Applications of three-dimensional imaging techniques in craniomaxillofacial surgery: A literature review

Abstract

Three-dimensional (3D) imaging technologies are increasingly used in craniomaxillofacial (CMF) surgery, especially to enable clinicians to get an effective approach and obtain better treatment results during different preoperative and postoperative phases, namely during image acquisition and diagnosis, virtual surgical planning (VSP), actual surgery, and treatment outcome assessment. The article presents an overview of 3D imaging technologies used in the aforementioned phases of the most common CMF surgery. We searched for relevant studies on 3D imaging applications in CMF surgery published over the past 10 years in the PubMed, ProQuest (Medline), Web of Science, Science Direct, Clinical Key, and Embase databases. A total of 2094 articles were found, of which 712 were relevant. An additional 26 manually searched articles were included in the analysis. The findings of the review demonstrated that 3D imaging technology is becoming increasingly popular in clinical practice and an essential tool for plastic surgeons. This review provides information that will help for researchers and clinicians consider the use of 3D imaging techniques in CMF surgery to improve the quality of surgical procedures and achieve satisfactory treatment outcomes.

Introduction

Craniomaxillofacial (CMF) deformities can be congenital or acquired and include dentofacial, congenital, posttumor resection–related, posttraumatic, and temporomandibular joint deformities [[1], [2], [3], [4]]. Surgical correction of CMF deformities is challenging because of the unique and complex nature of the three-dimensional (3D) anatomy of the skull and face, the presence of vital structures, and the variety of deficits. Compared with conventional two-dimensional (2D) imaging, 3D imaging provides a more accurate representation of complex 3D craniofacial anatomical structures [[5], [6], [7]]. In recent years, cost- and dose-effective cone-beam computed tomography (CBCT) has been frequently used for CMF modeling. Additionally, 3D photogrammetry with surface scanning is another rapid, nonradioactive, reliable, and reproducible imaging method suitable for quantitatively analyzing facial changes [[8], [9], [10], [11], [12], [13]]. Moreover, advances in 3D imaging technologies and software programs are valuable in the field of computer-assisted CMF surgery. Various 3D imaging technologies can empower clinicians to make critical decisions preoperatively and help them to achieve favorable results in different phases, including during image acquisition, diagnosis, virtual surgical planning (VSP), actual surgery, and treatment outcome evaluation [[14], [15], [16], [17], [18]].

For image acquisition and diagnosis, 3D images are acquired through CBCT, and images are stored in the digital imaging and communications in medicine (DICOM) format; these images are used to generate initial 3D virtual models. In addition, 3D surface images, such as craniofacial and digital dental images, can be obtained for some patients who require cephalometric analysis for diagnosis [[19], [20], [21], [22], [23], [24]]. Preoperatively, during VSP, computer-aided design and manufacturing (CAD/CAM) technology is used for computer-aided surgical simulation (CASS) or 3D surgical planning (virtual surgery) and intraoperative guided preparation [[25], [26], [27], [28], [29], [30]]. During the actual surgery, CAD/CAM-derived intraoperative guidance tools are usually adopted to accurately perform VSP; intraoperative guidance is valuable for complex CMF surgery [[31], [32], [33], [34], [35], [36], [37], [38], [39], [40]]. During the evaluation of treatment outcomes, the transfer accuracy of the surgical plan and treatment outcomes are postoperatively analyzed using 3D imaging [[41], [42], [43], [44], [45]].

This literature review discusses the current state of various 3D imaging techniques used in the aforementioned phases of the most common CMF surgeries. Physicians can benefit from the findings of this review by implementing them in their clinical practice and achieving favorable treatment outcomes for their patients.

Materials and methods

We searched for relevant studies that were published in English and within a period of 10 years in the following databases: PubMed, ProQuest (Medline), Web of Science, Science Direct, Clinical Key, and Embase. The following search keywords were used: three-dimensional images, three-dimensional imaging, 3D images, 3D imaging, three-dimensional photographs, and 3D facial scan in combination with facial plastic surgery, facial reconstruction, facial surgery, oral and maxilla-facial surgery, cranio-maxillofacial surgery, and orthognathic surgery. In addition, the reference lists of relevant articles have been manually searched to identify other potential studies. We included studies that were published in English after 2013 and in peer-reviewed journals; had full-length text openly available; and examined 3D images. After the removal of duplicate articles, 712 relevant articles were identified and included in the further analysis. The remaining reference articles were appraised to ensure all relevant publications were identified, and an additional 16 articles were hand-searched.

