PEDIATRIC/CRANIOFACIAL Reliability of Nasolabial Anthropometric Measures Using Three-Dimensional Stereophotogrammetry in Infants with Unrepaired Unilateral Cleft Lip Raymond Tse, M.D. Lindsay Booth, D.D.S. Kari Keys, M.D. Babette Saltzman, Ph.D. Erik Stuhaug, A.A.S. Hitesh Kapadia, D.D.S., Ph.D. Carrie Heike, M.D., M.S. Seattle, Wash.
Background: Surgical and orthodontic treatment of unilateral cleft lip +/– palate can produce dramatic changes in nasolabial form; however, the lack of ideal methods with which to objectively document three-dimensional form limits the ability to assess treatment outcomes. The purpose of this study was to determine the reliability of three-dimensional stereophotogrammetry for anthropometric assessment of the unilateral cleft lip +/– palate deformity in infants before cleft lip repair. Methods: Preoperative three-dimensional images were acquired from 26 consecutive patients with unrepaired unilateral cleft lip +/– palate. Three raters performed indirect anthropometry on each image on two separate occasions, with at least 1 week between rating sessions. One rater performed direct measurements on participants before surgery while in the operating room. Twenty-six linear and angular measurements were considered, and intrarater, interrater, and intermethod reliability were assessed. Results: Regarding intrarater and interrater reliability, most measurements had Pearson coefficients greater than 0.75, mean differences less than 0.8 mm, and mean proportional differences less than 0.1. For measurements involving vermilion height, nostril remnants, or Cupid’s bow width, Pearson coefficients ranged from 0.3 to 0.75, mean differences ranged from 0.4 to 0.9 mm, and mean proportional differences ranged from 0.1 to 0.3. Regarding intermethod reliability, correlation coefficients ranged from 0.4 to 0.75 for most measurements. The mean differences for nose and lip measurements were less than 1 mm and between 0.8 and 1.3 mm, respectively. Conclusion: Three-dimensional stereophotogrammetry provides a reliable method for many anthropometric measurements of nasolabial form in infants with unrepaired unilateral cleft lip +/– palate. (Plast. Reconstr. Surg. 133: 530e, 2014.) CLINICAL QUESTION/LEVEL OF EVIDENCE: Diagnostic, II.
M
ost cleft centers rely on conventional two-dimensional photography for longitudinal documentation of facial form during the treatment of unilateral cleft lip +/– palate. Although convenient, outcome analysis of three-dimensional form using two-dimensional images is limited by error from parallax. Farkas has previously addressed this problem using anthropometric analysis of facial form, but the burden of performing measurements on infants From the Division of Plastic and Craniofacial Surgery and the Division of Craniofacial Medicine, Seattle Children’s Hospital; the Departments of Surgery and Pediatrics, University of Washington School of Medicine; and the University of Washington School of Dentistry.
530e
and children makes it impractical in this population. Digital three-dimensional stereophotogrammetry may be an ideal method of capturing and measuring three-dimensional facial form of awake infants (Fig. 1), as it is rapid, noninvasive, and easily repeated. Although the reliability of anthropometric measurements on three-dimensional images has been established in adults without Received for publication August 27, 2013; accepted October 16, 2013. Presented at the 12th International Congress on Cleft Lip/Palate and Related Craniofacial Anomalies, in Orlando, Florida, May 5 through 10, 2013. Copyright © 2014 by the American Society of Plastic Surgeons DOI: 10.1097/PRS.0000000000000014
www.PRSJournal.com
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry
Fig. 1. Example of a three-dimensional image in multiple orientations.
craniofacial conditions,1,2 older children,3,4 and on dental casts,5 the reliability has yet to be evaluated in infants with unrepaired unilateral cleft lip +/– palate. The purpose of this study was to determine the reliability of anthropometric measurements specific to unilateral cleft lip +/– palate using digital three-dimensional stereophotogrammetry on infants before cleft lip repair.
PATIENTS AND METHODS We conducted a prospective study approved by the Seattle Children’s Hospital Institutional Review Board. Consecutive patients evaluated in our Craniofacial Center with unrepaired unilateral cleft lip +/– palate were invited to participate. Of the 28 eligible participants, 26 families agreed to participate. Written informed consent was obtained for each subject. The overall study design is illustrated in Figure 2. Data Collection We collected demographic data, including sex, ethnicity, associated anomalies, and age at time of anthropometric measurement. Cleft characteristics were recorded using a modified LAHSHAL Disclosure: None of the authors has a financial interest in any of the products or devices mentioned in this article.
system6: the anatomy is represented from the participant’s right to left (lip, alveolus, hard palate, soft palate) with capital letters representing complete, small letters representing incomplete, and asterisks (*) representing microform clefts; we added a plus sign (+) to document the presence of Simonart bands of the lip. Microform cleft lip was defined as a cleft limited to less than one-fourth of the upper lip height7; incomplete clefts extend beyond this. A complete cleft lip with Simonart band was differentiated from incomplete cleft lip by the presence of a complete cleft alveolus.8 Using the modified LAHSHAL system, a right complete cleft with Simonart band would be +LA, whereas a right incomplete cleft lip would be la. Three-dimensional images were captured according to standard guidelines3,9 within 1 week before cleft lip repair by a professional imaging technologist using the 3dMDCranial system (3dMD, Atlanta, Ga.). The 15-camera system (10 data cameras and five texture cameras) captures images in less than 2 msec. Synchronized image acquisition combined with stereo triangulation algorithms produce a continuous point cloud (with x, y, and z coordinates) represented as a geometric mesh with an overlying color texture (www.3dmd.com) (Fig. 1). The system is calibrated each clinic day according to the manufacturer’s protocol and checked before each image acquisition. The use of fixed dimensions and shapes on a calibration board allows algorithms to quantify
531e
Plastic and Reconstructive Surgery • April 2014
Fig. 2. Data collection and analysis. 3D, three-dimensional.
image data into accurate physical measurements. Data point density varies with shape and location of the surface captured and of the camera configuration. Our system is used in a high-volume craniofacial center and is designed for capture of both face and full cranial features. Data point density was approximately 115 points/cm2 on exposed surfaces. We wiped infants’ faces with a dry cloth to minimize artifact from oronasal secretions and placed participants in an infant chair before image capture.9 Images of the face included features from ear to ear and under the chin. Multiple images of each subject were captured, and the one with the best resolution and least facial animation was selected. In addition to the three-dimensional images, standard frontal, worm’s-eye, and lateral profile views were captured using conventional two-dimensional digital photography and were available for reference. Study images were deidentified and assigned a study code. We used a previously described system of anthropometric measurements (Fig. 3 and Table 1) that is recorded immediately before cleft lip repair to assist with planning and documentation of the cleft deformity.10–13 The measurement set includes nine dimensions of the nose and 12 dimensions of the lip. The Noordhoff point was defined as the point along the vermilion border of the lateral lip element where the vermilion is at
532e
its greatest height.14 This point was used to define the lateral segment Cupid’s bow peak (cphi′′), as the Noordhoff point is an identifiable landmark that does not vary with technique of cleft lip repair (i.e., rotation-advancement versus inferior triangle)12 or surgeon preference.15,16 On the medial segment, the cleft side Cupid’s bow peak (cphi′) was based on a mirror image of the noncleft side, so that cphi′ was equal distance to ls (i.e., cphi-ls = ls-cphi′). Direct anthropometric measurements were obtained under general anesthesia (according to standard clinical protocol) by the senior surgeon (R.T.) using 2.5× loupes and calipers for distances (to the nearest 0.5 mm) and a clear protractor for angles (to the nearest 5 degrees). These direct measurements were considered the criterion standard. At least 1 month after direct measurement by the senior surgeon, the senior surgeon and two other raters (senior plastic surgery resident and dental student) performed the same set of anthropometric measurements on the captured preoperative three-dimensional images using a 3dMD Vultus v.2.2.011 landmark analysis template (indirect measurements) on two occasions at least 1 week apart (Fig. 2). Images were displayed at 2 to 3× magnification to mimic loupe magnification, and landmark “dots” of less than 0.5 mm were used. Landmarks were identified using the textured
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry
Fig. 3. Preoperative measurements for incomplete (left) and complete (right) unilateral cleft lip. (Modified with permission from drawing of Lee-min Lee, Noordhoff Craniofacial Foundation; Fisher DM, Tse R, Marcus JR. Objective measurements for grading the primary unilateral cleft lip nasal deformity. Plast Reconstr Surg. 2008;122:874–880). Definitions of measurements are listed in Table 1.
surface image and using the raw triangular mesh (Fig. 4). All measurement sets were recorded separately to ensure that raters were blinded to prior results. Raters were calibrated using sample
two- and three-dimensional images of infants with cleft lip (who were not part of the study population). The senior surgeon trained the resident to recognize landmarks during a series of cleft lip
Table 1. Anthropometric Measurements Measurement (mm) Nose Nasal width Columellar deviation (degrees) Columellar width Nostril height, noncleft Nostril height, cleft (iCL only) Nostril width, noncleft Nostril width, cleft Cleft nostril remnant, medial (cCL only) Cleft nostril remnant, lateral (cCL only) Lip Lip width Cleft lip remnant height (iCL only) Medial lip height, noncleft Medial lip height, cleft Lateral lip height, noncleft Lateral lip height, cleft Lateral lip length, noncleft Lateral lip length, cleft Cupid’s bow width Vermilion height, noncleft Vermilion height, medial cleft Vermilion height, lateral cleft
No.*
Measurement Definition
19 2 8 4 5 6 7 7M
Widest horizontal dimension from alare to alare Vertical deviation of columellar axis Transverse width of columella at narrowest point Greatest vertical height of nostril Greatest vertical height of nostril Greatest horizontal width of nostril Greatest horizontal width of nostril Shelf from columella to vermilion edge
7L
Shelf from alar insertion to vermilion edge
20 3
Widest horizontal dimension from cheilion to cheilion Cutaneous remnant from nostril sill to cleft vermilion
12 13 10 11 14 15 9 16 17 18
Subnasale to upper edge of white roll over crista philtri Subnasale to upper edge of white roll over crista philtri Subalare to upper edge of white roll over crista philtri Subalare to upper edge of white roll over Noordhoff point Cheilion to crista philtri Cheilion to Noordhoff point Crista philtri to crista philtri Height of dry red lip at crista philtri Height of dry red lip at crista philtri Height of dry red lip at Noordhoff point
Anthropometric Notations al-al′ sn-sn′
ch-ch′ sn-cphi sn-cphi′ sbal-cphi sbal′-cphi′′ ch-cphi ch′-cphi′′ cphi-cphi′
iCL, incomplete cleft lip; cCL, complete cleft lip. *Corresponding to Figure 3.
533e
Plastic and Reconstructive Surgery • April 2014 repairs over the course of 6 consecutive months and both the resident and dental student to recognize landmarks on three-dimensional images. All images used in this study were collected for clinical purposes and none had landmarks placed at the time of image capture. To determine factors that may impact the ability of a rater to confidently measure dimensions on a three-dimensional image, raters made notes of image limitations and measurement difficulties for each subject at each measurement session. Statistical Analysis Precision (intrarater reliability) and fidelity (interrater reliability) were evaluated by three statistics (Fig. 2): (1) intraclass correlation coefficients with 95 percent confidence intervals (high reliability when close to 1 and low reliability when close to 0); (2) mean difference (the mean absolute difference between measurements); and (3) proportional difference (the mean absolute difference divided by the grand mean for a given variable). Accuracy (intermethod reliability) assessed the surgeon’s direct measurements against the surgeon’s indirect measurements (first session). Pearson correlation coefficients with 95 percent confidence intervals were estimated by bootstrapping methods (values between 0 and 1, with higher estimates indicating higher reliability),
mean difference, and proportional difference. Analyses were performed using Stata version 9.0 (Stata Corp., College Station, Texas).
