A new 3D approach to evaluate facial profile changes following BSSO

Latent disentanglement in mesh variational autoencoders improves the diagnosis of craniofacial syndromes and aids surgical planning

BACKGROUND AND OBJECTIVE

The use of deep learning to undertake shape analysis of the complexities of the human head holds great promise. However, there have traditionally been a number of barriers to accurate modelling, especially when operating on both a global and local level.

METHODS

In this work, we will discuss the application of the Swap Disentangled Variational Autoencoder (SD-VAE) with relevance to Crouzon, Apert and Muenke syndromes. The model is trained on a dataset of 3D meshes of healthy and syndromic patients which was increased in size with a novel data augmentation technique based on spectral interpolation. Thanks to its semantically meaningful and disentangled latent representation, SD-VAE is used to analyse and generate head shapes while considering the influence of different anatomical sub-units.

RESULTS

Although syndrome classification is performed on the entire mesh, it is also possible, for the first time, to analyse the influence of each region of the head on the syndromic phenotype. By manipulating specific parameters of the generative model, and producing procedure-specific new shapes, it is also possible to approximate the outcome of a range of craniofacial surgical procedures.

CONCLUSION

This work opens new avenues to advance diagnosis, aids surgical planning and allows for the objective evaluation of surgical outcomes. Our code is available at github.com/simofoti/CraniofacialSD-VAE.

Three-Dimensional Geometric Morphometry of Facial Soft Tissue Changes After Bilateral Sagittal Split Ramus Osteotomy

ABSTRACT

This study aimed to evaluate the performance of geometric morphometry (GM) to assess the changes in facial soft tissue after orthognathic surgery. Subjects were 27 patients (skeletal class III) who underwent bilateral sagittal split ramus osteotomy and 27 volunteers as a control group. Computed tomography images of each patient were obtained before surgery (T0) and 6 months after surgery (T1). Computed tomography images of 27 volunteers (skeletal class I) were also obtained as a control group. Using a three-dimensional (3D) modeling software, 3D models were created and exported to a 3D surface analyzing software for geometric morphometry and principal component (PC) analysis. Significant differences in facial soft tissue were found in the first and second of 15 PC. The first PC represented variation in the lower facial height, and the second PC represented variation in the anterior-posterior position of the chin. Comparing the pre- and post-operative images, they illustrated that lower facial height was decreased, and the chin and lower lip moved posteriorly. Geometric morphometry showed to be a successful tool to isolate surgery-related changes from interindividual morphological variations.

Revision of 116 orthognathic surgery patients operated on with the high-oblique sagittal osteotomy (HOSO): a retrospective case series (PROCESS-compliant article)

BACKGROUND

The high-oblique sagittal osteotomy (HOSO) is an alternative to a bilateral sagittal split osteotomy (BSSO). Due to its novelty, there are no long-term studies which have focused on describing the incidence and type of complications encountered in the post-operative follow-up. The aim of this retrospective study is to analyze patients operated on with this surgical technique and the post-operative complications encountered.

PATIENT AND METHODS

The electronic medical records of all patients treated with orthognathic surgery at the Department of Oral, Maxillofacial and Facial Plastic Surgery, University Hospital Frankfurt, Goethe University, Frankfurt, Germany, between the years 2009 and 2016 were retrospectively reviewed.

RESULTS

A total of 116 patients fulfilled the inclusion criteria. The cases operated on with the standard osteosynthesis (X, Y, and straight) showed a complication rate of 36.37% (n = 4/11). The cases operated on with the HOSO-dedicated plates (HOSO-DP) showed, in total, a complication rate of 6.67% (n = 7/105). The most common post-operative complication resulting from both fixation methods was a reduction in mouth opening and TMJ pain for 4.3%. During the first years of performing the surgery (2009-211), a variety of standard plates had material failure causing non-union or pseudarthrosis. No cases of material failure were observed in the cases operated on with the HOSO-DP. The statistical results showed a highly significant dependence of a reduction in OP-time over the years, when the HOSO was performed without additional procedures (R² > 0.83, P < 0.0015).

CONCLUSION

The rate of complications in the HOSO were shown to be comparable to the rate of complications from the BSSO reported in the literature. Moreover, the use of the ramus dedicated plate appears to provide enough stability to the bone segments, making the surgery safer.

CLINICAL RELEVANCE

The HOSO needs to be considered by surgeons as an alternative to BSSO. Once the use of the HOSO-DP was established, the rate of complications and the operation time reduced considerably.

Nasolabial shape and aesthetics in unilateral cleft lip and palate: an analysis of nasolabial shape using a mean 3D facial template

The aim of this study was to determine the amount of deviation in nasolabial shape in patients with a cleft compared with an average non-cleft face, and to assess whether this difference is related to nasolabial aesthetics. Three-dimensional stereophotogrammetric images of 60 patients with a unilateral cleft were used. To quantify shape differences, four average non-cleft faces were constructed from stereophotogrammetric images of 141 girls and 60 boys. Three-dimensional shape differences were calculated between superimposed cleft faces and the average non-cleft face for the same sex and age group. Nasolabial aesthetics were rated with the modified Asher-McDade Aesthetic Index using a visual analogue scale (VAS). Mean VAS scores ranged from 51.44 to 60.21 for clefts, with lower aesthetic ratings associated with increasing cleft severity. Shape differences were found between cleft faces and the average non-cleft face. No relationship was found for the VAS, age, and sex, except that a lower VAS was related to a higher nose and lip distance between the superimposed cleft and average non-cleft faces for nasal profile (P= 0.02), but the explained variance was low (R²=0.066). In conclusion, except for nasal profile, nasolabial aesthetics were not influenced by the extent of shape differences from the average non-cleft face.

