Heart monitoring reveals developmental age of premmies

Researchers at QIMR Berghofer Medical Research Institute are harnessing the power of artificial intelligence to track how premature babies are developing in neonatal intensive care, using the baby’s heart rate as a way to non-invasively estimate developmental age at the bedside.

By applying complex computer analysis to the vast amounts of data within long-duration recordings of a baby’s electrocardiogram (ECG) heart monitoring, the researchers and their international collaborators have been able to produce a simple readout of the infant’s developmental age. Their tool has been described in the journal Pediatric Research and should help to detect developmental delays, guide early intervention and measure treatment outcomes.

“Preterm babies are vulnerable to a range of complications, and it’s vital that any change in health status is caught early for the best possible outcome,” said QIMR Berghofer’s Dr Nathan Stevenson. “Small problems can develop into big problems as a baby gets older, but the right early intervention can have a significant and long-lasting impact on life.

“Comparing development [age] against a person’s true birth age is one of the best ways to track health. In the elderly, for example, being too advanced for your age is a concern. In babies, we’re worried about being too young for your age.”

Stevenson said it can be difficult to monitor for certain issues in preterm babies due to their fragility and small size. The clever thing about the new tool is that it piggybacks onto non-invasive ECG recordings that are already being gathered in the nursery which provide information on a person’s heart, breathing and even how their brain controls the heart. These measurements are also affected by medication, disease and injury.

“There’s a lot more subtle information embedded in the heart rate signal that human brings can’t extract visually,” Stevenson said. “So we use computer methods, we use artificial intelligence, to determine, to extract the real fine detail from that signal. If a single heartbeat is a letter, clinicians mostly will get the sentence, but we can get the book. Every heartbeat tells a story.”

QIMR Berghofer’s Dr Kartik Iyer, who led the research along with Stevenson, said ECG results change quite a bit in preterm babies in their earliest days and weeks as they develop and the connections between their heart and brain develop.

“By extracting subtle and diverse information from the ECG signal, we can provide an estimate of developmental age that can be different to a preterm baby’s chronological age,” Iyer said. “If their developmental age is lagging, clinicians can intervene with treatments that are personalised to the child.”

It is hoped that the tool will be in the clinic within five years, as work continues on even more complex analyses to extract a more complete picture of a baby’s wellbeing.

“Our next goal is to be able to extract information related to age, medication, disease and injury, and provide the best picture of a baby’s current health for doctors,” Stevenson said.

“The challenge is to make sure that not only do preterm infants survive, but they survive with the best possible health for their whole life. That’s where this precision medicine, this increased diagnostic information, really helps targeted clinical care.”

Image caption: Baby Reid, who was born at 28 weeks, in the NICU.