Understanding the effects of prenatal alcohol exposure using three-dimensional facial imaging.
Drinking of alcoholic beverages (Health aspects)
Drinking of alcoholic beverages (Research)
Fetal alcohol syndrome (Risk factors)
Fetal alcohol syndrome (Research)
Fetus (Effect of alcohol on)
Fetus (Risk factors)
|Publication:||Name: Alcohol Research & Health Publisher: U.S. Government Printing Office Audience: Academic; Professional Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2011 U.S. Government Printing Office ISSN: 1535-7414|
|Issue:||Date: Spring, 2011 Source Volume: 34 Source Issue: 1|
|Topic:||Event Code: 310 Science & research|
|Geographic:||Geographic Scope: United States Geographic Code: 1USA United States|
One of the (at least theoretically) most easily detectable features
of fetal alcohol syndrome (FAS) and fetal alcohol spectrum disorders
(FASD) is a distinct pattern of facial characteristics. However, in many
children prenatally exposed to alcohol, these characteristics are
expressed only subtly, making it difficult to correctly identify
children with these disorders. To date, several studies have used
conventional two-dimensional images to develop computerized programs
assisting in the identification of individuals with FAS or FASD.
However, many of the subtle features of prenatal alcohol exposure cannot
be visualized using two-dimensional images. Therefore, researchers at
the Collaborative Initiative on Fetal Alcohol Spectrum Disorders
(CIFASD; http://www.cifasd.org) have been using a special camera system
that can generate three-dimensional images, which allows them to explore
the advantages of using such images to identify subtle facial
differences between individuals who were exposed to alcohol prenatally
and individuals who were not. This approach may help investigators and
clinicians to better understand the complications that may arise from
prenatal alcohol exposure. For example, CIFASD researchers can use
facial measurements or shapes obtained from the three-dimensional images
to predict the presence of FAS, examine associations between facial
shapes and cognitive deficiencies, or better understand how the facial
growth of a person with FAS compares with facial growth in someone not
prenatally exposed to alcohol. Through an international consortium,
CIFASD has been addressing these questions in various age groups as well
as different ethnic groups.
The Three-Dimensional Camera System and Image Analysis
The camera system used to obtain the three-dimensional images consists of two pods attached at each end of a long arm that is mounted to a standard camera tripod (see figure 1). Each pod contains three cameras and two flashes so that a total of six photographs are generated. These six photographs, which eventually comprise the three-dimensional image, are obtained in 1.5 milliseconds, similar to normal flash photography. The attached laptop computer system uses special software to automatically stitch together the six photographs (see figure 2), generating the final three-dimensional image in less than 2 minutes (see figure 3). Although the camera system looks big and bulky, it can easily be taken apart and packed into two cases. As a result, CIFASD researchers have the ability to transport the camera anywhere in the world, take three-dimensional images of dozens of children within a day, and electronically transfer these images to a secure computer for analysis.
[FIGURE 1 OMITTED]
The three-dimensional images can be analyzed in several ways. One of the simpler analytic strategies is to measure the length, width, or height of various portions of the face, such as the length of the eye, width of the forehead, or height of the upper face (figure 4). CIFASD investigators have used these measurements to evaluate whether these parameters differ in people with and without prenatal alcohol exposure. The analyses found that by using a subset of these measurements as predictor variables in a statistical method called logistic regression, one can accurately classify children in a given sample into two groups: those with FAS and those who were not exposed to alcohol prenatally. For example, when studying children from Cape Town, South Africa, CIFASD researchers identified a set of measurements that could correctly classify 94 percent of children with FAS and 91 percent of children without prenatal alcohol exposure (Moore et al. 2007). Using the same statistical approach, but with a group of children from Helsinki, Finland, the researchers also identified a set of facial measurements that correctly classified 96 percent of children with FAS and 91 percent of those without prenatal alcohol exposure. Interestingly, the sets of facial measurements that best identified the respective groups differed between the South African and Finnish samples. However, in both groups small eye width was one of the parameters that helped to predict FAS (Moore et al. 2007). These results support previous observations that small eye widths are a key feature distinguishing individuals with FAS and without prenatal alcohol exposure.
[FIGURE 2 OMITTED]
CIFASD researchers now are seeking to understand why some unique facial features helped predict group membership in the South African and Finnish samples. One possibility is that these differences are caused by facial variation attributable to ethnicity. Alternatively, the differences may be related to age differences between the two samples because the South African children, on average, were much younger than the Finnish participants, and previous studies already noted that the facial characteristics of people with FAS change with age (Mutsvangwa et al. 2010; Streissguth et al. 1991).
