Photogrammetric Analysis in Orthognathic Surgery

div epub:type=”chapter” role=”doc-chapter”>

© Springer Nature Singapore Pte Ltd. 2021

J.-W. Choi, J. Y. LeeThe Surgery-First Orthognathic

10. Three-Dimensional Photogrammetric Analysis in Orthognathic Surgery

Jong-Woo Choi1   and Jang Yeol Lee2  

Department of Plastic Surgery, Asan Medical Center, Seoul, Korea (Republic of)

SmileAgain Orthodontic Center, Seoul, Korea (Republic of)
Jong-Woo Choi (Corresponding author)
Jang Yeol Lee

3D camera3D photogrammetryStructured lightPhotogrammetry3D simulationSoft tissue simulationOrthognathic surgeryFrontal soft tissue

10.1 Introduction

10.1.1 Two-Dimensional (2D) Versus Three-Dimensional (3D) Cameras

Two-dimensional photographs have been the standard for planning and evaluating facial aesthetic outcomes for a long time. However, technological advances have made the 3D camera a reality. Different from the conventional 2D photograph, 3D photogrammetry allows the measurement of exact values for various facial soft tissue landmarks; even precise measurements of areas and volumes are possible. I believe that this technology represents a paradigm shift for all medical fields dealing with the face and is something that I have anticipated for a long time.

With recent advances in technology, 3D photogrammetry allows objective and reliable data acquisition with fewer errors. In addition, it facilitates preoperative planning and the simulation of postoperative outcomes.

There are two types of 3D cameras available on the market. I will explain the theoretical differences between the structured light and the stereo methods (Fig. 10.1).

Fig. 10.1

Presently, there are two types of three-dimensional (3D) cameras available on the market. The structured light-type consists of a single projector and a single camera. This system provides highly accurate 3D mesh data but has difficulty capturing a movable object. The stereophotogrammetry type consists of two cameras. This system is able to capture movable objects accurately but the accuracy of the 3D mesh data is reduced because it is unable to find all corresponding points if points with characteristic features are absent, like in patients with smooth skin

The structured light type consists of a single projector and a single camera. This system is able to capture 3D mesh data accurately but has difficulty capturing movable objects. Conversely, the stereotype system consists of two cameras and is able to capture moveable objects; however, the accuracy of its 3D mesh data is reduced because it is unable to locate all the corresponding points on a structure if these points do not possess adequate characteristics, such as on a patient with a smooth face.

When we compare these systems in practice, the structured light type requires the use of an industrial camera to capture multiple images rapidly. Therefore, the texture quality is satisfactory, 3D mesh precision is high, and the speed of the 3D image processing is high.

As mentioned, above, the stereo system has a weakness in that it cannot find corresponding points on patients with smooth skin. Therefore, it requires supplementation with a high-resolution digital single-lens reflex camera to produce an image with very high texture quality and satisfactory 3D mesh precision, but the speed of the 3D image processing is moderate (Fig. 10.2).

Fig. 10.2

While the structured light-type three-dimensional camera provides an image of satisfactory texture quality, the three-dimensional photogrammetric camera produces an image of high texture quality

The price of the industrial cameras used in structured light photogrammetry is falling and the resolution is improving. Thus, the texture quality is getting better. While the stereotype systems are less likely to improve further because the texture quality is already high, the structured light-type systems have the possibility of continuous improvements in texture quality and mesh precision.

While the structured light-type 3D camera provides images with satisfactory texture quality, 3D photogrammetric cameras require 3D images of high texture quality. The structured light-type 3D camera can render an image in less than 25 s whereas the stereotype 3D camera requires more than 2 min to render an image (Fig. 10.3).

Fig. 10.3

Comparing 3D imaging processing speed, The structured light-type three-dimensional camera can process an image in less than 25 s, whereas the stereotype camera requires more than 2 min

However, why are we skeptical of this new technology in real practice? There have been some obstacles to the broad adoption of this technology. The resolution of the 3D camera is not the same as that of a standard digital single-lens reflex camera and the image processing time of 3D images is too long. Additionally, the absence of user-friendly software has contributed to the failure of this technology to be commonly adopted. Finally, how can we use the preoperative simulation images in real surgery?