Results

Fig. 1 presents the flowchart for study selection. After full-text analysis, 728 articles meeting the inclusion criteria were included in this review. These articles described the applications of various 3D imaging techniques in CMF surgery. Table 1 summarizes the characteristics of these articles.

Fig. 1.

Flowchart of study selection.

Table 1.

An overview of the characteristics of the reviewed articles (in a number of reviewed articles).

Category	Surgery						Total
	OGS	MR	CS	FF	HFM	Other
Data Acquisition	332	58	44	296	121	106	957
Virtual Surgical Planning (VSP)	216	36	31	148	86	98	615
Actual Surgery	182	29	20	131	62	77	501
Treatment Outcomes Assessment	290	45	11	207	79	80	712
The repeat articles can be categorized into two categories based on text content.

Abbreviations: OGS: Orthognathic Surgery; MR: Microtia Reconstruction, CS: Craniosynostosis Surgery; FF: Facial Fractures; HFM: Hemifacial Microsomia.

Data acquisition

Data acquisition is performed after diagnosing a deformity or problem through clinical evaluation (Fig. 2) and can be divided into three components: 3D image acquisition, replacement of the distorted dentition of the CBCT model, and orientation of the natural head position (NHP).

Fig. 2.

Three components of data acquisition (A) 3D image acquisition (B) replacement of the distorted dentition of the CBCT model (C) orientation of the NHP.

3D image acquisition

For patients with craniofacial abnormalities, 3D imaging modalities include multislice CT, CBCT, magnetic resonance imaging, 3D ultrasound, and 3D photography. Compared with conventional CT, 3D CBCT requires a lower radiation dose and has a higher spatial resolution, shorter scanning time, and lower costs and is thus widely used to obtain volumetric images in the fields of craniofacial, dental, and oral. The key advantages of CBCT are that the reconstruction of CBCT data can be performed at a native resolution by using a personal computer, and the data can be reorientated to reflect true spatial relationships. Moreover, the application-specific software of CBCT can provide an accurate 3D representation of a patient's craniofacial anatomy; the creation of a virtual patient can help clinicians in establishing an appropriate diagnosis, planning treatment, and evaluating treatment outcomes [[46], [47], [48], [49]]. During image acquisition, patients are asked to relax their facial musculature, close their eyes tightly, and place the teeth in habitual occlusion. After CBCT imaging, patient data are stored in DICOM format and are thresholded for segmentation by delineating and identifying anatomical structures of interest in the CBCT images for 3D model reconstruction. The resulting CBCT models are 3D representations of the soft or bone tissues of each patient and provide a range of threshold values (Hounsfield units, HU); the results can be used for diagnosis and treatment planning. Another commonly used 3D imaging modality is 3D photography, which replicates the 3D surface of soft tissue by using highly accurate craniofacial spatial data obtained from stereo surface imaging systems, including 3D digital stereophotogrammetry and 3D laser scanning. Several scanners are used as 3D facial data acquisition devices, functioning as either mobile scanners, such as the Vectra H1 (Canfield Scientific Inc., Parsippany, USA), M4D Scan (Rodin4D, Mérignac, France), and Artec Eva (Artec 3D, Luxembourg), or as stationary scanning devices, such as the 3dMD Face System (3dMD LLC, Atlanta, USA), FaceScan3D (3D-Shape GmbH, Erlangen, Germany), and Vectra M3/XT (Canfield Scientific Inc., Parsippany, USA) [[50], [51], [52], [53], [54]]. Compared with obtaining and using CT data, using the aforementioned devices to capture data for 3D reconstruction has advantages such as minimal invasiveness, no radiation dose requirement, high repeatability, easy reference point determination, rapid image capture, and ease of use; the 3D reconstructed data can even be combined with CT reconstructed osseous volume data to perform reliable soft tissue simulation and prediction and objectively evaluate craniofacial treatment changes [55,56]. During 3D digital stereophotogrammetry (e.g., 3dMDface system and M4D Scan), patients are instructed to maintain their NHP, place their teeth in their habitual occlusion, and relax their facial musculature after swallowing. Intraoral scanners use 3D laser scanning, which has the advantages of reducing the patient's swallowing reflex and working time, eliminating the deformation of impression material or plaster, and being easy to repeat [57]. Intraoral scanners have become an increasingly popular alternative to conventional methods. Moreover, intraoral scanners can provide details that are not obtained from 3D digital stereophotogrammetry, including information on the nostril and the shape of the outer ear (for microtia reconstruction).