[FIGURE 3 OMITTED]
Another way to analyze three-dimensional images is to place specific landmarks on the images and then connect the landmarks with lines that generate shapes (see figure 5) (Mutsvangwa and Douglas 2007), a method that originally was proposed by Clarren and colleagues (1987). Using an approach called morphometrics, one then can look at differences between the shapes found in the faces of children prenatally exposed to alcohol and those found in the faces of nonexposed children. To achieve this, the shapes obtained from the faces of all subjects are aligned statistically so that they then can be compared between the two groups (i.e., people with FAS and control individuals) (Douglas and Mutsvangwa 2010). Similarly, one can compare the shapes between younger and older individuals (Mutsvangwa et al. 2010). Using this approach, researchers can examine what information specific shapes can provide about an individual's face--such as the fullness of the face, slower growth in certain facial areas, or asymmetries between the two sides of the face. Thus, CIFASD researchers have found that differences exist in the shape of particular facial regions between children with FAS and controls (Klingenberg et al. 2010). In addition, certain differences in facial shapes appear to be related to performance on tests that measure cognitive function (i.e., IQ) (Wetherill et al. 2009). Findings such as these are important because they begin to allow researchers to better understand how prenatal alcohol exposure affects development, both in the face and in the brain.
Potential Applications of the Three-Dimensional Imaging System
by doctors and clinicians who specialize in the effects of prenatal alcohol exposure that the facial features of children with FAS become less obvious as the children age.
[FIGURE 4 OMITTED]
Another application currently being explored by members of CIFASD is the opportunity to study changes in the faces of younger children. To this end, CIFASD is working with a team of pediatricians at Tygerberg Hospital, a part of Stellenbosch University Medical School, South Africa, to take three-dimensional images of babies at 1 month of age and again at 12 months. The investigators hope to enroll about 1,200 babies in the study, about 600 of whom will have been prenatally exposed to alcohol. Extensive data is available on the children's prenatal alcohol exposure, which will allow the research team to study how the babies' faces change during the first year depending on the amount and timing of the prenatal alcohol exposure.
[FIGURE 5 OMITTED]
Finally, studies involving participants of differing ages, races, and ethnicities will allow the CIFASD team to devise new ways to identify children with prenatal alcohol exposure. Because it is essential for the children's prognosis to initiate interventions as early as possible, this approach using the three-dimensional camera and image analysis can lead to earlier detection of and intervention for those at greatest risk.
Preparation of this manuscript was supported by National Institutes of Health grants U01AA--014809 and P60AA--007611.
CLARREN, S.K.; SAMPSON, P.D.; LARSEN, J.; ET AL. Facial effects of fetal alcohol exposure: Assessment by photographs and morphometric analysis. American Journal of Medical Genetics 26(3):651-666, 1987. PMID: 3565480
DOUGLAS, T.S., AND MUTSVANGWA, T.E. A review of facial image analysis for delineation of facial phenotype associated with fetal alcohol syndrome. American Journal of Medical Genetics 152A(2):528-536, 2010. PMID: 20101703
KLINGENBERG, C.P.; WETHERILL, L.; ROGERS, J.; ET AL. Prenatal alcohol exposure alters the patterns of facial asymmetry. Alcohol 44(78):649-657, 2010. PMID: 20060678
MOORE, E.S.; WARD, R.E.; WETHERILL, L.F.; ET AL. Unique facial features distinguish fetal alcohol syndrome patients and controls in diverse ethnic populations. Alcoholism: Clinical and Experimental Research 31(10):1707-1713, 2007. PMID: 17850644
MUTSVANGWA, T., AND DOUGLAS, T.S. Morphometric analysis of facial landmark data to characterize the facial phenotype associated with fetal alcohol syndrome. Journal of Anatomy 210(2):209-220, 2007. PMID: 17310546
MUTSVANGWA, T.E.; MEINTJES, E.M.; VILJOEN, D.L.; AND DOUGLAS, T.S. Morphometric analysis and classification of the facial phenotype associated with fetal alcohol syndrome in 5- and 12-year old children. American Journal of Medical Genetics. Part A 152A(1):32-41, 2010. PMID: 20014122
STREISSGUTH, A.P.; AASE J.M.; CLARREN S.K.; ET AL. Fetal alcohol syndrome in adolescents and adults. JAMA: Journal of the American Medical Association 265(15): 1961--1967, 1991. PMID: 2008025
WETHERILL, L.; HOYME, H.E.; ROBINSON, L.K.; ET AL. Longitudinal changes in facial measurements from 3D images in children with heavy prenatal alcohol exposure. Alcoholism: Clinical and Experimental Research 34(Suppl. S2):101A:P361, 2010.
WETHERILL, L.; KLINGENBERG, C.P.; ROGERS, J.; ET AL. Covariation of facial shape and neurocognitive variables in fetal alcohol syndrome individuals vs controls. Alcoholism: Clinical and Experimental Research 33(Suppl. S1):131A:P482, 2009.
Leah Wetherill, M.S., is a statistician, and Tatiana Foroud, Ph.D., is a professor in the Department of Medical and Molecular Genetics, Indiana University School of Medicine, Indianapolis, Indiana.
|Gale Copyright:||Copyright 2011 Gale, Cengage Learning. All rights reserved.|