In this chapter, I will explain how I have been using 3D photogrammetry in my practice.

10.1.2 3D Photogrammetry in Orthognathic Surgery

Orthognathic surgery, used to correct dentofacial anomalies, has two key goals: correction of malocclusion and good postoperative facial aesthetics. Orthognathic surgery can alter facial soft tissue contours by changing skeletal tissue and, therefore, can be used to create a more attractive face. However, the effect of skeletal surgery on soft tissue profiles is not easy to predict [1].

Although 2D cephalometry has been used for soft tissue analysis, it can only assess the lateral profile; it cannot be used for anteroposterior frontal analyses, especially for facial soft tissues (Fig. 10.4) Thus, 2D cephalometry emphasizes hard tissue landmarks because their reproducibility is better than those for soft tissue landmarks [24]. Other methods of soft tissue analysis include anthropometry, photography, stereophotogrammetry, photocephalometry, and Moire topography [57], but all of these approaches have major limitations, such as the time required, poor reproducibility, or possible errors in translation [5].

Fig. 10.4

Traditional two-dimensional lateral cephalometric analysis. Preoperative and postoperative views

These shortcomings have resulted in an increase in the use of 3D imaging techniques. For example, the visible facial soft tissue volume changes observed with an optical 3D sensor have been evaluated after midface distraction or LeFort I maxillary advancement [810]. In addition, 3D computed tomography, with volume rendering, has been used for soft tissue analysis following orthognathic surgery [1114], but these techniques have serial measurement limitations imposed by potential radiation hazards and their poorer resolution than conventional photogrammetry [15, 16] (Fig. 10.5).

Fig. 10.5

Traditional two-dimensional photogrammetry versus three-dimensional photogrammetry. Three-dimensional photogrammetry provides precise objective measurements of various soft tissue landmarks

In 2007, my practice introduced its first commercial 3D camera for analyzing soft tissue landmarks. Due to its high resolution, similar to that of conventional photogrammetry, it yielded accurate and reproducible data. The 3D camera allows for frontal view analyses not previously possible using cephalometry. Moreover, 3D camera images can be rotated, translated, and enlarged, providing realistic simulations of the effects of the planned orthodontic and surgical treatment. In contrast, the results of conventional photogrammetry cannot be similarly manipulated, thus preventing serial anthropometric analyses, despite using multilateral pictures of natural head positions (Fig. 10.6).

Fig. 10.6

Three-dimensional camera based on stereophotogrammetry. (a) A three-dimensional stereophotogrammetric camera system (second generation, Vectra, Canfield, USA). (b) The process of axis calibration. During this procedure, correct frontal views can be obtained. The yaw, pitch, and roll of three-dimensional images led to these calibrations

Although 3D camera systems have been shown to yield reliable and reproducible results [1719], their usefulness after orthognathic surgery has not been assessed. Therefore, we tested the ability of a 3D camera system to analyze soft tissue landmarks in patients with skeletal Class III dentofacial deformities who underwent two-jaw rotations with maxillary posterior impaction, but without maxillary advancement. This is a surgical approach that results in better aesthetic outcomes than conventional methods, including maxillary advancement and mandibular setback, in Asian skeletal Class III patients [20]. We utilized the new 3D camera system to quantitatively analyze soft tissue changes, with a focus on facial proportions, including vertical and horizontal dimensions, mid- and lower-facial surface areas, and frontal soft tissue landmarks (Fig. 10.7).

Fig. 10.7

Pre- and postoperative three-dimensional camera images based on the white structured light method. Although the three-dimensional photographs were not taken with natural head positions, they could be corrected, unlike two-dimensional photographs

10.2 Methods

From now on, I will introduce one of my investigations associated with the use of 3D camera in terms of the evaluation of the soft tissue changes before and after the orthognathic surgery. The study involved 25 consecutive patients with skeletal Class III dentofacial deformities. Between January 2008 and December 2009, these patients underwent two-jaw rotational setback surgery, using posterior maxillary impaction without maxillary advancement, at the Seoul Asan Medical Center. The patients were all Asians and had a mean age of 22 years (range, 17–32 years). Patients who underwent conventional maxillary advancement and mandibular setback and those who underwent anterior maxillary vertical reduction were excluded, as were patients with syndromic or disease-initiated dentofacial anomalies, such as secondary cleft-related dentofacial deformities [2].

