AI diagnoses heart defects in babies

An artificial intelligence (AI) developed by Zurich researchers recognises a specific heart defect in newborns. According to the researchers, it has the potential to reduce the number of missed diagnoses of pulmonary hypertension.

a baby in a newborn ward

ETH Zurich announced that it is important to recognise and treat severe cases of this disease quickly. This is because the sooner treatment is administered, the better the prospects for the newborn child. In pulmonary hypertension, the pulmonary arteries remain narrowed after birth or close again in the first few days or weeks after birth. This can lead to severe respiratory distress in newborn babies. During treatment, the arteries to the lungs are opened by administering highly concentrated oxygen.

Four out of five AI diagnoses correct

The problem: recognising this disease is very complex and requires a great deal of experience, according to ETH Zurich. To remedy this, the researchers at ETH Zurich, together with researchers from the Kuno Klinik St. Hedwig in Regensburg, Germany, trained an AI model with video recordings of cardiac ultrasound examinations of 192 newborns.

The model trained in this way was then able to suggest the correct diagnosis in around 80 to 90 per cent of cases and determine the correct severity of the disease in around 65 to 85 per cent of cases.

However, the diagnosis is still made by a human being, emphasised developer Julia Vogt from ETH Zurich. According to the professor of medical data science, it is therefore important that people can understand the criteria on which the AI makes its decisions.

Written by: sda

Photos: Pixabay

Read more