RESULTS Demographics Participants in this study demonstrated typical cleft characteristics (Tables 2 and 3). At our Craniofacial Center, cleft lip repair is typically performed when patients are between 3 and 6 months of age. Two children underwent repair at 10 and 12 months of age because of associated medical conditions. Two participants established care in our Center at a late age after international adoption and underwent cleft lip repair at 24 months of age. Intrarater Reliability Intrarater reliability estimates are summarized in Table 4. The mean difference between sessions in the measured distances was always less than 1 mm. The proportional difference was less than 0.1 for most measured distances but rose to just over 0.25 for the five measurements that involved dimensions less than 4 mm (nostril remnants, measurements 7M and 7L; cleft lip remnant height, measurement 3; and vermilion heights, measurements 16, 17, and 18).
Fig. 4. Close-up of example three-dimensional image. The vermilion-mucosal border is difficult to visualize. Data holes, artifacts, and loss of resolution within cleft nostril make measurements of nostril remnants difficult.
534e
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry Table 2. Demographic and Clinical Characteristics of the 26 Patients with Cleft Lip Who Were Included in This Study Characteristics
No. (%)
Demographic Age, mo Mean Median Range Sex Male Race/ethnicity Caucasian Asian Latino African American Clinical Side of cleft Right Left Type of cleft Microform Incomplete Complete with Simonart band Complete Associated syndromes No Yes
6.9 5.5 3–24 14 (54) 4 (15) 2 (23) 1 (4) 1 (4) 15 (58) 11 (42) 0 (0) 12 (46) 4 (15) 10 (38) 24 (92) 2 (8)*
*Popliteal pterygium and craniofacial microsomia.
Correlation coefficients were greater than 0.75 for most measurements. However, estimates were 0.4 to 0.75 for measurements of small dimensions,
including columellar width (measurement 8, sn-sn′), cleft nostril height (measurement 5), cleft nostril remnants (measurements 7M and 7L), cleft lip remnant height (measurement 3), vermilion heights (measurements 16, 17, 18), and Cupid’s bow width (measurement 9, cphi-cphi′). Estimates of the correlation coefficient were less than 0.4 for the medial cleft vermilion height (measurement 17) for one rater; however, the mean difference was only 0.5 mm, which suggests that this result was related to the small dimension of the measurement. The mean difference between sessions in measured angles (columellar deviation, 2) was always less than 6 degrees. The proportional difference was approximately 0.15 and the correlation coefficient was approximately 0.9. Interrater Reliability Estimates for the interrater reliability were similar (Table 4). The mean difference between raters was less than 1 mm for most measurements, with a maximum difference of 1.3 mm for the larger measurements of lip width (measurement 20, ch-ch′) and nose width (measurement 19, al-al′) (mean dimension, 33 and 29 mm, respectively). The proportional difference was close to 0.1 for most measured distances but ranged between 0.15 and 0.32 for measurements that involved small dimensions
Table 3. Cleft Characteristics of the 26 Patients with Cleft Lip According to a Modified LAHSHAL Notation Subject No.* 3 14 13 16 19 21 26 5 1 24 12 9 23 25 11 15 20 8 10 18 17 22 6 7 2 4
LAHSHAL Classification Si
+
L
A
l l l l l l l l L L L L L L
a a a a a a A A A A A A
H
S
H
A
L
H H H H H H h h
A A A A A A a a A A a a
L L L L L L l l L L l l
Si
s h h H H H
S S S S S S S S S S S
+ + +
Si, Simonart band; L, lip; A, alveolus; H, hard palate; S, soft palate. *Subject number is in order of consecutive enrollment.
535e
Plastic and Reconstructive Surgery • April 2014 Table 4. Intrarater and Interrater Reliability Measurement Intrarater reliability† Nasal width Columellar deviation (degrees) Columellar width Nostril height Noncleft Cleft (iCL only) Nostril width Noncleft Cleft Cleft nostril remnant Medial (cCL only) Lateral (cCL only) Lip width Cleft lip remnant height (iCL only) Medial lip height Noncleft Cleft Lateral lip height Noncleft Cleft Lateral lip length Noncleft Cleft Cupid’s bow width Vermilion height Noncleft Medial cleft Lateral cleft Interrater reliability‡ Nasal width Columellar deviation (degrees) Columellar width Nostril height Noncleft Cleft (iCL only) Nostril width Noncleft Cleft Cleft nostril remnant Medial (cCL only) Lateral (cCL only) Lip width Cleft lip remnant height (iCL only) Medial lip height Noncleft Cleft Lateral lip height Noncleft Cleft Lateral lip length Noncleft Cleft Cupid’s bow width Vermilion height Noncleft Medial cleft Lateral cleft
Mean (mm)*
SD
Mean Difference
Proportional Difference
Correlation Coefficient (95% CI)
29.4 30 4.5
2.7 14.2 0.9
0.4 4.4 0.5
0.01 0.15 0.11
0.99 (0.98–1.00) 0.93 (0.88–0.98) 0.69 (0.50–0.89)
5.9 5.6
0.8 0.6
0.2 0.3
0.03 0.05
0.92 (0.87–0.98) 0.74 (0.51–0.98)
6.7 11.8
1.2 3.3
0.3 0.7
0.04 0.06
0.95 (0.91–0.99) 0.97 (0.95–0.99)
3 3.1 34 2.5
0.6 1.2 4.5 1.2
0.7 0.6 0.6 0.4
0.23 0.19 0.02 0.16
0.44 (0.00–0.87) 0.74 (0.51–0.98) 0.98 (0.97–1.00) 0.93 (0.85–1.00)
8.3 5
1.4 1.2
0.6 0.4
0.07 0.08
0.87 (0.78–0.96) 0.90 (0.83–0.97)
9.4 7.2
1.7 1.7
0.5 0.5
0.05 0.07
0.93 (0.88–0.98) 0.91 (0.85–0.98)
17.4 14.8 6.3
2.6 2 1
0.7 0.6 0.6
0.04 0.04 0.1
0.93 (0.88–0.98) 0.92 (0.86–0.98) 0.76 (0.60–0.92)
3.4 2 3
0.7 0.6 0.7
0.5 0.4 0.6
0.15 0.2 0.2
0.69 (0.49–0.88) 0.46 (0.17–0.76) 0.59 (0.35–0.84)
28.8 35.6 4.4
2.8 14.3 0.9
1.3 6.9 0.5
0.05 0.19 0.11
0.86 (0.77–0.94) 0.82 (0.71–0.92) 0.77 (0.65–0.90)
5.8 5.5
0.8 0.7
0.4 0.4
0.07 0.07
0.81 (0.71–0.92) 0.63 (0.37–0.89)
6.8 11.5
1.1 3.2
0.4 0.8
0.06 0.07
0.90 (0.83–0.96) 0.94 (0.90–0.98)
3 2.8 33.4 2.6
1.1 1.2 3.9 0.9
0.7 0.9 1.2 0.7
0.23 0.32 0.04 0.27
0.61 (0.34–0.88) 0.65 (0.40–0.90) 0.94 (0.91–0.98) 0.71 (0.50–0.93)
8.7 5.4
1.4 1.5
0.8 0.7
0.09 0.13
0.77 (0.65–0.90) 0.75 (0.61–0.88)
10 7.4
1.8 1.6
0.8 0.9
0.08 0.12
0.86 (0.78–0.94) 0.74 (0.61–0.88)
16.9 14.4 7.4
2.3 1.6 1.1
0.9 1 1
0.05 0.07 0.14
0.89 (0.83–0.96) 0.72 (0.58–0.87) 0.34 (0.10–0.58)
3.4 2.3 2.6
0.7 0.5 0.5
0.5 0.5 0.5
0.15 0.22 0.19
0.64 (0.46–0.82) 0.48 (0.26–0.70) 0.52 (0.30–0.73)
iCL, incomplete cleft lip; cCL, complete cleft lip. *All linear measurements were in millimeters and all angular measurements were in degrees. †The intrarater reliability estimates for one of the three raters (the estimates were qualitatively similar for each of the three raters). ‡The correlation between the three raters during the first rating session.
(e.g., nostril remnants, measurements 7M and 7L; cleft lip remnant height, measurement 3; and vermilion heights, measurements 16, 17, and 18).
536e
Correlation coefficients were greater than 0.75 for most measurements. However, estimates for measurements of relatively small dimensions were 0.4
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry to 0.75 (e.g., columellar width, measurement 8, sn-sn′; cleft nostril height, measurement 5; cleft lip remnant height, measurement 3; and vermilion height, measurements 16, 17, and 18). Correlation coefficients were less than 0.4 for the cleft medial nostril remnant and the Cupid’s bow width. The mean difference between raters in the columellar deviation angle (measurement 2) was less than 8 degrees. The proportional difference was approximately 0.2 and the correlation coefficient was 0.8. Intermethod Reliability Correlation coefficients were greater than 0.75 for nasal width, columellar deviation, nostril widths, lip width, and noncleft lateral lip height. They were 0.4 to 0.75 for columellar width, nostril heights, medial lip heights, cleft lateral lip height, and lateral lip lengths. The mean difference between direct and indirect measurements was less than 1.5 mm for most measurements, with a maximum difference of 3.9 mm for lip width (measurement 20, ch-ch′), corresponding to a 10 percent proportional difference. Similar to intrarater and interrater reliability, the intermethod reliability was less than 0.4 for the cleft nostril remnants (measurements 7M and 7L), the vermilion heights (measurements 16, 17, and 18), and the Cupid’s bow width (measurement 9, cphi-cphi′). The mean difference between direct and indirect measurements was within 1 mm for most measurements and up to 1.9 for the Cupid’s bow width (measurement 9, cphi-cphi′).
Qualitative Observations The raters recorded the following findings that impacted their ability to confidently place landmarks: (1) some three-dimensional images included data holes and artifacts within the nostrils or the cleft (Fig. 4); (2) dry vermilion was often difficult to differentiate from wet mucosa; (3) mesh and texture resolution was better for older children; and (4) images of children often displayed some degree of facial animation (Fig. 5). To subjectively evaluate image quality in younger compared with older infants, we systematically reviewed the three-dimensional images for all subjects, comparing their initial three-dimensional images at the time of presentation (usually at 1 to 2 weeks of age) to those captured as part of this preoperative study protocol. Form and texture definition were better with older age. We also noted that 24 of the 26 study images were captured with a child who was awake.
DISCUSSION Treatment of children with unilateral cleft lip +/– palate spans childhood; occurs over multiple stages; requires collaboration among multiple specialists; and is associated with variable outcomes related to patient characteristics, surgical technique, and surgical provider. Understanding the relative effect of individual treatments, techniques, or protocols relies on accurate and objective documentation of form over the course of a child’s treatment. Most cleft centers use serial two-dimensional photographs to record facial changes. However, measurements performed on such images are
Fig. 5. Comparison of facial animation with direct and indirect measurements. Direct measurements were performed under anesthetic immediately before cleft lip repair. Three-dimensional images were captured in the clinic at the preoperative visit, generally, with an awake infant. Animation distorts nasolabial features.