Replacement of the distorted dentition of CBCT model

This step consists of initial registration, superimposition, and replacement, then a computerized skull–dental composite model is created. However, this model is typically not necessary for most CMF surgeries. In the presence of metallic restorations, crowns, brackets, and implants, reconstructed CT images can show streaky artifacts that prevent accurate identification of the teeth. Replacing the CBCT scanned dentition with a digital dental model is necessary to determine the optimized interdental relationships. Digital models developed using data obtained from extraoral or intraoral scanners are reliable and present accurate occlusal relationships similar to conventional casts with a high degree of accuracy and reproducibility. Many studies have proposed methods of combining CT models and digital dental models to achieve simultaneous 3D rendering of teeth, skeletal structures, and occlusion. Regardless of the defects in the imaging obtained when using CBCT, fiducial markers such as softened gutta-percha, ceramic spheres, titanium spheres, acrylic, and facebow can be used to replace the distorted dental images [[58], [59], [60]]. Certain methods of voxel- and surface-based superimpositions are superior to landmark-based superimposition in terms of accuracy, precision, reproducibility, and efficiency after clinical validation. These superimposition methods are based on an iterative nearest point algorithm that determines the error by calculating the 3D distance between surface points on the two images and manually selecting the corresponding region [61,62].

Orientation of NHP

The NHP is a standardized and reproducible upright positioning of the head of a standing or sitting patient with eyes looking straight at a mark. The patient's head position during CT should be identical to the NHP to obtain a precise 3D simulation result in CASS, especially for orthognathic surgery (OGS). However, patients have difficulty maintaining the NHP when the head is in a supine position during conventional CT or immobilized by various headrests during CBCT. Various methods have been proposed to determine the orientation of the patient's head during CBCT, including the use of laser scanning, facial markers along the laser line, and digital orientation sensing devices to record and reproduce the NHP. However, these methods are impractical in routine clinical practice. Therefore, reorientation of CBCT images through the use of stable cranial landmarks and reference systems may be a practical alternative. Most studies have proposed unique methods to orient the CBCT model to approximate the NHP (simulated NHP) by adjusting the orientation of the Frankfort horizontal, midsagittal, and transporionic planes to match the axial, sagittal, and coronal planes, respectively. Constructing a 3D anatomical reference system is the most crucial step before adjusting the CBCT/computerized skull–dental composite model reorientation [63,64]. For the landmark-based reference planes, different coordinate systems are provided, and the remaining reference planes were determined based on the standardized reference planes first selected by the investigators. The horizontal plane is determined as the standard plane, followed by the midsagittal plane and coronal plane to form a 3D reference system [65,66].

VSP

VSP includes four steps (Fig. 3): 3D image preprocessing, 3D cephalometric analysis, surgical simulation, and intraoperative guidance design.

Fig. 3.

Four components of VSP (A) 3D image preprocessing (B) 3D cephalometric analysis (C) surgical simulation (D) design and preparation of intraoperative guidance.

3D image preprocessing

This step involves determining the regions of interest (ROIs) from a CBCT image (also known as segmentation) by assigning a label to every voxel in a 3D image. This step is used to visualize the surface representation of the bone anatomy and the soft tissue or facial anatomy. After ROI segmentation, unwanted objects are manually removed to obtain clean objects for surgical simulation [46,67]. For OGS and jaw or mandible angle reduction, coronal CT images are carefully investigated to examine the anatomical location of the inferior alveolar nerve (IAN) from the entry to the mandibular bone (mandibular foramen) to the exit (mental foramen) for the management and prevention of IAN injury during osteotomy [68,69].

