Artificial intelligence for early diagnosis of fetal alcohol spectrum disorder

Fetal alcohol spectrum disorder is a disease caused by exposure to alcohol during pregnancy. It encompasses a large number of physical, mental, behavioural and cognitive abnormalities. The most severe form within this spectrum is Fetal Alcohol Syndrome (FAS). It is characterised by morphological malformations (especially craniofacial defects), growth retardation and cognitive, behavioural, socialisation and learning disorders (due to problems with the development of the central nervous system).

Researchers from the Hospital Clínic de Barcelona and the August Pi i Sunyer Biomedical Research Institute (IDIBAPS) in Barcelona have coordinated a study that demonstrates the efficiency of a new method that allows for the early diagnosis of fetal alcohol spectrum disorder, improving the quality of life of those affected and their families.

Specialists from the La Paz University Hospital in Madrid, the University of Barcelona (UB) and the International University of Valencia in Spain also worked on the study.

The team, led by Anna Ramos-Triguero, from IDIBAPS and the UB, has verified the effectiveness of a machine learning algorithm (a type of artificial intelligence) in the early diagnosis of fetal alcohol spectrum disorder.

The chosen algorithm was trained using sociodemographic, clinical and psychological variables, from a database made up entirely of data from children diagnosed at the Hospital Clínic de Barcelona. The researchers analysed clinical and neuropsychological data to develop models that can identify early signs of fetal alcohol spectrum disorder in children.

Machine learning algorithms (a form of artificial intelligence) can help in the early diagnosis of fetal alcohol spectrum disorder. (Illustration: Amazings / NCYT)

Diagnosing fetal alcohol spectrum disorder can sometimes be a challenge due to its similarity to other disorders, such as autism or hyperactivity. Therefore, this new tool could help achieve a faster and more accurate diagnosis, which would allow for timely interventions to improve the quality of life of those affected, avoiding erroneous diagnoses.

The results show that patients with fetal alcohol spectrum disorder have specific physical and psychological impairments, and the algorithms allow identifying patterns by fetal alcohol spectrum disorder subtype, such as Fetal Alcohol Syndrome (FAS), partial FAS (pFAS) and Alcohol-Related Neurodevelopmental Disorder (ARND).

“Preventing fetal alcohol spectrum disorder is everyone’s responsibility and involves avoiding alcohol consumption during pregnancy and from the moment a woman decides to become pregnant. There is no safe amount of alcohol to consume during pregnancy,” says Dr. Oscar García-Algar, Research and Innovation Coordinator of the Neonatology Service at Hospital Clínic.

The study is titled “Machine learning algorithms for the early diagnosis of fetal alcohol spectrum disorders.” It has been published in the academic journal Frontiers in Neuroscience. (Source: Hospital Clínic de Barcelona)