537e
Plastic and Reconstructive Surgery • April 2014 prone to error secondary to variations in camera angle relative to the subject (i.e., parallax).17 Direct anthropometric measurement eliminates parallax, and Farkas published valuable reference data using this approach; however, this method is time consuming and difficult to perform on awake infants. Reports of direct measurements on children younger than 1 year are rare18–22 and involve limited select facial dimensions23 or require an asleep baby with10,24 or without19 anesthetic. Nasolabial casts provide a three-dimensional record of form25,26; however, taking impressions on an awake child is burdensome, time-intensive, and involves risks of aspiration. Infants are invariably distressed when the impressions are taken, and this leads to distortion of facial form. Finally, storage and management of casts can be costly. Three-dimensional laser scanning can produce accurate digital three-dimensional images27 for reliable anthropometric analysis.28 Such data have been used to study facial morphology29,30 and postoperative outcomes of cleft surgery.31,32 However, the relatively long image acquisition time is a barrier to using the technology with infants, and image quality has been found to vary inversely with subject age.30 Three-dimensional optical scanning is faster and can be applied to capture facial form in older children.33–36 A recent pilot study has demonstrated reliable anthropometric measurements in infants with unilateral cleft lip +/– palate37; however, the set of dimensions evaluated was limited and does not include all of the dimensions of interest that we have examined. In addition, the lack of texture resolution on the generated images limits the application of this type of imaging. Three-dimensional stereophotogrammetry has advantages related to its rapid image capture speed and noninvasive nature. Compared with other methods, there is no burden and there is no distortion from parallax. The accuracy of three-dimensional stereophotogrammetry has been documented on mannequin heads,38 on normal adults,1,2 and on children and adults with noncleft facial dysmorphology.3,4 However, the reliability of three-dimensional stereophotogrammetry has yet to be established directly on infants who have unrepaired unilateral cleft lip +/– palate. Given the young age, small dimensions, and unique form of the unrepaired unilateral cleft lip +/– palate, the accuracy of three-dimensional stereophotogrammetry cannot be assumed in this population based on previous reports. Our study examined the reliability of a detailed set
538e
of anthropometric measurements that are used specifically for cleft lip repair.11–13 In contrast to other studies, these measurements have not been evaluated on a three-dimensional model before, and we measured reliability within the context of a standard clinical routine. We used the entire data set and did not exclude any image or subject. As such, measured variations reflect a typical clinical scenario rather than a highly selected data set in which lower quality images or more difficult dimensions to measure have been excluded. We found that most but not all of the 21 measurements were reliable. Measurements involving vermilion heights, Cupid’s bow width, and nostril remnants were less reliable. The mean difference in vermilion heights was on the order of 0.5 mm; however, with dimensions of 2 to 4 mm, the proportional difference was large, and this was reflected in the Pearson correlation coefficients. Vermilion height can be challenging to measure, as the junction of dry vermilion and wet mucosa can be subtle. Without tactile feedback or the ability to change light reflections with tissue movement, this junction becomes even more difficult to identify on conventional two-dimensional photography. Although a three-dimensional image can be rotated as a whole, the lip cannot be moved for exposure, and we found that the texture resolution of three-dimensional images was limited. As a result, measurements of vermilion height and dimensions based on the Noordhoff point (cphi″; identified by the maximal vermilion height) are more prone to error than other measurements. Reliable measurement of Cupid’s bow width can be difficult because of flattening of the curvature of Cupid’s bow peak, especially on the cleft side. Some surgeons use this lack of definition to their advantage for cleft lip repair by cheating the point of closure medially or laterally.39 Given the limitations of texture resolution on three-dimensional images, we were not surprised by variations in measurements based on chpi′. Nostril remnants along the nasal sill on either side of a cleft can be better exposed by manipulation of the lip and nose. On static three-dimensional images, these shelves can be difficult to visualize, with data holes, artifacts, and loss of data resolution that are often present in these dark areas (Fig. 4). The mean difference in measurement of these areas was within 1 mm; however, the small dimensions of 2 to 3 mm made
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry the proportional differences large and consequently the correlation coefficients unfavorable. Some degree of intermethod variation was expected given that direct anthropometric measurements were performed while subjects were asleep under general anesthesia immediately before cleft lip repair, whereas the three-dimensional images were captured while subjects were awake during their preoperative clinic visit (Fig. 5). Although we attempted to capture infants with a neutral facial expression, most awake infants demonstrate some degree of animation during image acquisition. On review of the three-dimensional images, all but two appeared to have been captured while the subject was awake. Anthropometric measurements are prone to stretch and tone of the face with animation, especially those based off of the cheilion. Consistent with this, all proportional differences were higher, and intermethod correlation values were lower, especially those measured from landmarks prone to movement associated with animation (such as lip width, lateral lip lengths, and lateral lip heights). Capturing images while subjects are asleep would more closely mimic our reference standard using direct measurements; however, doing so during a routine clinic visit would be difficult and impractical. Three-dimensional image capture under anesthesia would potentially have also improved our ability to measure intermethod correlation; however, bringing the three-dimensional system into the operating room is not practical and would deviate from our current clinical practice. For the same reasons, we could not determine the senior surgeon’s variance performing direct measurements. We assumed that direct measurement would have higher reliability, as the tactile feedback and manipulation of tissues improves visualization and recognition of anatomical landmarks (especially vermilion, nostril remnants, and Cupid’s bow). These differences likely contribute to the observed intermethod variability. Three-dimensional stereophotogrammetry has been used to measure normal facial morphology,40 to compare children with unrepaired cleft lip to normal controls,41–43 and to document form after orthodontic44,45 and surgical17,46–49 treatment of unilateral cleft lip +/– palate. Although use of this technology is attractive and becoming more popular, understanding the benefits and limitations in infants with unrepaired unilateral cleft lip +/– palate is important: 1. For gross nasolabial dimensions measured in previous studies,9,23,50 we found that the reliability was very good.
2. For measurements that are important descriptors of cleft severity,10 such as columellar angle and nostril width ratio, we found that the reliability was similarly very good. 3. Certain measurements are key in the design of cleft lip repair, and these were found to be reliable. The relative medial lip heights define the degree to which the lip needs to be corrected by means of rotation (i.e., rotation-advancement approaches), back-cut (i.e., inferior triangle approaches), or Rose-Thompson effect. The lateral lip heights and lengths define the amount of tissue available for medial augmentation by means of advancement and/or laterally based flaps. Fisher has demonstrated how these lateral lip dimensions can have significant impacts depending on the design used for repair.13 Ultimately, all of these measurements define the “defect” to be reconstructed and are relevant to surgical outcomes. 4. Other measurements that can be used in cleft lip repair are subtler and in our study were found to be less reliable. Relative vermilion heights determine the augmentation necessary to avoid leaving chronically exposed mucosa at the site of repair,14,51–53 and the size of the nostril remnants impacts on the design for appropriate nasal sill and floor closure.51 Cleft surgeons have an obligation to audit their results, and three-dimensional images provide a potentially powerful means of doing so. Farkas provided us with a framework to objectively analyze facial form; however, the burden of performing measurements has limited the use of this approach. Three-dimensional imaging may allow widespread use so that we can be more objective in analyzing our outcomes. Our center’s imaging technologist captures three-dimensional images in conjunction with standard two-dimensional photographic records according to our center’s protocol. Once the three-dimensional system is calibrated (which requires 10 minutes at the start of clinic) and a child is positioned,9 additional three-dimensional image capture is as rapid as conventional photography. The generated three-dimensional model can be moved and rotated in much the same way as we appreciate facial form by varying our viewing angles and perspectives. The set of anthropometric measurements used in this study takes 5 to 10 minutes for
539e
Plastic and Reconstructive Surgery • April 2014 Table 5. Intermethod Reliability Mean Measurement (mm)* Measurement Nose Nasal width Columellar deviation (degrees) Columellar width Nostril height Noncleft Cleft (iCL only) Nostril width Noncleft Cleft Cleft nostril remnant Medial (cCL only) Lateral (cCL only) Lip Lip width Cleft lip remnant height (iCL only) Medial lip height Noncleft Cleft Lateral lip height Noncleft Cleft Lateral lip length Noncleft Cleft Cupid’s bow width Lateral lip length Noncleft Medial cleft Lateral cleft
Indirect 3D
Direct Caliper
Mean Difference
Proportional Difference
Correlation Coefficient (95% CI)
29.57 29.75 4.76
30.42 27.31 4.97
1.31 8.91 0.49
0.04 0.33 0.1
0.9 (0.8–1.0) 0.8 (0.6–0.9) 0.7 (0.6–0.9)
5.9 5.65
5.87 4.69
0.72 1
0.12 0.21
0.4 (0.1–0.7) 0.5 (0.0–0.9)
6.44 11.33
5.6 10.71
0.89 0.85
0.16 0.08
0.9 (0.8–1.0) 1.0 (0.9–1.0)
2.97 3.01
3.61 4.36
0.84 1.66
0.23 0.38
0.3 (−0.3–0.9) 0.3 (−0.4–0.9)
34.23 2.55
38.08 2.98
3.88 1.1
0.1 0.37
0.8 (0.6–0.9) 0.3 (−0.4–0.9)
8.45 5.18
7.99 4.73
1.12 0.8
0.14 0.17
0.4 (0.2–0.7) 0.6 (0.4–0.9)
9.63 7.36
9.25 7.77
0.94 1.27
0.1 0.16
0.8 (0.6–0.9) 0.6 (0.4–0.8)
17.65 14.89 6.48
17.93 15.02 8.08
1.53 1.79 1.86
0.09 0.12 0.23
0.7 (0.5–0.9) 0.5 (0.2–0.8) 0.0 (−0.3–0.4)
3.42 2.05 3.08
4.33 2.28 4.11
0.95 0.66 1.11
0.22 0.29 0.27
0.7 (0.4–0.9) 0.2 (−0.3–0.7) 0.5 (0.2–0.8)
iCL, incomplete cleft lip; cCL, complete cleft lip. *All linear measurements were in millimeters and all angular measurements were in degrees.
an experienced rater and can be performed at any time once the images have been stored. In addition, we have found the three-dimensional model to be helpful for preoperative planning and for resident teaching. In spite of the advantages of three-dimensional imaging, current systems are expensive and cannot replace standardized two-dimensional photography.54 Limitations in texture resolution on three-dimensional systems compared with two-dimensional images may not allow for reliable scar assessment. Although our experience using three-dimensional technology for longitudinal assessment is limited, we adhere to a standard protocol for image acquisition9 so that future comparisons can be made. The three-dimensional system used in our study is optimized for image capture of the pediatric face and head in the setting of a busy craniofacial center. Although resolution may be improved using a different configuration dedicated to capture of facial features only, the need to recalibrate the system makes multiple configurations in a single clinic impractical. Our measures of reliability could also be improved by rejecting lower
540e
quality images; however, we felt that evaluation of this system in a normal clinical setting would be important. Just as digital photography has replaced conventional film-based photography at many institutions, with further refinements, three-dimensional imaging technology may one day replace two-dimensional photography. Until then, it is important to understand the limitations of this technology.
CONCLUSIONS Three-dimensional stereophotogrammetry provides a rapid-capture, noninvasive, low-burden method of documenting nasolabial form without problems of parallax. The current report provides an evaluation of the many features that can be reliably captured using this method, and those that may not. Three-dimensional stereophotogrammetry is a valuable tool that may contribute to the current tools used for comprehensive documentation of prerepair unilateral cleft lip +/– palate form.