3D cephalometric analysis

Cephalometric analysis is often performed to make critical surgical planning decisions. First, certain anatomical and 3D specific cephalometric landmarks are defined. Then, 3D measurements are performed using points, lines, planes, and angles to assess CMF deformities such as facial skeleton asymmetry, occlusion, soft tissue angles, bimaxillary line, maxillary contour angle, mandibular sulcus contour angle and tooth exposure during smile.²⁶ The results of studies obtained using 3D models can improve the understanding of surgeons and patients prior to surgical planning [70,71].

Surgical simulation

(1) OGS: The traditional CASS protocol for OGS is known as hybrid 3D digital planning because dental impressions, dental casts, model surgery, manual occlusal splints, and 2D-based cephalometric norms (normative data) are still used for 3D surgical simulation and planning. A new CASS protocol involves the use of integrated digital impressions, virtual occlusions, digital splints, and 3D norms in full 3D digital planning. Once the skull–dental composite model is generated and the final occlusion is set up, similar to real surgery, the surgeon can perform patient-specific surgical simulation (3D surgical planning) on the composite model according to 3D normative data by using simulation software programs [19,25,[46], [47], [48]]. A study constructed an average 3D skeletofacial model by using data obtained from the normal population as a template for VSP to provide a simple and flexible alternative to conventional 3D cephalometric norm-based methods [72]. The popular commercial software programs used for VSP in OGS include 3D Systems' VSP (3D Systems, Rock Hill, South Carolina, USA), Mimics (Materialise, Leuven, Belgium), SimPlant OMS (Materialise Dental, Leuven, Belgium), ProPlan CMF (ProPlan CMF (Depuy Synthes, Solothurn, Switzerland, and. Materialise, Leuven, Belgium), Maxilim (Medicim, Bruges, Belgium), and Dolphin (Dolphin Imaging, Chatsworth, California, USA). The total airway volume, airway length and mandibular position are also evaluated regularly during simulation [73]. For soft tissue simulation, most of the aforementioned programs are not validated. Thus, mesh deformation methods still need to be improved to obtain desired soft tissue changes and the corresponding bony repositioning plan. (2) Microtia reconstruction: If a patient has a unaffected ear on the unaffected side (unilateral), after 3D image scanning, the normal side is mirrored to the defect side on the basis of the midsagittal plane (MSP). Then, the microtic ear is removed, and the optimal orientation and position of the reconstructed ear are determined on the basis of the frontal view and adjusted until symmetry is achieved; the data are then exported for modification. For bilateral microtia, the ear of one of the patient's parents (possibly uniformly scaled, if required) or an ear template are used as the reference for the reconstructed ear. The 3D imaging measurements generated using the 3dMD Vultus software package (3dMD LLC, Atlanta, GA, USA) have been shown to be accurate and reproducible in the reconstruction of microtia. Additionally, the Geomagic studio software package (3D system, Rock Hill, SC, USA), SimPlant O&O, and ProPlan CMF can also be used [74]. (3) Craniosynostosis surgery: Virtual craniosynostosis surgery includes calculating a preoperative patient-specific normative reference shape, estimating optimal bone fragments, and computing the most appropriate configuration of fragments used for osteotomy and remodeling [75,76]. (4) Facial Fractures: For zygomatic arch fractures, the anatomical region that is unaffected by a deformity can be mirrored to the lesion side and overlaid against the affected anatomical region by using the MSP (optimal symmetry planes, OSPs) to acquire a harmonized and balanced facial structure [37,48,67,77]. (5) Reconstructive surgery for hemifacial microsomia: Conventional treatment for patients with hemifacial microsomia (HFM) involves OGS and distraction osteogenesis of the mandible [78,79]. For mandibular distraction osteogenesis, virtual osteotomy positions and configurations, including linear oblique, inverted L, and multiangular, are customized for each patient. The position and direction of the distractor are repeatedly adjusted until the ideal status is achieved. If required, a finite element method (FEM) analysis of biomechanics is performed to investigate the effect of simulated models [80]. Grafts are used to supplement the reconstruction of CMF deformities involving the temporomandibular joint and ramus [81]. Soft tissue correction is needed after facial bone realignment. Autologous fat grafting techniques are often used because they are versatile, simple, effective, and inexpensive. 3D photographic images of the NHP and habitual dental occlusion are taken prior to fat grafting performed. Virtual surgery and planning are used to estimate the amount and location of fat injection required. Localization of the 3D face image is critical for determining the OSP and anterior–posterior planes on both sides of the face. The location and MSP of the 3D face images are critical, especially for patients with HFM deformities and Romberg disease, where the midline landmarks from the skull base to the facial are oblique and not on the same plane. The MSP influences the required amount and location of fat injections. In such patients, the NHP is used with the MSP across the nasion, and yaw rotation is adjusted as needed. The other side of the face is selected using the MSP and mirrored to the lesioned side to obtain a harmoniously balanced facial contour. The mirrored copy is then registered to the original face. Then, the volume difference (facial defect) between the upper surface shells of the superimposed (combined) images is calculated and the amount of fat required is estimated by adding the possible safe and effective absorption rate to the calculated facial defect volume. Finally, a colormap showing the differences in contour between the superimposed images is generated to guide the fat injection into the region and calculate the amount of fat required for each injection layer. If HFM patients present with malformations involving the pinna and external auditory (EAC) which require ear reconstruction, virtual simulations are performed after the microtia reconstruction procedure described above.

