The aim of this study was to develop a reference frame for three dimensional (3D) facial soft tissue growth analysis in children and to determine its reproducibility. Two observers twice placed the reference frame on 39 3D-stereophotogrammetry facial images of children with orofacial clefts and control children. The observers’ performances were analyzed by calculating mean distance, distance variability, and P95 between the same facial surfaces at two different time points. Correlations between observers were analyzed with Pearson’s correlation coefficient. The influence of presence of a cleft, absence of one ear in the photograph, and age on the reproducibility of the reference frame was checked using Student’s t test. Results of intraobserver comparisons showed a mean distance of <0.40 mm, distance variability of <0.51 mm, and P95 of <0.80 mm. For interobserver reliability, the mean distance was <0.52 mm, distance variability was <0.53 mm, and P95 was <1.10 mm. Presence of a cleft, age, and absence of one ear on the 3D photograph did not have a significant influence on the reproducibility of placing the reference frame. The children’s reference frame is a reproducible method to superimpose on 3D soft tissue stereophotogrammetry photographs of growing individuals with and without orofacial clefts.
Medical imaging has advanced from two dimensional (2D) representations to more sophisticated three dimensional (3D) techniques. 3D imaging is of particular value in recording and evaluating the complex morphology of orofacial clefts (OFC), which have a profound effect on facial growth. 3D methods to capture facial images to study growth may be broadly classified as 3D cephalometry, laser scanning, moiré techniques, 3D morphometry, stereophotogrammetry, patterned light techniques, conventional computed tomography (CT) scans, 3D cone beam CT (CBCT) scans, and 3D magnetic resonance images.
Stereophotogrammetry is a noninvasive technique based on the principle of photographing a 3D object with two pairs of identical cameras separated by a known base distance. The result is a stereo pair of photographs of the face taken from two different positions at the same time. These two photo images are combined to form a 3D model. The technique permits acquisition of a true representation of the human face without exposure to ionizing radiation.
A few cross-sectional studies have used stereophotogrammetry to characterize facial soft tissue features in babies. There is little data regarding 3D facial growth for very young Caucasian children in the age range of 0–6 years with or without orofacial clefts. White et al. used stereophotogrammetry to capture images of 41 boys and 42 girls without clefts at approximately 3 months of age. Hood et al. used the same method to characterize the soft tissue features of 23 children with a cleft presurgically, and 21 non-cleft controls at approximately 3 months of age.
For longitudinal studies, facial growth can be evaluated with stereophotogrammetry by superimposing 3D stereophotogrammetry pictures from different time points and measuring changes in 3D facial volume and changes in anatomical landmark positions over time. Standardized positioning of the patient is crucial for objective assessment. In 2D cephalometrics, the cranial base is used as an anatomical structure for superimposition. On a 3D soft-tissue picture there are no so-called stable structures on which superimpositions can be made to measure growth. Swennen et al. developed a Cartesian CB CT-based reference frame that can be used for superimpositions of nongrowing individuals. This reference frame comprises three planes. The ‘horizontal plane’ is the Frankfurt horizontal plane. The ‘vertical plane’ is perpendicular to the horizontal plane and at a fixed distance 12 cm posterior to the point (pupil reconstructed point) located at the midline of the nose at the level of the interpupillary line. The ‘median plane’ is placed perpendicular to the horizontal plane and vertical planes through the midline of the face, but this reference frame cannot be used for superimpositions of growing individuals, because of the fixed distance of the vertical plane 12 cm posterior to the pupil reconstructed point. When superimpositions are made on this frame, facial growth is quantified more in a backward direction due to this fixed distance. Therefore, in the present study, the authors modified the reference frame of Swennen and Schutyser so it can be used to superimpose 3D soft-tissue pictures of growing individuals. This modified reference frame is called the children’s reference frame ( Fig. 1 ). The aim of this study was to determine the intra- and interobserver reproducibility of this reference frame when used for children with and without OFC.
Subjects and methods
The study protocol was approved by the medical ethical commission of the institution in which the study was carried out, and the subjects’ parents gave their informed consent.
A convenience sample of 39 3D photographs was taken from the 3D database of a larger longitudinal study and divided into three groups. Group 1 included 13 Caucasian babies with nonsyndromic complete unilateral cleft lip and palate (UCLP), aged 3 months. Group 2 comprised 16 control Caucasian babies without orofacial cleft, aged 3 months. Group 3 comprised 10 control Caucasian babies without orofacial cleft, aged 12 months. The photographs were randomly numbered 1 through 39. Complete imaging of the face and at least one ear was mandatory to enable setup of the reference frame. On 33 photographs, both ears were fully visible, while the six remaining photographs had only one fully imaged ear.
All the 3D photographs were acquired by a trained photographer using a 3D stereophotogrammetric camera setup and the software program Modular System v1.0 (3dMDface™ System, 3dMD Ltd., Atlanta, GA, USA). The camera setup consisted of two pods, each equipped with three digital cameras and a flash. Prior to its use, the camera was calibrated to define a 3D coordinate system for the 3D photograph, which was referred to as the original 3D coordinate system. To isolate the region of interest on the 3D photographs, the neck and background were trimmed using 3dMDpatient V3.0.1 software. The 3D photograph was exported as a wavefront object file (.obj) and imported into Maxilim ® version 184.108.40.206 (Medicim NV, Mechelen, Belgium).
Children’s reference frame
The cephalometry plug-in tool of Maxilim ® was used to build the children’s reference frame. The canthal line was marked by selecting both exocanthia ( Fig. 2 ). The patient was rotated around this canthal line to the canthal–superaural line ( Fig. 3 ), and the pupil reconstructed point (defined as the point located at the midline of the nose at the level of the interpupillary line) was marked ( Fig. 4 ).
These three steps were used to create the three planes of the Cartesian 3D photograph-based reference frame ( Fig. 1 ). The horizontal plane was placed through the pupil reconstructed point and 7° below the canthal–superaural line to define the Frankfurt horizontal. The vertical plane was placed perpendicular to the horizontal plane 12 cm posterior to the pupil reconstructed point. The median plane was placed through the pupil reconstructed point perpendicular to the horizontal and vertical planes.
This frame could not be used for superimpositions of growing individuals, due to the fixed distance of the vertical plane 12 cm posterior to the pupil reconstructed point. Therefore, the frame was extended by placing a plane parallel to the vertical plane through the pretragion point. The pretragion point was marked at the middle of tragus on the attachment of the ear to the face on both the right and left sides of the face ( Fig. 5 ). When the right and left pretragion points were not in the same antero-posterior position, Maxilim ® was used to calculate the mean of the left and the right side and the plane was placed through this mean. When it was impossible to place both the right and left pretragion points due to incomplete imaging of both ears in the 3D photograph only one pretragion point was placed. This plane through the pretragion point was called the ‘centre plane’ and simulated the plane (centre of gravity) from which growth is described ( Fig. 6 ). The centre, horizontal, and median planes together formed the children’s reference frame, which was used to superimpose 3D pictures of growing individuals to study relative growth. To determine intra- and interobserver reproducibility of the children’s reference frame, two extensively trained observers placed the frame independent of one other on each of the 39 photographs on two occasions (series 1 and series 2) with an interval of 4 weeks to prevent memory effects.