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry Raymond Tse, M.D. Division of Plastic and Craniofacial Surgery Department of Surgery Seattle Children’s Hospital 4800 Sand Point Way NE M/S OB.9.527 Seattle, Wash. 98105 [email protected]
ACKNOWLEDGMENT
This work was supported by the University of Washington Dental Alumni Fund. PATIENT CONSENT
Parents or guardians provided written consent for the use of patients’ images. REFERENCES 1. Weinberg SM, Scott NM, Neiswanger K, Brandon CA, Marazita ML. Digital three-dimensional photogrammetry: Evaluation of anthropometric precision and accuracy using a Genex 3D camera system. Cleft Palate Craniofac J. 2004;41:507–518. 2. Wong JY, Oh AK, Ohta E, et al. Validity and reliability of craniofacial anthropometric measurement of 3D digital photogrammetric images. Cleft Palate Craniofac J. 2008;45:232–239. 3. Heike CL, Cunningham ML, Hing AV, Stuhaug E, Starr JR. Picture perfect? Reliability of craniofacial anthropometry using three-dimensional digital stereophotogrammetry. Plast Reconstr Surg. 2009;124:1261–1272. 4. Aldridge K, Boyadjiev SA, Capone GT, DeLeon VB, Richtsmeier JT. Precision and error of three-dimensional phenotypic measures acquired from 3dMD photogrammetric images. Am J Med Genet A 2005;138:247–253. 5. Ayoub A, Garrahy A, Hood C, et al. Validation of a vision-based, three-dimensional facial imaging system. Cleft Palate Craniofac J. 2003;40:523–529. 6. Kriens OE. LAHSHAL: A concise documentation system for cleft lip, alveolus and palate diagnosis. In: Kriens OE, ed. What Is a Cleft Lip and Palate? A Multidisciplinary Update. New York: Thieme; 1989:30–34. 7. Onizuka T, Hosaka Y, Aoyama R, Takahama H, Jinnai T, Usui Y. Operations for microforms of cleft lip. Cleft Palate Craniofac J. 1991;28:293–300; discussion 300. 8. Semb G, Shaw WC. Simonart’s band and facial growth in unilateral clefts of the lip and palate. Cleft Palate Craniofac J. 1991;28:40–46; discussion 46. 9. Heike CL, Upson K, Stuhaug E, Weinberg SM. 3D digital stereophotogrammetry: A practical guide to facial image acquisition. Head Face Med. 2010;6:18. 10. Fisher DM, Tse R, Marcus JR. Objective measurements for grading the primary unilateral cleft lip nasal deformity. Plast Reconstr Surg. 2008;122:874–880. 11. Noordhoff MS. The Surgical Technique for the Unilateral Cleft Lip-Nasal Deformity. Taipei: Noordhoff Craniofacial Foundation; 1997. 12. Tse R. Unilateral cleft lip: Principles and practice of surgical management. Semin Plast Surg. 2012;26:145–155. 13. Boorer CJ, Cho DC, Vijayasekaran VS, Fisher DM. Presurgical unilateral cleft lip anthropometrics: Implications
for the choice of repair technique. Plast Reconstr Surg. 2011;127:774–780. 14. Noordhoff MS. Reconstruction of vermilion in unilateral and bilateral cleft lips. Plast Reconstr Surg. 1984;73:52–61. 15. Losee JE, Selber JC, Arkoulakis N, Serletti JM. The cleft lateral lip element: Do traditional markings result in secondary deformities? Ann Plast Surg. 2003;50:594–600. 16. Cutting CB, Dayan JH. Lip height and lip width after extended Mohler unilateral cleft lip repair. Plast Reconstr Surg. 2003;111:17–23; discussion 24. 17. Oh TS, Choi JW, Koh KS. Upper lip asymmetry perception using three-dimensional anthropometry in patients with unilateral cleft lip deformity. J Craniofac Surg. 2011;22:2080–2083. 18. Farkas LG, Posnick JC, Hreczko TM, Pron GE. Growth patterns of the nasolabial region: A morphometric study. Cleft Palate Craniofac J. 1992;29:318–324. 19. Cho BC, Kim JY, Yang JD, Chung HY, Park JW, Hwang JH. Anthropometric study of the upper lip and the nose of infants less than a year of age. J Craniofac Surg. 2006;17:57–61. 20. Kim SK, Lee JH, Lee KC, Park JM. Mulliken method of bilateral cleft lip repair: Anthropometric evaluation. Plast Reconstr Surg. 2005;116:1243–1251; discussion 1252. 21. Yeow VK, Huang MH, Lee ST, Fook Chong SM. An anthropometric analysis of indices of severity in the unilateral cleft lip. J Craniofac Surg. 2002;13:68–74. 22. Vegter F, Hage JJ. Facial anthropometry in cleft patients: A historical appraisal. Cleft Palate Craniofac J. 2001;38:577–581. 23. Farkas LG, Hajnis K, Posnick JC. Anthropometric and anthroposcopic findings of the nasal and facial region in cleft patients before and after primary lip and palate repair. Cleft Palate Craniofac J. 1993;30:1–12. 24. Mulliken JB, Burvin R, Farkas LG. Repair of bilateral complete cleft lip: Intraoperative nasolabial anthropometry. Plast Reconstr Surg. 2001;107:307–314. 25. Xu Y, Li J, Zhao S, Shi B, Zheng Q, Wang Y. Accuracy of a plastic facial cast fabricated with a custom tray in comparison with cone beam computed tomography. Oral Surg Oral Med Oral Pathol Oral Radiol. 2014;117:238–245. 26. Germec-Cakan D, Canter HI, Nur B, Arun T. Comparison of facial soft tissue measurements on three-dimensional images and models obtained with different methods. J Craniofac Surg. 2010;21:1393–1399. 27. Baik HS, Lee HJ, Lee KJ. A proposal for soft tissue landmarks for craniofacial analysis using 3-dimensional laser scan imaging. World J Orthod. 2006;7:7–14. 28. Toma AM, Zhurov A, Playle R, Ong E, Richmond S. Reproducibility of facial soft tissue landmarks on 3D laser-scanned facial images. Orthod Craniofac Res. 2009;12:33–42. 29. Mori A, Nakajima T, Kaneko T, Sakuma H, Aoki Y. Analysis of 109 Japanese children’s lip and nose shapes using 3-dimensional digitizer. Br J Plast Surg. 2005;58:318–329. 30. Schwenzer-Zimmerer K, Chaitidis D, Boerner I, et al. Systematic contact-free 3D topometry of the soft tissue profile in cleft lips. Cleft Palate Craniofac J. 2008;45:607–613. 31. Nakamura N, Okawachi T, Nishihara K, Hirahara N, Nozoe E. Surgical technique for secondary correction of unilateral cleft lip-nose deformity: Clinical and 3D observations of preoperative and postoperative nasal forms. J Oral Maxillofac Surg. 2010;45:2248–22570. 32. Schwenzer-Zimmerer K, Chaitidis D, Berg-Boerner I, et al. Quantitative 3D soft tissue analysis of symmetry prior to and after unilateral cleft lip repair compared with non-cleft persons (performed in Cambodia). J Craniomaxillofac Surg. 2008;36:431–438. 33. Bilwatsch S, Kramer M, Haeusler G, et al. Nasolabial symmetry following Tennison-Randall lip repair: A three-dimensional
541e
Plastic and Reconstructive Surgery • April 2014 approach in 10-year-old patients with unilateral clefts of lip, alveolus and palate. J Craniomaxillofac Surg. 2006;34:253–262. 34. Nkenke E, Lehner B, Kramer M, et al. Determination of facial symmetry in unilateral cleft lip and palate patients from threedimensional data: Technical report and assessment of measurement errors. Cleft Palate Craniofac J. 2006;43:129–137. 35. Nkenke E, Langer A, Laboureux X, et al. Validation of in vivo assessment of facial soft-tissue volume changes and clinical application in midfacial distraction: A technical report. Plast Reconstr Surg. 2003;112:367–380. 36. Stauber I, Vairaktaris E, Holst A, et al. Three-dimensional analysis of facial symmetry in cleft lip and palate patients using optical surface data (in German). J Orofac Orthop. 2008;69:268–282. 37. Li G, Wei J, Wang X, et al. Three-dimensional facial anthropometry of unilateral cleft lip infants with a structured light scanning system. J Plast Reconstr Aesthet Surg. 2013;66:1109–1116. 38. Weinberg SM, Naidoo S, Govier DP, Martin RA, Kane AA, Marazita ML. Anthropometric precision and accuracy of digital three-dimensional photogrammetry: Comparing the Genex and 3dMD imaging systems with one another and with direct anthropometry. J Craniofac Surg. 2006;17:477–483. 39. Koh KS, Hwang CH, Kim EK. Modified design of Cupid’s bow in the repair of unilateral microform cleft lip. J Craniofac Surg. 2009;20:1367–1369. 40. Yamada T, Mori Y, Minami K, Mishima K, Tsukamoto Y. Three-dimensional analysis of facial morphology in normal Japanese children as control data for cleft surgery. Cleft Palate Craniofac J. 2002;39:517–526. 41. Hood CA, Hosey MT, Bock M, White J, Ray A, Ayoub AF. Facial characterization of infants with cleft lip and palate using a three-dimensional capture technique. Cleft Palate Craniofac J. 2004;41:27–35. 42. Krimmel M, Kluba S, Bacher M, Dietz K, Reinert S. Digital surface photogrammetry for anthropometric analysis of the cleft infant face. Cleft Palate Craniofac J. 2006;43:350–355. 43. Zreaqat M, Hassan R, Halim AS. Facial dimensions of Malay children with repaired unilateral cleft lip and palate: A three dimensional analysis. Int J Oral Maxillofac Surg. 2012;41:783–788.
542e
44. Singh GD, Levy-Bercowski D, Santiago PE. Three-dimensional nasal changes following nasoalveolar molding in patients with unilateral cleft lip and palate: Geometric morphometrics. Cleft Palate Craniofac J. 2005;42:403–409. 45. Simanca E, Morris D, Zhao L, Reisberg D, Viana G. Measuring progressive soft tissue change with nasoalveolar molding using a three-dimensional system. J Craniofac Surg. 2011;22:1622–1625. 46. Devlin MF, Ray A, Raine P, Bowman A, Ayoub AF. Facial symmetry in unilateral cleft lip and palate following alar base augmentation with bone graft: A three-dimensional assessment. Cleft Palate Craniofac J. 2007;44:391–395. 47. Ayoub A, Garrahy A, Millett D, et al. Three-dimensional assessment of early surgical outcome in repaired unilateral cleft lip and palate: Part 1. Nasal changes. Cleft Palate Craniofac J. 2011;48:571–577. 48. Ayoub A, Garrahy A, Millett D, et al. Three-dimensional assessment of early surgical outcome in repaired unilateral cleft lip and palate: Part 2. Lip changes. Cleft Palate Craniofac J. 2011;48:578–583. 49. Singh GD, Levy-Bercowski D, Yáñez MA, Santiago PE. Three-dimensional facial morphology following surgical repair of unilateral cleft lip and palate in patients after nasoalveolar molding. Orthod Craniofac Res. 2007;10:161–166. 50. Farkas LG, Lindsay WK. Morphology of the adult face following repair of unilateral cleft lip and palate in childhood. Plast Reconstr Surg. 1973;52:652–655. 51. Fisher DM. Unilateral cleft lip repair: An anatomical subunit approximation technique. Plast Reconstr Surg. 2005;116:61–71. 52. Mulliken JB, Pensler JM, Kozakewich HP. The anatomy of Cupid’s bow in normal and cleft lip. Plast Reconstr Surg. 1993;92:395–403; discussion 404. 53. Sitzman TJ, Fisher DM. Presurgical unilateral cleft lip anthropometrics: Incidence of vermilion height asymmetry. Plast Reconstr Surg. 2013;131:935e–937e. 54. Birgfeld CB, Saltzman BS, Luquetti DV, Latham K, Starr JR, Heike CL. Comparison of two-dimensional and three-dimensional images for phenotypic assessment of craniofacial microsomia. Cleft Palate Craniofac J. 2013;50:305–314.