Intraoperative guidance design and preparation

(1) OGS: After the final VSP, the results are transferred to the operation room with the help of intraoperative guiding to ensure the accuracy and efficiency of surgical procedures. Surgical splints, osteotomy guides, repositioning guides, and real-time navigation systems are commonly used intraoperative guidance tools to translate virtual plan results into actual surgery in the OGS. For CAD/CAM surgical splints, virtual surgical planning is an alternative to designing and fabricating surgical splints based on traditional impressions or plaster models. In CASS, digital occlusion is generated by scanning a conventional impression or plaster model and used to design a digital surgical splint. First, the results of surgical simulation are exported to stereolithography (STL) format and imported into CAD software such as Geomagic Wrap (3D System, USA), 3-Matic (Materialise NV, Belgium), or Tizian Creativ RT (Schütz Dental). The splint is then designed digitally by performing Boolean operations and trimming, based on intermediate and final tooth positions, to achieve the actually planned specifications [[82], [83], [84]]. Some specific modifications of the digital occlusal splint can be designed by adding extension bars or targets to form a composite occlusal splint that provides a useful intraoperative guide for positioning the maxillomandibular complex (MMC) during surgery. Finally, digital splint files are obtained and prepared for 3D printing. The accuracy and reliability of the CAD/CAM-generated splint have been shown to reproduce 3D VSP results during real surgery and replace traditional occlusal splints [85]. Recently, a 3D intraoral scanning device was used for electronic occlusal setup (virtual occlusion) and applied for fully digital 3D OGS planning [86]. A standard occlusion setting protocol is based on a normal overjet, overbite, and arch coordination (occlusal symmetry). After setting the virtual occlusion, certain adjustments can be made. For example, in patients with an anterior open bite, the final digital bite can have enough anterior bite and a posterior open bite to facilitate postoperative orthodontic treatment. Once the final virtual occlusion has been confirmed, an occlusal splint can be designed using the process described above. The mandible is superimposed on the lower dentition to form the MMC and surgical simulation is performed by moving the MMC. In addition, osteotomy and repositioning guides are designed and manufactured. Osteotomy guides are used to ensure that osteotomy in LeFort I is placed identically to that in digital planning; this can ensure the accurate placement of the repositioning guide in the desired position in the bone segment [18,[87], [88], [89], [90]]. Another type of repositioning guide is achieved by moving the bone segment without reference to the osteotomy line. A pair of CAD/CAM surface templates is introduced to reduce the effort, operating time, and errors that can occur using traditional approach [91]. The use of spacers is another approach for achieving facial symmetry and harmony in LeFort I when lengthening the vertical dimension of the maxilla in bilateral sagittal split osteotomy (BSSO) after rotating or shifting of the mandible to maintain or increase symmetry or cheek contour correction and in vertical lengthening procedures in genioplasty [92]. Spaces should be maintained as planned during plate and screw fixation. Once the design is finalized, surgical guides and templates (splints) are manufactured using selective laser sintering (additive manufacturing) and polyamide material. All 3D-printed models are sterilized before surgery by using standard autoclave protocols. In CASS, a real-time navigation system is used to determine the final bone position without positioning guidance or as an aid to guide the osteotomy and check the final position of the bone movement [19,39]. The preoperative navigation plan is created preoperatively. First, VSP results and CT DICOM data are imported simultaneously into the navigation planning software program. Pre-defined dental landmarks and fiducial markers (identifiable bones) are represented as reference points on the 3D model, which are used for establishing a link between the patient and the virtual model by using registration. The osteotomy landmarks determined from the VSP results are marked on the 3D model to guide the osteotomy. Finally, intraoperative verification points are identified on the surgical simulation image for guiding the bone movement. (2) Microtia reconstruction: After determining the reconstructed ear, the STL file is imported into commercial software, such as Geomagic Freeform Software (3D Systems, Rock Hill, S.C.) for modifying the extraneous artifacts and subtracting the skin thickness. Finally, the auricular framework beneath the skin envelope (resembling the convolutions of a normal ear) is obtained for manufacture as an intraoperative guide by using a 3D printer with biocompatible material [74,93]. (3) Craniosynostosis surgery: After surgical simulation, the 3D models of customized templates and cutting guides are defined and designed to assist intraoperative bone remodeling. An intraoperative navigation system with image-guided surgery is used when necessary; this system is based on optical tracking for real-time positioning and 3D visualization. (4) Facial fractures: A guide plate design can be used to match the fracture region with the shape and position of a preformed titanium plate; this can help in the preliminary fixation of a fracture and the placement of the guide plate [16,37,84,94]. An intraoperative navigation system allows for real-time confirmation of bone positioning and implant placement, thus solving the visibility problem and shortening surgery time [34,95]. Several types of surgical navigation systems are available, including optical tracking systems, screen-based systems (Columbia scientific SIM/Plant software, Columbia Scientific, USA; VISIT surgical navigation software, Vienna, Austria; optical tracking system microscope, Radionics, Tyco Healthcare Group, USA; Vector vision, Brainlab, Munich, Germany; and Stryker leibinger navigation system, Stryker, Leibinger, Freiburg, Germany), projected image–based systems (surgical microscope navigation system, SMN, Zeiss, Oberkochen, Germany and surgical segment navigator, Carl Zeiss, Germany), electromagnetic tracking systems, heat-mounted display systems (ARTMA, Instatrak, GE HealthCare, Buckinghamshire, UK), and mechanical tracking systems. (5) Reconstructive surgery for HFM: The surgical plan is exported as an STL format file into a CAD program to prepare a surgical guide, which is specific for the mandible.

