Researchers at Boston Children’s Hospital can now use AI to better predict the risk of vesicoureteral reflux (VUR) in infants. The new algorithm leverages stored postnatal ultrasound data to identify patients at highest risk for VUR. As a result, burdensome diagnostic tests such as VCUG can be regularly avoided. This not only reduces the burden on patients and their families but also leads to significant cost savings. The tool analyzes factors such as gender, ureteral dilation and kidney tissue thickness to more accurately assess the risks of VUR.
Glass sphere
This innovation clearly demonstrates the potential of AI in newborn care. Accurate risk assessments, such as for the existence of VUR, improve babies’ well-being and allow treatment plans to be better tailored. This paves the way for early intervention and treatment, which is essential to prevent complications.
The developed AI tool is user-friendly and specially designed for clinical settings so that medical staff can make the right decisions. Wang and his team are now working on an even more robust model, based on 15,000 patient records, to predict the course of hydronephrosis: “We hope to build a crystal ball for urologists so that children do not have to undergo unnecessary tests,” Wang concludes.
AI and newborns
Artificial intelligence (AI) is increasingly finding its way into healthcare, especially in the most vulnerable patients: newborns. In addition to this latest innovation in early detection of urological problems in Boston, there are several innovative examples in this area. A first example can be found at the Stanford School of Medicine. An algorithm has been developed here that, based on the electronic health records of mothers and babies, can predict which newborns are at risk for health in their first two months of life. This model identifies babies who may develop serious complications of preterm birth, including chronic lung disease and retinopathy of prematurity, before these symptoms arise.
A second example comes from a researcher at the Utrecht Medical Center. They developed an AI model that can predict the long-term development of babies born much too early. This model estimates, among other things, the risk of developing intellectual disabilities and other complications. The findings of this Utrecht study have been published in The Lancet Digital Health.