Background: Surgical and orthodontic treatment of unilateral cleft lip +/– palate can produce dramatic changes in nasolabial form; however, the lack of ideal methods with which to objectively document three-dimensional form limits the ability to assess treatment outcomes. The purpose of this study was to determine the reliability of three-dimensional stereophotogrammetry for anthropometric assessment of the unilateral cleft lip +/– palate deformity in infants before cleft lip repair. Methods: Preoperative three-dimensional images were acquired from 26 consecutive patients with unrepaired unilateral cleft lip +/– palate. Three raters performed indirect anthropometry on each image on two separate occasions, with at least 1 week between rating sessions. One rater performed direct measurements on participants before surgery while in the operating room. Twenty-six linear and angular measurements were considered, and intrarater, interrater, and intermethod reliability were assessed. Results: Regarding intrarater and interrater reliability, most measurements had Pearson coefficients greater than 0.75, mean differences less than 0.8 mm, and mean proportional differences less than 0.1. For measurements involving vermilion height, nostril remnants, or Cupid’s bow width, Pearson coefficients ranged from 0.3 to 0.75, mean differences ranged from 0.4 to 0.9 mm, and mean proportional differences ranged from 0.1 to 0.3. Regarding intermethod reliability, correlation coefficients ranged from 0.4 to 0.75 for most measurements. The mean differences for nose and lip measurements were less than 1 mm and between 0.8 and 1.3 mm, respectively. Conclusion: Three-dimensional stereophotogrammetry provides a reliable method for many anthropometric measurements of nasolabial form in infants with unrepaired unilateral cleft lip +/– palate. (Plast. Reconstr. Surg. 133: 530e, 2014.) CLINICAL QUESTION/LEVEL OF EVIDENCE: Diagnostic, II.
M
ost cleft centers rely on conventional two-dimensional photography for longitudinal documentation of facial form during the treatment of unilateral cleft lip +/– palate. Although convenient, outcome analysis of three-dimensional form using two-dimensional images is limited by error from parallax. Farkas has previously addressed this problem using anthropometric analysis of facial form, but the burden of performing measurements on infants From the Division of Plastic and Craniofacial Surgery and the Division of Craniofacial Medicine, Seattle Children’s Hospital; the Departments of Surgery and Pediatrics, University of Washington School of Medicine; and the University of Washington School of Dentistry.
530e
and children makes it impractical in this population. Digital three-dimensional stereophotogrammetry may be an ideal method of capturing and measuring three-dimensional facial form of awake infants (Fig. 1), as it is rapid, noninvasive, and easily repeated. Although the reliability of anthropometric measurements on three-dimensional images has been established in adults without Received for publication August 27, 2013; accepted October 16, 2013. Presented at the 12th International Congress on Cleft Lip/Palate and Related Craniofacial Anomalies, in Orlando, Florida, May 5 through 10, 2013. Copyright © 2014 by the American Society of Plastic Surgeons DOI: 10.1097/PRS.0000000000000014
www.PRSJournal.com
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry
Fig. 1. Example of a three-dimensional image in multiple orientations.
craniofacial conditions,1,2 older children,3,4 and on dental casts,5 the reliability has yet to be evaluated in infants with unrepaired unilateral cleft lip +/– palate. The purpose of this study was to determine the reliability of anthropometric measurements specific to unilateral cleft lip +/– palate using digital three-dimensional stereophotogrammetry on infants before cleft lip repair.
PATIENTS AND METHODS We conducted a prospective study approved by the Seattle Children’s Hospital Institutional Review Board. Consecutive patients evaluated in our Craniofacial Center with unrepaired unilateral cleft lip +/– palate were invited to participate. Of the 28 eligible participants, 26 families agreed to participate. Written informed consent was obtained for each subject. The overall study design is illustrated in Figure 2. Data Collection We collected demographic data, including sex, ethnicity, associated anomalies, and age at time of anthropometric measurement. Cleft characteristics were recorded using a modified LAHSHAL Disclosure: None of the authors has a financial interest in any of the products or devices mentioned in this article.
system6: the anatomy is represented from the participant’s right to left (lip, alveolus, hard palate, soft palate) with capital letters representing complete, small letters representing incomplete, and asterisks (*) representing microform clefts; we added a plus sign (+) to document the presence of Simonart bands of the lip. Microform cleft lip was defined as a cleft limited to less than one-fourth of the upper lip height7; incomplete clefts extend beyond this. A complete cleft lip with Simonart band was differentiated from incomplete cleft lip by the presence of a complete cleft alveolus.8 Using the modified LAHSHAL system, a right complete cleft with Simonart band would be +LA, whereas a right incomplete cleft lip would be la. Three-dimensional images were captured according to standard guidelines3,9 within 1 week before cleft lip repair by a professional imaging technologist using the 3dMDCranial system (3dMD, Atlanta, Ga.). The 15-camera system (10 data cameras and five texture cameras) captures images in less than 2 msec. Synchronized image acquisition combined with stereo triangulation algorithms produce a continuous point cloud (with x, y, and z coordinates) represented as a geometric mesh with an overlying color texture (www.3dmd.com) (Fig. 1). The system is calibrated each clinic day according to the manufacturer’s protocol and checked before each image acquisition. The use of fixed dimensions and shapes on a calibration board allows algorithms to quantify
531e
Plastic and Reconstructive Surgery • April 2014
Fig. 2. Data collection and analysis. 3D, three-dimensional.
image data into accurate physical measurements. Data point density varies with shape and location of the surface captured and of the camera configuration. Our system is used in a high-volume craniofacial center and is designed for capture of both face and full cranial features. Data point density was approximately 115 points/cm2 on exposed surfaces. We wiped infants’ faces with a dry cloth to minimize artifact from oronasal secretions and placed participants in an infant chair before image capture.9 Images of the face included features from ear to ear and under the chin. Multiple images of each subject were captured, and the one with the best resolution and least facial animation was selected. In addition to the three-dimensional images, standard frontal, worm’s-eye, and lateral profile views were captured using conventional two-dimensional digital photography and were available for reference. Study images were deidentified and assigned a study code. We used a previously described system of anthropometric measurements (Fig. 3 and Table 1) that is recorded immediately before cleft lip repair to assist with planning and documentation of the cleft deformity.10–13 The measurement set includes nine dimensions of the nose and 12 dimensions of the lip. The Noordhoff point was defined as the point along the vermilion border of the lateral lip element where the vermilion is at
532e
its greatest height.14 This point was used to define the lateral segment Cupid’s bow peak (cphi′′), as the Noordhoff point is an identifiable landmark that does not vary with technique of cleft lip repair (i.e., rotation-advancement versus inferior triangle)12 or surgeon preference.15,16 On the medial segment, the cleft side Cupid’s bow peak (cphi′) was based on a mirror image of the noncleft side, so that cphi′ was equal distance to ls (i.e., cphi-ls = ls-cphi′). Direct anthropometric measurements were obtained under general anesthesia (according to standard clinical protocol) by the senior surgeon (R.T.) using 2.5× loupes and calipers for distances (to the nearest 0.5 mm) and a clear protractor for angles (to the nearest 5 degrees). These direct measurements were considered the criterion standard. At least 1 month after direct measurement by the senior surgeon, the senior surgeon and two other raters (senior plastic surgery resident and dental student) performed the same set of anthropometric measurements on the captured preoperative three-dimensional images using a 3dMD Vultus v.2.2.011 landmark analysis template (indirect measurements) on two occasions at least 1 week apart (Fig. 2). Images were displayed at 2 to 3× magnification to mimic loupe magnification, and landmark “dots” of less than 0.5 mm were used. Landmarks were identified using the textured
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry
Fig. 3. Preoperative measurements for incomplete (left) and complete (right) unilateral cleft lip. (Modified with permission from drawing of Lee-min Lee, Noordhoff Craniofacial Foundation; Fisher DM, Tse R, Marcus JR. Objective measurements for grading the primary unilateral cleft lip nasal deformity. Plast Reconstr Surg. 2008;122:874–880). Definitions of measurements are listed in Table 1.
surface image and using the raw triangular mesh (Fig. 4). All measurement sets were recorded separately to ensure that raters were blinded to prior results. Raters were calibrated using sample
two- and three-dimensional images of infants with cleft lip (who were not part of the study population). The senior surgeon trained the resident to recognize landmarks during a series of cleft lip
Table 1. Anthropometric Measurements Measurement (mm) Nose Nasal width Columellar deviation (degrees) Columellar width Nostril height, noncleft Nostril height, cleft (iCL only) Nostril width, noncleft Nostril width, cleft Cleft nostril remnant, medial (cCL only) Cleft nostril remnant, lateral (cCL only) Lip Lip width Cleft lip remnant height (iCL only) Medial lip height, noncleft Medial lip height, cleft Lateral lip height, noncleft Lateral lip height, cleft Lateral lip length, noncleft Lateral lip length, cleft Cupid’s bow width Vermilion height, noncleft Vermilion height, medial cleft Vermilion height, lateral cleft
No.*
Measurement Definition
19 2 8 4 5 6 7 7M
Widest horizontal dimension from alare to alare Vertical deviation of columellar axis Transverse width of columella at narrowest point Greatest vertical height of nostril Greatest vertical height of nostril Greatest horizontal width of nostril Greatest horizontal width of nostril Shelf from columella to vermilion edge
7L
Shelf from alar insertion to vermilion edge
20 3
Widest horizontal dimension from cheilion to cheilion Cutaneous remnant from nostril sill to cleft vermilion
12 13 10 11 14 15 9 16 17 18
Subnasale to upper edge of white roll over crista philtri Subnasale to upper edge of white roll over crista philtri Subalare to upper edge of white roll over crista philtri Subalare to upper edge of white roll over Noordhoff point Cheilion to crista philtri Cheilion to Noordhoff point Crista philtri to crista philtri Height of dry red lip at crista philtri Height of dry red lip at crista philtri Height of dry red lip at Noordhoff point
Anthropometric Notations al-al′ sn-sn′
ch-ch′ sn-cphi sn-cphi′ sbal-cphi sbal′-cphi′′ ch-cphi ch′-cphi′′ cphi-cphi′
iCL, incomplete cleft lip; cCL, complete cleft lip. *Corresponding to Figure 3.
533e
Plastic and Reconstructive Surgery • April 2014 repairs over the course of 6 consecutive months and both the resident and dental student to recognize landmarks on three-dimensional images. All images used in this study were collected for clinical purposes and none had landmarks placed at the time of image capture. To determine factors that may impact the ability of a rater to confidently measure dimensions on a three-dimensional image, raters made notes of image limitations and measurement difficulties for each subject at each measurement session. Statistical Analysis Precision (intrarater reliability) and fidelity (interrater reliability) were evaluated by three statistics (Fig. 2): (1) intraclass correlation coefficients with 95 percent confidence intervals (high reliability when close to 1 and low reliability when close to 0); (2) mean difference (the mean absolute difference between measurements); and (3) proportional difference (the mean absolute difference divided by the grand mean for a given variable). Accuracy (intermethod reliability) assessed the surgeon’s direct measurements against the surgeon’s indirect measurements (first session). Pearson correlation coefficients with 95 percent confidence intervals were estimated by bootstrapping methods (values between 0 and 1, with higher estimates indicating higher reliability),
mean difference, and proportional difference. Analyses were performed using Stata version 9.0 (Stata Corp., College Station, Texas).