Actual surgery

For all surgeries, the surgical guides are sterilized before use in the operating room. If required, navigation planning is imported to a commercial navigation system [34,39,49,76]. First, screw retained digital reference frame is used for intraoperative tracking. Then the surface-matching or landmark-based registration is used to obtain the mathematical relationship between the patient's physical spatial coordinates and the CT imaging. Registration errors may be encountered; the surface-matching registration method has been shown that it is not suitable for bimaxillary surgery because the use of nasotracheal intubation (NTI) may result in the skin surface distortion and displacement. Landmark registration has been reported to be a more reliable method to meet the clinical requirement. For fat grafting, micro autologous fat grafting is performed layer by layer using an injection gun that marks the contour lines of each injected layer based on a colored map and pre-calculated required amount [81].

Treatment outcome assessment

The most commonly used methods for postoperative outcome assessment include validating the predictability and accuracy of treatment plan transfer by comparing VSP and postoperative outcomes, assessing postoperative skeletal stability and facial morphology (or comparing the surgical outcome between surgical techniques), and investigating the relationship between facial soft tissue changes and underlying bone movement. (1) Validating the predictability and accuracy of treatment plan transfer by comparing VSP and postoperative results: Two methods are used for validation, one based on color difference metrics and the other on descriptive statistical analysis. CBCT images are acquired in the early postoperative period. For chromatic aberration index, the differences between VSP and postoperative images after initial registration of the cranial base simulation images with the actual postsurgical images is evaluated by using a visual overlapping degree of two surface models. Model fusion tools are commonly applied to the automatically display of the size, direction, and location of discrepancy between the two models; this is accomplished by color-coding the visualization and root-mean-square deviation (RMSD) values [18,20,44,78]. For descriptive statistical analyses, 3D cephalometric analyses are performed using anatomical landmarks, reference planes, and measurements (distance and angle) [33,90]. To determine reliability and accuracy, statistical analyses are conducted to examine whether there are significant difference in various cephalometric parameters between actual and virtual postoperative images. For most measurements, Inter- and intra-examiner reliability is often calculated. (2) Assessing postoperative skeletal stability and facial morphology (or comparing the surgical outcome of surgical techniques): To examine skeletal stability at different postoperative periods, angular changes in pitch, roll, and yaw directions are measured by using a some defined plane of correlation of bone segments [96,97]. Additionally, RMSD values can be used as another comparison index. Except for traditional 3D cephalometric analysis, 3D objective assessment of the facial contour asymmetry outcome is a common method used to evaluate facial morphology. Various methods are used to achieve facial symmetry. Facial asymmetry based on 3D facial landmarks can be quantitatively evaluated by using the reference plane method (asymmetry index and asymmetry rate) and Euclidean distance matrix analysis (constructing the difference matrix of the linear distance ratio between each landmark of the left face and right face and extracting relevant ratios to construct the facial asymmetry index). RMSD values are widely used as indicators of 3D facial asymmetry, which