical workflow of Materialise Enlight v7.0, a series of advanced functionalities become available during the intermediate ‘Reposition’ phase. Among the most frequently used are the ‘Move,’ ‘Mirror,’ and ‘Align’ tools, which allow for precise manipulation of individual anatomical segments. These functions significantly increase flexibility in treatment planning and are particularly valuable not only in orthognathic surgery but also in the management of complex craniofacial trauma, where accurate repositioning of skeletal fragments is essential. ,,,,
In addition to these manipulation tools, the software provides an ‘STL Import’ function that enables the integration of supplementary digital models into the virtual twin environment. This feature supports the import of additional IOS, either with or without prosthetic restorations, as well as various facial scans representing the patient in different conditions, including neutral rest position, spontaneous smile, and lip-retracted expression.
Once imported, these facial scans are registered to the CT-derived soft-tissue model within Enlight. Alignment is achieved through a combination of initial positional adjustments and surface-based registration using stable anatomical reference regions such as the forehead, nasal bridge, and zygomatic arches. This process ensures a precise match between the surface-based facial scans and the DICOM-based soft-tissue envelope. , ( Fig. 5 ).