RESULTS Demographics Participants in this study demonstrated typical cleft characteristics (Tables 2 and 3). At our Craniofacial Center, cleft lip repair is typically performed when patients are between 3 and 6 months of age. Two children underwent repair at 10 and 12 months of age because of associated medical conditions. Two participants established care in our Center at a late age after international adoption and underwent cleft lip repair at 24 months of age. Intrarater Reliability Intrarater reliability estimates are summarized in Table 4. The mean difference between sessions in the measured distances was always less than 1 mm. The proportional difference was less than 0.1 for most measured distances but rose to just over 0.25 for the five measurements that involved dimensions less than 4 mm (nostril remnants, measurements 7M and 7L; cleft lip remnant height, measurement 3; and vermilion heights, measurements 16, 17, and 18).
Fig. 4. Close-up of example three-dimensional image. The vermilion-mucosal border is difficult to visualize. Data holes, artifacts, and loss of resolution within cleft nostril make measurements of nostril remnants difficult.
534e
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry Table 2. Demographic and Clinical Characteristics of the 26 Patients with Cleft Lip Who Were Included in This Study Characteristics
No. (%)
Demographic Age, mo Mean Median Range Sex Male Race/ethnicity Caucasian Asian Latino African American Clinical Side of cleft Right Left Type of cleft Microform Incomplete Complete with Simonart band Complete Associated syndromes No Yes
6.9 5.5 3–24 14 (54) 4 (15) 2 (23) 1 (4) 1 (4) 15 (58) 11 (42) 0 (0) 12 (46) 4 (15) 10 (38) 24 (92) 2 (8)*
*Popliteal pterygium and craniofacial microsomia.
Correlation coefficients were greater than 0.75 for most measurements. However, estimates were 0.4 to 0.75 for measurements of small dimensions,
including columellar width (measurement 8, sn-sn′), cleft nostril height (measurement 5), cleft nostril remnants (measurements 7M and 7L), cleft lip remnant height (measurement 3), vermilion heights (measurements 16, 17, 18), and Cupid’s bow width (measurement 9, cphi-cphi′). Estimates of the correlation coefficient were less than 0.4 for the medial cleft vermilion height (measurement 17) for one rater; however, the mean difference was only 0.5 mm, which suggests that this result was related to the small dimension of the measurement. The mean difference between sessions in measured angles (columellar deviation, 2) was always less than 6 degrees. The proportional difference was approximately 0.15 and the correlation coefficient was approximately 0.9. Interrater Reliability Estimates for the interrater reliability were similar (Table 4). The mean difference between raters was less than 1 mm for most measurements, with a maximum difference of 1.3 mm for the larger measurements of lip width (measurement 20, ch-ch′) and nose width (measurement 19, al-al′) (mean dimension, 33 and 29 mm, respectively). The proportional difference was close to 0.1 for most measured distances but ranged between 0.15 and 0.32 for measurements that involved small dimensions
Table 3. Cleft Characteristics of the 26 Patients with Cleft Lip According to a Modified LAHSHAL Notation Subject No.* 3 14 13 16 19 21 26 5 1 24 12 9 23 25 11 15 20 8 10 18 17 22 6 7 2 4
LAHSHAL Classification Si
+
L
A
l l l l l l l l L L L L L L
a a a a a a A A A A A A
H
S
H
A
L
H H H H H H h h
A A A A A A a a A A a a
L L L L L L l l L L l l
Si
s h h H H H
S S S S S S S S S S S
+ + +
Si, Simonart band; L, lip; A, alveolus; H, hard palate; S, soft palate. *Subject number is in order of consecutive enrollment.
535e
Plastic and Reconstructive Surgery • April 2014 Table 4. Intrarater and Interrater Reliability Measurement Intrarater reliability† Nasal width Columellar deviation (degrees) Columellar width Nostril height Noncleft Cleft (iCL only) Nostril width Noncleft Cleft Cleft nostril remnant Medial (cCL only) Lateral (cCL only) Lip width Cleft lip remnant height (iCL only) Medial lip height Noncleft Cleft Lateral lip height Noncleft Cleft Lateral lip length Noncleft Cleft Cupid’s bow width Vermilion height Noncleft Medial cleft Lateral cleft Interrater reliability‡ Nasal width Columellar deviation (degrees) Columellar width Nostril height Noncleft Cleft (iCL only) Nostril width Noncleft Cleft Cleft nostril remnant Medial (cCL only) Lateral (cCL only) Lip width Cleft lip remnant height (iCL only) Medial lip height Noncleft Cleft Lateral lip height Noncleft Cleft Lateral lip length Noncleft Cleft Cupid’s bow width Vermilion height Noncleft Medial cleft Lateral cleft
Mean (mm)*
SD
Mean Difference
Proportional Difference
Correlation Coefficient (95% CI)
29.4 30 4.5
2.7 14.2 0.9
0.4 4.4 0.5
0.01 0.15 0.11
0.99 (0.98–1.00) 0.93 (0.88–0.98) 0.69 (0.50–0.89)
5.9 5.6
0.8 0.6
0.2 0.3
0.03 0.05
0.92 (0.87–0.98) 0.74 (0.51–0.98)
6.7 11.8
1.2 3.3
0.3 0.7
0.04 0.06
0.95 (0.91–0.99) 0.97 (0.95–0.99)
3 3.1 34 2.5
0.6 1.2 4.5 1.2
0.7 0.6 0.6 0.4
0.23 0.19 0.02 0.16
0.44 (0.00–0.87) 0.74 (0.51–0.98) 0.98 (0.97–1.00) 0.93 (0.85–1.00)
8.3 5
1.4 1.2
0.6 0.4
0.07 0.08
0.87 (0.78–0.96) 0.90 (0.83–0.97)
9.4 7.2
1.7 1.7
0.5 0.5
0.05 0.07
0.93 (0.88–0.98) 0.91 (0.85–0.98)
17.4 14.8 6.3
2.6 2 1
0.7 0.6 0.6
0.04 0.04 0.1
0.93 (0.88–0.98) 0.92 (0.86–0.98) 0.76 (0.60–0.92)
3.4 2 3
0.7 0.6 0.7
0.5 0.4 0.6
0.15 0.2 0.2
0.69 (0.49–0.88) 0.46 (0.17–0.76) 0.59 (0.35–0.84)
28.8 35.6 4.4
2.8 14.3 0.9
1.3 6.9 0.5
0.05 0.19 0.11
0.86 (0.77–0.94) 0.82 (0.71–0.92) 0.77 (0.65–0.90)
5.8 5.5
0.8 0.7
0.4 0.4
0.07 0.07
0.81 (0.71–0.92) 0.63 (0.37–0.89)
6.8 11.5
1.1 3.2
0.4 0.8
0.06 0.07
0.90 (0.83–0.96) 0.94 (0.90–0.98)
3 2.8 33.4 2.6
1.1 1.2 3.9 0.9
0.7 0.9 1.2 0.7
0.23 0.32 0.04 0.27
0.61 (0.34–0.88) 0.65 (0.40–0.90) 0.94 (0.91–0.98) 0.71 (0.50–0.93)
8.7 5.4
1.4 1.5
0.8 0.7
0.09 0.13
0.77 (0.65–0.90) 0.75 (0.61–0.88)
10 7.4
1.8 1.6
0.8 0.9
0.08 0.12
0.86 (0.78–0.94) 0.74 (0.61–0.88)
16.9 14.4 7.4
2.3 1.6 1.1
0.9 1 1
0.05 0.07 0.14
0.89 (0.83–0.96) 0.72 (0.58–0.87) 0.34 (0.10–0.58)
3.4 2.3 2.6
0.7 0.5 0.5
0.5 0.5 0.5
0.15 0.22 0.19
0.64 (0.46–0.82) 0.48 (0.26–0.70) 0.52 (0.30–0.73)
iCL, incomplete cleft lip; cCL, complete cleft lip. *All linear measurements were in millimeters and all angular measurements were in degrees. †The intrarater reliability estimates for one of the three raters (the estimates were qualitatively similar for each of the three raters). ‡The correlation between the three raters during the first rating session.
(e.g., nostril remnants, measurements 7M and 7L; cleft lip remnant height, measurement 3; and vermilion heights, measurements 16, 17, and 18).
536e
Correlation coefficients were greater than 0.75 for most measurements. However, estimates for measurements of relatively small dimensions were 0.4
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry to 0.75 (e.g., columellar width, measurement 8, sn-sn′; cleft nostril height, measurement 5; cleft lip remnant height, measurement 3; and vermilion height, measurements 16, 17, and 18). Correlation coefficients were less than 0.4 for the cleft medial nostril remnant and the Cupid’s bow width. The mean difference between raters in the columellar deviation angle (measurement 2) was less than 8 degrees. The proportional difference was approximately 0.2 and the correlation coefficient was 0.8. Intermethod Reliability Correlation coefficients were greater than 0.75 for nasal width, columellar deviation, nostril widths, lip width, and noncleft lateral lip height. They were 0.4 to 0.75 for columellar width, nostril heights, medial lip heights, cleft lateral lip height, and lateral lip lengths. The mean difference between direct and indirect measurements was less than 1.5 mm for most measurements, with a maximum difference of 3.9 mm for lip width (measurement 20, ch-ch′), corresponding to a 10 percent proportional difference. Similar to intrarater and interrater reliability, the intermethod reliability was less than 0.4 for the cleft nostril remnants (measurements 7M and 7L), the vermilion heights (measurements 16, 17, and 18), and the Cupid’s bow width (measurement 9, cphi-cphi′). The mean difference between direct and indirect measurements was within 1 mm for most measurements and up to 1.9 for the Cupid’s bow width (measurement 9, cphi-cphi′).
Qualitative Observations The raters recorded the following findings that impacted their ability to confidently place landmarks: (1) some three-dimensional images included data holes and artifacts within the nostrils or the cleft (Fig. 4); (2) dry vermilion was often difficult to differentiate from wet mucosa; (3) mesh and texture resolution was better for older children; and (4) images of children often displayed some degree of facial animation (Fig. 5). To subjectively evaluate image quality in younger compared with older infants, we systematically reviewed the three-dimensional images for all subjects, comparing their initial three-dimensional images at the time of presentation (usually at 1 to 2 weeks of age) to those captured as part of this preoperative study protocol. Form and texture definition were better with older age. We also noted that 24 of the 26 study images were captured with a child who was awake.
DISCUSSION Treatment of children with unilateral cleft lip +/– palate spans childhood; occurs over multiple stages; requires collaboration among multiple specialists; and is associated with variable outcomes related to patient characteristics, surgical technique, and surgical provider. Understanding the relative effect of individual treatments, techniques, or protocols relies on accurate and objective documentation of form over the course of a child’s treatment. Most cleft centers use serial two-dimensional photographs to record facial changes. However, measurements performed on such images are
Fig. 5. Comparison of facial animation with direct and indirect measurements. Direct measurements were performed under anesthetic immediately before cleft lip repair. Three-dimensional images were captured in the clinic at the preoperative visit, generally, with an awake infant. Animation distorts nasolabial features.