is based on the original mirror image alignment of the left and right sides of the face, with lower RMSD values indicating a more symmetrical face [98]. In recent years, the template mapping strategy has received considerable attention and has been used in facial asymmetry analysis. A 3D contour line–based machine learning method can serve as a general, useful, and automated decision-making tool with human-like efficiency for the objective assessment of facial symmetry [99]. (3) Investigating the relationship between facial soft tissue changes and underlying skeletal movement: To accurately evaluate soft tissue changes associated with bone tissue movements, the first step is to determine the surgically induced changes in the facial images. Pre- and post-operative 3D image models based on the cranial base are stable and unaffected by surgery and are superimposed using the surface registration method. After registration, 3D measurements of pre- and post-operative models of soft- and hard-tissue models can be performed [100]. For 3D FEM analysis, the personalization of the general model based on CBCT data and mathematical transformations is performed using constructed 3D FE models of craniofacial soft and bone tissues. A patient-specific 3D facial FE model including distinct layers of soft tissues can be predicted after surgery. For anatomical landmark–based measurements, reference planes are defined for linear measurements between reference planes and soft tissue points. For region-based measurements, 3D pre- and post-operative models of soft and hard tissues can be used to determine volumetric differences to further measure the mean movement of soft or hard tissue (underlying bone). All the aforementioned approaches require examination of the correlation between hard tissue and soft tissue rate of change to assess the predictive power.

Discussion

This review demonstrates that 3D imaging technology has been extensively used in CMF surgery, especially for image acquisition and diagnosis, VSP, actual surgery, and treatment outcome assessment. These techniques can be extended to related procedures in the field of plastic, oral and maxillofacial, and head/neck surgery. Incorporating 3D images for cephalometric analysis in diagnostic and treatment planning enables clinicians to comprehensively assess maxillofacial structures and surgical planning. 3D VSP enables surgeons to accurately reproduce treatment plans on a computer, perform interactive operations in a real-time 3D virtual environment, and simulate different surgical procedures. Additionally, simulation results can be used to determine optimal outcomes, establish clear and objective treatment plans for the correcting facial deformity, and reduce surgical difficulty through problem detection and plan modification before surgery. Intraoperative guidance tools, such as CAD/CAM surgical guides and real-time navigation systems, can assist surgeons to accurately transfer the results of VSP to actual surgery, simplify the positioning procedure, eliminate intraoperative human errors, reduce operation time, and achieve satisfactory esthetic facial appearance, improve patient care, and reduce costs. Finally, 3D imaging is an essential tool for the objective and quantitative assessment of treatment outcomes, and it is particularly important for clinician in evaluating the accuracy and execution of surgical planning, and assessing treatment outcomes during follow-up.

Conclusions

We believe that 3D imaging will become increasingly popular in clinical practice and an essential tool for the modern plastic surgeons. This review provides information to help researchers and clinicians consider the use of 3D imaging for CMF surgery in the future.

Conflicts of interest

The authors have no financial conflicts of interest.