10.2.1 Imaging Methods

A 3D stereophotogrammetric camera and software system was used for frontal soft tissue analysis (Vectra, Canfield Scientific, Parsippany-Troy Hills, NJ, USA; Fig. 10.8a). The camera setup consisted of three digital cameras, a flash, and control bodies. Prior to use, the camera was calibrated to define a 3D coordinate system for the photographs. The 3D photographs were taken with the patients maintaining natural head position; each patient was looking into a mirror and had a natural facial expression (Fig. 10.8b). To test the reliability of our 3D photogrammetric tool, precision and accuracy testing was performed. The test involved 10 normal adults (2 males, 8 females) and three observers. Six images were taken of each subject and repeated twice for each observer; seven linear measurements and four angular measurements were completed for each 3D image. The precision testing revealed that the mean absolute difference of the linear measurements was within 1.2 mm, which is considered very precise compared with other measurement tools. A Kruskal–Wallis test failed to demonstrate any statistically significant differences among the observers or calibrations. The accuracy testing showed a 1.4-mm difference between measurements. The Pearson’s correlation coefficients were so high that the measured 3D values were regarded as having very acceptable accuracy and precision. The 3D photogrammetry results were very similar to other reports using different measurement tools.

Fig. 10.8

Precision measurement validation of the three-dimensional camera. Three-dimensional camera measurement errors were investigated using a skull model to compare the differences between the actual and the three-dimensional measurements. For example, actual skull model lengths of 20 mm, 30 mm, and 50 mm were calculated to be 20.164 mm, 30.241 mm. and 50.567 mm, respectively, on the three-dimensional images

10.2.2 Landmark Identification

Prior to landmark placements, the axes were calibrated by yawing, rolling, and pitching of the 3D images (Fig. 10.9). Two observers each indicated the landmarks on the facial soft tissue images, twice. The soft tissue landmarks were similar to those previously described, but were modified to fit the 3D analysis, according to previous reports [17, 18]. These landmarks included the trichion, nasion, nasal tip, subnasale, stomion, and menton for vertical measurements as well as the bizygomatic points, bigonial points, medial and lateral canthus, and oral commissures. To position these landmarks correctly, we enlarged and/or rotated the 3D images, while correlating the axes with those previously identified.

Fig. 10.9

Three-dimensional soft tissue landmarks. These landmarks included the trichion, nasion, nasal tip, subnasale, stomion, and menton for vertical measurements, and the bizygomatic points, bigonial points, medial and lateral canthus, and oral commissures

10.2.3 Measurement of Actual Distances and Surface Areas on the 3D Images

We used the software program to measure the actual distances between soft tissue landmarks before surgery and at least 6 months postoperatively (Table 10.1). Vertical parameters included the lengths of the upper, middle (from the nasion to the subnasale), and lower (from the subnasale to the chin) thirds of the face, and the lengths of the upper and lower lips. The transverse parameters included the bizygomatic and bigonial widths; the bizygomatic widths were used to assess intra- and inter-rater errors because these values were not changed by orthognathic surgery unless a zygomatic reduction procedure was performed. The vertical maxillary length was defined as extending from the subnasale to the stomion and the vertical mandibular length as extending from the stomion to the menton. The nose profile analysis included measurements of the alar width, nasal tip, and columellar height; the lip analysis included measurements of their horizontal and vertical lengths. The surface areas of the middle and lower thirds of the face were measured using a 3D software program (Vectra, Canfield). To determine facial asymmetry, the distances between the medial and lateral canthus and the oral commissure were compared, pre- and postoperatively. Finally, the cheek soft tissue convexity was measured. Paired t-tests and the Wilcoxon-signed rank test were used for statistical analyses. If a normality test, like the Kolmogorov–Smirnova test, was plausible, a parametric paired t-test was performed. If a Kolmogorov–Smirnova test was not plausible, or if the number of samples was less than 20, a nonparametric Wilcoxon-signed rank test was performed. All statistical analyses were performed using the SPSS statistical package (version 18.0, SPSS, Chicago, IL, USA).