537e
Plastic and Reconstructive Surgery • April 2014 prone to error secondary to variations in camera angle relative to the subject (i.e., parallax).17 Direct anthropometric measurement eliminates parallax, and Farkas published valuable reference data using this approach; however, this method is time consuming and difficult to perform on awake infants. Reports of direct measurements on children younger than 1 year are rare18–22 and involve limited select facial dimensions23 or require an asleep baby with10,24 or without19 anesthetic. Nasolabial casts provide a three-dimensional record of form25,26; however, taking impressions on an awake child is burdensome, time-intensive, and involves risks of aspiration. Infants are invariably distressed when the impressions are taken, and this leads to distortion of facial form. Finally, storage and management of casts can be costly. Three-dimensional laser scanning can produce accurate digital three-dimensional images27 for reliable anthropometric analysis.28 Such data have been used to study facial morphology29,30 and postoperative outcomes of cleft surgery.31,32 However, the relatively long image acquisition time is a barrier to using the technology with infants, and image quality has been found to vary inversely with subject age.30 Three-dimensional optical scanning is faster and can be applied to capture facial form in older children.33–36 A recent pilot study has demonstrated reliable anthropometric measurements in infants with unilateral cleft lip +/– palate37; however, the set of dimensions evaluated was limited and does not include all of the dimensions of interest that we have examined. In addition, the lack of texture resolution on the generated images limits the application of this type of imaging. Three-dimensional stereophotogrammetry has advantages related to its rapid image capture speed and noninvasive nature. Compared with other methods, there is no burden and there is no distortion from parallax. The accuracy of three-dimensional stereophotogrammetry has been documented on mannequin heads,38 on normal adults,1,2 and on children and adults with noncleft facial dysmorphology.3,4 However, the reliability of three-dimensional stereophotogrammetry has yet to be established directly on infants who have unrepaired unilateral cleft lip +/– palate. Given the young age, small dimensions, and unique form of the unrepaired unilateral cleft lip +/– palate, the accuracy of three-dimensional stereophotogrammetry cannot be assumed in this population based on previous reports. Our study examined the reliability of a detailed set
538e
of anthropometric measurements that are used specifically for cleft lip repair.11–13 In contrast to other studies, these measurements have not been evaluated on a three-dimensional model before, and we measured reliability within the context of a standard clinical routine. We used the entire data set and did not exclude any image or subject. As such, measured variations reflect a typical clinical scenario rather than a highly selected data set in which lower quality images or more difficult dimensions to measure have been excluded. We found that most but not all of the 21 measurements were reliable. Measurements involving vermilion heights, Cupid’s bow width, and nostril remnants were less reliable. The mean difference in vermilion heights was on the order of 0.5 mm; however, with dimensions of 2 to 4 mm, the proportional difference was large, and this was reflected in the Pearson correlation coefficients. Vermilion height can be challenging to measure, as the junction of dry vermilion and wet mucosa can be subtle. Without tactile feedback or the ability to change light reflections with tissue movement, this junction becomes even more difficult to identify on conventional two-dimensional photography. Although a three-dimensional image can be rotated as a whole, the lip cannot be moved for exposure, and we found that the texture resolution of three-dimensional images was limited. As a result, measurements of vermilion height and dimensions based on the Noordhoff point (cphi″; identified by the maximal vermilion height) are more prone to error than other measurements. Reliable measurement of Cupid’s bow width can be difficult because of flattening of the curvature of Cupid’s bow peak, especially on the cleft side. Some surgeons use this lack of definition to their advantage for cleft lip repair by cheating the point of closure medially or laterally.39 Given the limitations of texture resolution on three-dimensional images, we were not surprised by variations in measurements based on chpi′. Nostril remnants along the nasal sill on either side of a cleft can be better exposed by manipulation of the lip and nose. On static three-dimensional images, these shelves can be difficult to visualize, with data holes, artifacts, and loss of data resolution that are often present in these dark areas (Fig. 4). The mean difference in measurement of these areas was within 1 mm; however, the small dimensions of 2 to 3 mm made
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry the proportional differences large and consequently the correlation coefficients unfavorable. Some degree of intermethod variation was expected given that direct anthropometric measurements were performed while subjects were asleep under general anesthesia immediately before cleft lip repair, whereas the three-dimensional images were captured while subjects were awake during their preoperative clinic visit (Fig. 5). Although we attempted to capture infants with a neutral facial expression, most awake infants demonstrate some degree of animation during image acquisition. On review of the three-dimensional images, all but two appeared to have been captured while the subject was awake. Anthropometric measurements are prone to stretch and tone of the face with animation, especially those based off of the cheilion. Consistent with this, all proportional differences were higher, and intermethod correlation values were lower, especially those measured from landmarks prone to movement associated with animation (such as lip width, lateral lip lengths, and lateral lip heights). Capturing images while subjects are asleep would more closely mimic our reference standard using direct measurements; however, doing so during a routine clinic visit would be difficult and impractical. Three-dimensional image capture under anesthesia would potentially have also improved our ability to measure intermethod correlation; however, bringing the three-dimensional system into the operating room is not practical and would deviate from our current clinical practice. For the same reasons, we could not determine the senior surgeon’s variance performing direct measurements. We assumed that direct measurement would have higher reliability, as the tactile feedback and manipulation of tissues improves visualization and recognition of anatomical landmarks (especially vermilion, nostril remnants, and Cupid’s bow). These differences likely contribute to the observed intermethod variability. Three-dimensional stereophotogrammetry has been used to measure normal facial morphology,40 to compare children with unrepaired cleft lip to normal controls,41–43 and to document form after orthodontic44,45 and surgical17,46–49 treatment of unilateral cleft lip +/– palate. Although use of this technology is attractive and becoming more popular, understanding the benefits and limitations in infants with unrepaired unilateral cleft lip +/– palate is important: 1. For gross nasolabial dimensions measured in previous studies,9,23,50 we found that the reliability was very good.
2. For measurements that are important descriptors of cleft severity,10 such as columellar angle and nostril width ratio, we found that the reliability was similarly very good. 3. Certain measurements are key in the design of cleft lip repair, and these were found to be reliable. The relative medial lip heights define the degree to which the lip needs to be corrected by means of rotation (i.e., rotation-advancement approaches), back-cut (i.e., inferior triangle approaches), or Rose-Thompson effect. The lateral lip heights and lengths define the amount of tissue available for medial augmentation by means of advancement and/or laterally based flaps. Fisher has demonstrated how these lateral lip dimensions can have significant impacts depending on the design used for repair.13 Ultimately, all of these measurements define the “defect” to be reconstructed and are relevant to surgical outcomes. 4. Other measurements that can be used in cleft lip repair are subtler and in our study were found to be less reliable. Relative vermilion heights determine the augmentation necessary to avoid leaving chronically exposed mucosa at the site of repair,14,51–53 and the size of the nostril remnants impacts on the design for appropriate nasal sill and floor closure.51 Cleft surgeons have an obligation to audit their results, and three-dimensional images provide a potentially powerful means of doing so. Farkas provided us with a framework to objectively analyze facial form; however, the burden of performing measurements has limited the use of this approach. Three-dimensional imaging may allow widespread use so that we can be more objective in analyzing our outcomes. Our center’s imaging technologist captures three-dimensional images in conjunction with standard two-dimensional photographic records according to our center’s protocol. Once the three-dimensional system is calibrated (which requires 10 minutes at the start of clinic) and a child is positioned,9 additional three-dimensional image capture is as rapid as conventional photography. The generated three-dimensional model can be moved and rotated in much the same way as we appreciate facial form by varying our viewing angles and perspectives. The set of anthropometric measurements used in this study takes 5 to 10 minutes for
539e
Plastic and Reconstructive Surgery • April 2014 Table 5. Intermethod Reliability Mean Measurement (mm)* Measurement Nose Nasal width Columellar deviation (degrees) Columellar width Nostril height Noncleft Cleft (iCL only) Nostril width Noncleft Cleft Cleft nostril remnant Medial (cCL only) Lateral (cCL only) Lip Lip width Cleft lip remnant height (iCL only) Medial lip height Noncleft Cleft Lateral lip height Noncleft Cleft Lateral lip length Noncleft Cleft Cupid’s bow width Lateral lip length Noncleft Medial cleft Lateral cleft
Indirect 3D
Direct Caliper
Mean Difference
Proportional Difference
Correlation Coefficient (95% CI)
29.57 29.75 4.76
30.42 27.31 4.97
1.31 8.91 0.49
0.04 0.33 0.1
0.9 (0.8–1.0) 0.8 (0.6–0.9) 0.7 (0.6–0.9)
5.9 5.65
5.87 4.69
0.72 1
0.12 0.21
0.4 (0.1–0.7) 0.5 (0.0–0.9)
6.44 11.33
5.6 10.71
0.89 0.85
0.16 0.08
0.9 (0.8–1.0) 1.0 (0.9–1.0)
2.97 3.01
3.61 4.36
0.84 1.66
0.23 0.38
0.3 (−0.3–0.9) 0.3 (−0.4–0.9)
34.23 2.55
38.08 2.98
3.88 1.1
0.1 0.37
0.8 (0.6–0.9) 0.3 (−0.4–0.9)
8.45 5.18
7.99 4.73
1.12 0.8
0.14 0.17
0.4 (0.2–0.7) 0.6 (0.4–0.9)
9.63 7.36
9.25 7.77
0.94 1.27
0.1 0.16
0.8 (0.6–0.9) 0.6 (0.4–0.8)
17.65 14.89 6.48
17.93 15.02 8.08
1.53 1.79 1.86
0.09 0.12 0.23
0.7 (0.5–0.9) 0.5 (0.2–0.8) 0.0 (−0.3–0.4)
3.42 2.05 3.08
4.33 2.28 4.11
0.95 0.66 1.11
0.22 0.29 0.27
0.7 (0.4–0.9) 0.2 (−0.3–0.7) 0.5 (0.2–0.8)
iCL, incomplete cleft lip; cCL, complete cleft lip. *All linear measurements were in millimeters and all angular measurements were in degrees.
an experienced rater and can be performed at any time once the images have been stored. In addition, we have found the three-dimensional model to be helpful for preoperative planning and for resident teaching. In spite of the advantages of three-dimensional imaging, current systems are expensive and cannot replace standardized two-dimensional photography.54 Limitations in texture resolution on three-dimensional systems compared with two-dimensional images may not allow for reliable scar assessment. Although our experience using three-dimensional technology for longitudinal assessment is limited, we adhere to a standard protocol for image acquisition9 so that future comparisons can be made. The three-dimensional system used in our study is optimized for image capture of the pediatric face and head in the setting of a busy craniofacial center. Although resolution may be improved using a different configuration dedicated to capture of facial features only, the need to recalibrate the system makes multiple configurations in a single clinic impractical. Our measures of reliability could also be improved by rejecting lower
540e
quality images; however, we felt that evaluation of this system in a normal clinical setting would be important. Just as digital photography has replaced conventional film-based photography at many institutions, with further refinements, three-dimensional imaging technology may one day replace two-dimensional photography. Until then, it is important to understand the limitations of this technology.
CONCLUSIONS Three-dimensional stereophotogrammetry provides a rapid-capture, noninvasive, low-burden method of documenting nasolabial form without problems of parallax. The current report provides an evaluation of the many features that can be reliably captured using this method, and those that may not. Three-dimensional stereophotogrammetry is a valuable tool that may contribute to the current tools used for comprehensive documentation of prerepair unilateral cleft lip +/– palate form.