A machine learning framework for automated diagnosis and computer-assisted planning in plastic and reconstructive surgery

Current computational tools for planning and simulation in plastic and reconstructive surgery lack sufficient precision and are time-consuming, thus resulting in limited adoption. Although computer-assisted surgical planning systems help to improve clinical outcomes, shorten operation time and reduce cost, they are often too complex and require extensive manual input, which ultimately limits their use in doctor-patient communication and clinical decision making. Here, we present the first large-scale clinical 3D morphable model, a machine-learning-based framework involving supervised learning for diagnostics, risk stratification, and treatment simulation. The model, trained and validated with 4,261 faces of healthy volunteers and orthognathic (jaw) surgery patients, diagnoses patients with 95.5% sensitivity and 95.2% specificity, and simulates surgical outcomes with a mean accuracy of 1.1 ± 0.3 mm. We demonstrate how this model could fully-automatically aid diagnosis and provide patient-specific treatment plans from a 3D scan alone, to help efficient clinical decision making and improve clinical understanding of face shape as a marker for primary and secondary surgery.

Esthetic evaluation of facial cheek volume: A study using 3D stereophotogrammetry

OBJECTIVES

To investigate the influence of cheek volume on facial esthetics judged by orthodontists and non-specialists.

MATERIALS AND METHODS

A 25-year-old female's natural and smiling face was captured by 3D stereophotogrammetry. Cheek volume of the 3D image was altered to different degrees three-dimensionally. For the natural and smiling face, seven groups of facial images were created: decreased grade I/II/III, increased grade I/II/III, and the original one. Thirty orthodontists and 30 nonspecialists were invited to perform esthetic evaluation of the original and transformed images using a questionnaire. Data were calculated with one-way analysis of variance (least significant difference test) and independent samples t test.

RESULTS

Compared to nonspecialists, orthodontists gave lower esthetic scores to the decreased grade III facial images (maximum deformation degree: 7.500 mm and 7.327 mm in natural and smiling face-oriented image groups, respectively). The decreased grade III facial images also received the highest age ranks. However, the increased grade III facial images received the lowest scores and highest age ranks from nonspecialists (maximum deformation degree: 6.994 mm and 5.300 mm in natural and smiling face-oriented image groups, respectively) ( P < .01).

CONCLUSIONS

Orthodontists and nonspecialists showed different esthetic evaluation of varied cheek volume. The influence of cheek volume in orthodontic diagnostic analysis needs further consideration.

Three-Dimensional Imaging of the Face: A Comparison Between Three Different Imaging Modalities

BACKGROUND

Three-dimensional (3D) imaging of the face is being used extensively in medicine for clinical decision making, surgical planning, and research. Nowadays, several companies are offering a broad range of 3D imaging systems, varying in price, method, and mobility. However, most planning and evaluation methods are created and validated solely with one imaging system. Therefore, it is important to analyze possible differences in the 3D surface reconstruction between different systems.

OBJECTIVES

The objective of this study was to analyze differences in the 3D surface reconstruction between three systems: 3dMDface system, Vectra XT, and Artec Eva.

METHODS

Three-dimensional images of the face were acquired from 15 healthy patients with each imaging system. Reproducibility of each device was calculated and a comparison of the Vectra XT and Artec Eva with the 3dMDface was made.

RESULTS

All 3D imaging devices showed high reproducibility, with a mean difference of 0.18 ± 0.15 mm (3dMDface system), 0.15 ± 0.15 mm (Vectra XT), and 0.26 ± 0.24 mm (Artec Eva). No significant difference in reproducibility was found between the Vectra XT and 3dMDface, while a significant difference was found between 3dMDface and Artec Eva, and between Vectra XT and Artec Eva. The mean difference between 3dMDface and Vectra XT was 0.32 ± 0.26 mm. The mean difference between 3dMDface and Artec Eva was 0.44 ± 1.09 mm.

CONCLUSIONS

All three imaging devices showed high reproducibility and accuracy. Although the Artec Eva showed a significant lower reproducibility, the difference found was not clinically relevant. Therefore, using these different systems alongside each other in clinical and research settings is possible.

LEVEL OF EVIDENCE 3

Three-dimensional virtual simulation of alar width changes following bimaxillary osteotomies

The aim of this study was to evaluate the accuracy of three-dimensional (3D) soft tissue simulation of nose width changes following bimaxillary osteotomies and to identify patient- and surgery-related factors that may affect the accuracy of simulation. Sixty patients (mean age 26 years) who underwent bimaxillary osteotomies participated in this study. Cone beam computed tomography scans were acquired preoperatively and at 1-year postoperative. The 3D hard and soft tissue rendered preoperative and postoperative virtual head models were superimposed, after which the maxilla and mandible were segmented and aligned to the postoperative position. The postoperative changes in alar width were simulated using a mass tensor model (MTM)-based algorithm and compared with the postoperative outcome. 3D cephalometric analyses were used to quantify the simulation error. The postoperative alar width was increased by 1.6±1.1mm and the mean error between the 3D simulation and the actual postoperative alar width was 1.0±0.9mm. The predictability was not correlated to factors such as age, sex, alar cinch suture, VY closure, maxillary advancement, or a history of surgically assisted rapid maxillary expansion. The MTM-based simulation model of postoperative alar width change was found to be reasonably accurate, although there is room for further improvement.