Table 10.1

Definitions of facial soft tissue landmarks




Facial proportion

Upper 1/3

Trichion ~ Upper margin of eyebrow

Mid 1/3

Upper margin of eyebrow ~ Subnasale

Lower 1/3

Subnasale ~ Menton

Mx & Mn

Mx. Height

Subnasale ~ Stomion

Mandible height

Stomion ~ Menton

Transverse width

Zygomatic width

The length between the most lateral points in zygomatic arch (suborbitale level)

Bigonial width

The length between the most lateral points in mandible angle (stomion level)


Alar width

The length between the alar

Nasion—nasal tip

Nasion ~ Nasal tip

Nasal tip—subnasale

Nasal tip ~ Subnasale


Upper vermilion area

Redline in midpoint of upper lip ~ Stomion

Lower vermilion area

Stomion ~ Redline of lower lip

Lip length

The length between oral commissures

Upper vermilion height

The upper margin of upper red vermilion ~ stomion in midline

Lower vermilion height

Stomion ~ the lower margin of lower red vermilion in midline

Surface areas

Upper facial area

The facial surface areas from subbrow to stomion level

Lower facial area

The facial surface areas from Stomion to menton level


Medial canthus- Rt. Oral commissure

Endpoint of medial canthus—most latera point of oral commissure

Medial canthus—Lt. oral commissure

Endpoint of medial canthus—most latera point of oral commissure

Lateral canthus—Rt. Oral commissure

Endpoint of lateral canthus—most latera point of oral commissure

Lateral canthus—Lt. Oral commissure

Endpoint of lateral canthus—most latera point of oral commissure

The landmarks were determined based on previous reports and were slightly modified for the three-dimensional analysis

10.3 Results

10.3.1 Cephalometric Changes

I present my research data about the application of 3D photogrammetric analysis in clock wise rotational orthognathic surgery based on PNS impaction. This result will help you understand the impact of the 3D camera in orthognathic surgery. It allows me not only to analyze the lateral soft tissue profile, but also to analyze the frontal soft tissue profile objectively. I analyzed the surgical outcomes of 25 patients who underwent clock wise rotational orthognathic surgery based on PNS impaction ranging from 3mm to 8 mm. [21]. Jaw rotational orthognathic surgery resulted in satisfactory postoperative results in all 25 patients. The average mandibular setback was 10.7 mm (range, 5–17 mm), and there was no evidence of major upper airway hindrance. In 7 cases, reduction or advancing genioplasty procedures were performed, with an average reduction of 2.31 mm and an average advancement of 2.02 mm. The average maxillary posterior impaction was 4.5 mm. The pivot points of the two-jaw rotations were mostly A points. None of the patients underwent maxillary advancement. Thus, the average SNA increased from 77.4° to 77.8° although the average SNB decreased from 89.2° to 81.1°. The average occlusal plane increased from 8.7° to 11.4° because of the two-jaw rotations with posterior maxillary impaction (Fig. 4a–d).

10.3.2 Vertical Facial Proportions (Table 10.2)

The length of the upper third of the face was unchanged by jaw surgery and the middle third decreased a small amount, from 58.8 mm to 57.8 mm (p = 0.059), probably because the anterior maxillary height decreased slightly during the procedures; only the posterior maxillary height was decreased using the posterior impaction procedure for clockwise rotations of the maxilla. In contrast, the length of the lower third of the face decreased significantly, from 70.4 mm to 68.2 mm (p = 0.0006). We found that the maxillary vertical height increased significantly, from 22.7 mm to 23.7 mm (p = 0.023) while the mandibular vertical height decreased significantly, from 47.9 mm to 44.2 mm (p < 0.0001). The ratio of the maxillary to mandibular vertical length changed from a preoperative ratio of 1:2.11 to 1:1.86, postoperatively.

Mar 5, 2021 | Posted by in Orthodontics | Comments Off on Photogrammetric Analysis in Orthognathic Surgery
Premium Wordpress Themes by UFO Themes