Volume 133, Number 4 • Three-Dimensional Stereophotogrammetry Raymond Tse, M.D. Division of Plastic and Craniofacial Surgery Department of Surgery Seattle Children’s Hospital 4800 Sand Point Way NE M/S OB.9.527 Seattle, Wash. 98105 [email protected]
ACKNOWLEDGMENT
This work was supported by the University of Washington Dental Alumni Fund. PATIENT CONSENT
Parents or guardians provided written consent for the use of patients’ images. REFERENCES 1. Weinberg SM, Scott NM, Neiswanger K, Brandon CA, Marazita ML. Digital three-dimensional photogrammetry: Evaluation of anthropometric precision and accuracy using a Genex 3D camera system. Cleft Palate Craniofac J. 2004;41:507–518. 2. Wong JY, Oh AK, Ohta E, et al. Validity and reliability of craniofacial anthropometric measurement of 3D digital photogrammetric images. Cleft Palate Craniofac J. 2008;45:232–239. 3. Heike CL, Cunningham ML, Hing AV, Stuhaug E, Starr JR. Picture perfect? Reliability of craniofacial anthropometry using three-dimensional digital stereophotogrammetry. Plast Reconstr Surg. 2009;124:1261–1272. 4. Aldridge K, Boyadjiev SA, Capone GT, DeLeon VB, Richtsmeier JT. Precision and error of three-dimensional phenotypic measures acquired from 3dMD photogrammetric images. Am J Med Genet A 2005;138:247–253. 5. Ayoub A, Garrahy A, Hood C, et al. Validation of a vision-based, three-dimensional facial imaging system. Cleft Palate Craniofac J. 2003;40:523–529. 6. Kriens OE. LAHSHAL: A concise documentation system for cleft lip, alveolus and palate diagnosis. In: Kriens OE, ed. What Is a Cleft Lip and Palate? A Multidisciplinary Update. New York: Thieme; 1989:30–34. 7. Onizuka T, Hosaka Y, Aoyama R, Takahama H, Jinnai T, Usui Y. Operations for microforms of cleft lip. Cleft Palate Craniofac J. 1991;28:293–300; discussion 300. 8. Semb G, Shaw WC. Simonart’s band and facial growth in unilateral clefts of the lip and palate. Cleft Palate Craniofac J. 1991;28:40–46; discussion 46. 9. Heike CL, Upson K, Stuhaug E, Weinberg SM. 3D digital stereophotogrammetry: A practical guide to facial image acquisition. Head Face Med. 2010;6:18. 10. Fisher DM, Tse R, Marcus JR. Objective measurements for grading the primary unilateral cleft lip nasal deformity. Plast Reconstr Surg. 2008;122:874–880. 11. Noordhoff MS. The Surgical Technique for the Unilateral Cleft Lip-Nasal Deformity. Taipei: Noordhoff Craniofacial Foundation; 1997. 12. Tse R. Unilateral cleft lip: Principles and practice of surgical management. Semin Plast Surg. 2012;26:145–155. 13. Boorer CJ, Cho DC, Vijayasekaran VS, Fisher DM. Presurgical unilateral cleft lip anthropometrics: Implications
for the choice of repair technique. Plast Reconstr Surg. 2011;127:774–780. 14. Noordhoff MS. Reconstruction of vermilion in unilateral and bilateral cleft lips. Plast Reconstr Surg. 1984;73:52–61. 15. Losee JE, Selber JC, Arkoulakis N, Serletti JM. The cleft lateral lip element: Do traditional markings result in secondary deformities? Ann Plast Surg. 2003;50:594–600. 16. Cutting CB, Dayan JH. Lip height and lip width after extended Mohler unilateral cleft lip repair. Plast Reconstr Surg. 2003;111:17–23; discussion 24. 17. Oh TS, Choi JW, Koh KS. Upper lip asymmetry perception using three-dimensional anthropometry in patients with unilateral cleft lip deformity. J Craniofac Surg. 2011;22:2080–2083. 18. Farkas LG, Posnick JC, Hreczko TM, Pron GE. Growth patterns of the nasolabial region: A morphometric study. Cleft Palate Craniofac J. 1992;29:318–324. 19. Cho BC, Kim JY, Yang JD, Chung HY, Park JW, Hwang JH. Anthropometric study of the upper lip and the nose of infants less than a year of age. J Craniofac Surg. 2006;17:57–61. 20. Kim SK, Lee JH, Lee KC, Park JM. Mulliken method of bilateral cleft lip repair: Anthropometric evaluation. Plast Reconstr Surg. 2005;116:1243–1251; discussion 1252. 21. Yeow VK, Huang MH, Lee ST, Fook Chong SM. An anthropometric analysis of indices of severity in the unilateral cleft lip. J Craniofac Surg. 2002;13:68–74. 22. Vegter F, Hage JJ. Facial anthropometry in cleft patients: A historical appraisal. Cleft Palate Craniofac J. 2001;38:577–581. 23. Farkas LG, Hajnis K, Posnick JC. Anthropometric and anthroposcopic findings of the nasal and facial region in cleft patients before and after primary lip and palate repair. Cleft Palate Craniofac J. 1993;30:1–12. 24. Mulliken JB, Burvin R, Farkas LG. Repair of bilateral complete cleft lip: Intraoperative nasolabial anthropometry. Plast Reconstr Surg. 2001;107:307–314. 25. Xu Y, Li J, Zhao S, Shi B, Zheng Q, Wang Y. Accuracy of a plastic facial cast fabricated with a custom tray in comparison with cone beam computed tomography. Oral Surg Oral Med Oral Pathol Oral Radiol. 2014;117:238–245. 26. Germec-Cakan D, Canter HI, Nur B, Arun T. Comparison of facial soft tissue measurements on three-dimensional images and models obtained with different methods. J Craniofac Surg. 2010;21:1393–1399. 27. Baik HS, Lee HJ, Lee KJ. A proposal for soft tissue landmarks for craniofacial analysis using 3-dimensional laser scan imaging. World J Orthod. 2006;7:7–14. 28. Toma AM, Zhurov A, Playle R, Ong E, Richmond S. Reproducibility of facial soft tissue landmarks on 3D laser-scanned facial images. Orthod Craniofac Res. 2009;12:33–42. 29. Mori A, Nakajima T, Kaneko T, Sakuma H, Aoki Y. Analysis of 109 Japanese children’s lip and nose shapes using 3-dimensional digitizer. Br J Plast Surg. 2005;58:318–329. 30. Schwenzer-Zimmerer K, Chaitidis D, Boerner I, et al. Systematic contact-free 3D topometry of the soft tissue profile in cleft lips. Cleft Palate Craniofac J. 2008;45:607–613. 31. Nakamura N, Okawachi T, Nishihara K, Hirahara N, Nozoe E. Surgical technique for secondary correction of unilateral cleft lip-nose deformity: Clinical and 3D observations of preoperative and postoperative nasal forms. J Oral Maxillofac Surg. 2010;45:2248–22570. 32. Schwenzer-Zimmerer K, Chaitidis D, Berg-Boerner I, et al. Quantitative 3D soft tissue analysis of symmetry prior to and after unilateral cleft lip repair compared with non-cleft persons (performed in Cambodia). J Craniomaxillofac Surg. 2008;36:431–438. 33. Bilwatsch S, Kramer M, Haeusler G, et al. Nasolabial symmetry following Tennison-Randall lip repair: A three-dimensional
541e
Plastic and Reconstructive Surgery • April 2014 approach in 10-year-old patients with unilateral clefts of lip, alveolus and palate. J Craniomaxillofac Surg. 2006;34:253–262. 34. Nkenke E, Lehner B, Kramer M, et al. Determination of facial symmetry in unilateral cleft lip and palate patients from threedimensional data: Technical report and assessment of measurement errors. Cleft Palate Craniofac J. 2006;43:129–137. 35. Nkenke E, Langer A, Laboureux X, et al. Validation of in vivo assessment of facial soft-tissue volume changes and clinical application in midfacial distraction: A technical report. Plast Reconstr Surg. 2003;112:367–380. 36. Stauber I, Vairaktaris E, Holst A, et al. Three-dimensional analysis of facial symmetry in cleft lip and palate patients using optical surface data (in German). J Orofac Orthop. 2008;69:268–282. 37. Li G, Wei J, Wang X, et al. Three-dimensional facial anthropometry of unilateral cleft lip infants with a structured light scanning system. J Plast Reconstr Aesthet Surg. 2013;66:1109–1116. 38. Weinberg SM, Naidoo S, Govier DP, Martin RA, Kane AA, Marazita ML. Anthropometric precision and accuracy of digital three-dimensional photogrammetry: Comparing the Genex and 3dMD imaging systems with one another and with direct anthropometry. J Craniofac Surg. 2006;17:477–483. 39. Koh KS, Hwang CH, Kim EK. Modified design of Cupid’s bow in the repair of unilateral microform cleft lip. J Craniofac Surg. 2009;20:1367–1369. 40. Yamada T, Mori Y, Minami K, Mishima K, Tsukamoto Y. Three-dimensional analysis of facial morphology in normal Japanese children as control data for cleft surgery. Cleft Palate Craniofac J. 2002;39:517–526. 41. Hood CA, Hosey MT, Bock M, White J, Ray A, Ayoub AF. Facial characterization of infants with cleft lip and palate using a three-dimensional capture technique. Cleft Palate Craniofac J. 2004;41:27–35. 42. Krimmel M, Kluba S, Bacher M, Dietz K, Reinert S. Digital surface photogrammetry for anthropometric analysis of the cleft infant face. Cleft Palate Craniofac J. 2006;43:350–355. 43. Zreaqat M, Hassan R, Halim AS. Facial dimensions of Malay children with repaired unilateral cleft lip and palate: A three dimensional analysis. Int J Oral Maxillofac Surg. 2012;41:783–788.
542e
44. Singh GD, Levy-Bercowski D, Santiago PE. Three-dimensional nasal changes following nasoalveolar molding in patients with unilateral cleft lip and palate: Geometric morphometrics. Cleft Palate Craniofac J. 2005;42:403–409. 45. Simanca E, Morris D, Zhao L, Reisberg D, Viana G. Measuring progressive soft tissue change with nasoalveolar molding using a three-dimensional system. J Craniofac Surg. 2011;22:1622–1625. 46. Devlin MF, Ray A, Raine P, Bowman A, Ayoub AF. Facial symmetry in unilateral cleft lip and palate following alar base augmentation with bone graft: A three-dimensional assessment. Cleft Palate Craniofac J. 2007;44:391–395. 47. Ayoub A, Garrahy A, Millett D, et al. Three-dimensional assessment of early surgical outcome in repaired unilateral cleft lip and palate: Part 1. Nasal changes. Cleft Palate Craniofac J. 2011;48:571–577. 48. Ayoub A, Garrahy A, Millett D, et al. Three-dimensional assessment of early surgical outcome in repaired unilateral cleft lip and palate: Part 2. Lip changes. Cleft Palate Craniofac J. 2011;48:578–583. 49. Singh GD, Levy-Bercowski D, Yáñez MA, Santiago PE. Three-dimensional facial morphology following surgical repair of unilateral cleft lip and palate in patients after nasoalveolar molding. Orthod Craniofac Res. 2007;10:161–166. 50. Farkas LG, Lindsay WK. Morphology of the adult face following repair of unilateral cleft lip and palate in childhood. Plast Reconstr Surg. 1973;52:652–655. 51. Fisher DM. Unilateral cleft lip repair: An anatomical subunit approximation technique. Plast Reconstr Surg. 2005;116:61–71. 52. Mulliken JB, Pensler JM, Kozakewich HP. The anatomy of Cupid’s bow in normal and cleft lip. Plast Reconstr Surg. 1993;92:395–403; discussion 404. 53. Sitzman TJ, Fisher DM. Presurgical unilateral cleft lip anthropometrics: Incidence of vermilion height asymmetry. Plast Reconstr Surg. 2013;131:935e–937e. 54. Birgfeld CB, Saltzman BS, Luquetti DV, Latham K, Starr JR, Heike CL. Comparison of two-dimensional and three-dimensional images for phenotypic assessment of craniofacial microsomia. Cleft Palate Craniofac J. 2013;50:305–314.
