Contact us / Members' area    

Home > Abstracts > Index > 2012 > Thorup K, et al


Feasibility study using facial analysis software to document facial features associated with fetal alcohol syndrome in newborn infants

Presented at the Neonatal Society 2012 Summer Meeting (programme).

Thorup K1, Monk V1, Gourlay E1, Aiton N2

1 Brighton and Sussex Medical School, UK
2 Brighton and Sussex University Hospital Trust, UK

Background: Fetal Alcohol Syndrome (FAS) comprises a triad of growth impairment, central nervous system dysfunction and characteristic facial features. Diagnosis is complex and often not recognised at an early age. The three facial features: short palpebral fissures, smooth philtrum and thin upper lip, are unique to FAS. Clinical examination is inherently subjective and apart from palpebral fissure length, minimal reference data is available in neonates. Establishing a standardised method and normal reference range would promote an objective assessment. Earlier diagnosis would enable earlier effective interventions thus reducing the risk of adverse life outcomes. The purpose of this study was to determine the feasibility of using facial analysis software to measure the three facial features compatible with FAS in neonates. We also aimed to define the normal reference ranges of the facial features in term and appropriate for-gestational age Caucasian neonates. Ethical approval was obtained for this study.

Methods: 33 normal term Caucasian neonates and their mothers were recruited from Royal Sussex County Hospital, Brighton. Digital facial photographs were taken according to a standardized method. Mothers completed anonymous questionnaires about alcohol consumption during pregnancy. Photographs were assessed using Facial Analysis Software to obtain values for palpebral fissure length (PFL) and upper lip circularity (LC). Upper lip thinness and philtrum smoothness were ranked according to 5-point Likert-Scale.

Results: 29 infants were analysed, of which 58% were male (17 male: 12 female). Mean gestational age was 40.3 weeks (SD 1.3), mean weight 3556g (SD 525). 23 (79%) had no prenatal alcohol exposure whilst 6 had minimal exposure (1-2units/week). PFL measurements could be obtained from 21 photographs (72%) with mean of 15.6 mm (SD 1.1). Upper lip and philtrum values could be determined in 24 (83%). The mean LC was 57.2 (SD 19.8). Mean rank scores for upper lip and philtrum were 2.71 (SD 0.69) and 2.83 (SD 0.70) respectively. PFL and LC had normal distribution with Shapiro-Wilk test. Reassessment of the quality of the photographs reduced the sample size to 15. Factors affecting quality of photographs were rotation of the head > 8 (45.5%), inadequate facial expressions (54.5%), or poor focus and exposure (3%). Of the remaining 15, PFL measurements could be obtained in 9 (27%) of photographs whilst upper lip and philtrum could be assessed in 12 (36%) of photographs.

Conclusion: Identification of neonates exposed to alcohol prenatally is challenging but the use of photographs and facial analysis software could aid the recognition of an FAS phenotype. It has been possible to gain measurements of facial features in approximately 1/3 of neonates studied, showing the potential of this technique in this age-group. Adjustments to the technique to increase quality of the photographs, may improve feasibility. Further results are needed to establish reference ranges.

Corresponding author:

Astley SJ, Clarren SK. A case definition and photographic screening tool for the facial phenotype of fetal alcohol syndrome. J Pediatr. 1996; 129(1):33-41

Home / The Neonatal Society, founded 1959 / 2001 - 2018