An initiative to solve a several thousand-year-old medical mystery—what causes sudden or unexplained infant deaths (SUID)—has led to new insights that recently were presented at the 6th annual SIDS Summit.
The initiative brings together the world’s leading researchers on sudden infant death syndrome (SIDS) and infant deaths with private-sector data scientists, who together mix data science, AI, and medical research techniques. Their goal is to share insights, inspire new hypotheses, and find the elusive causes of SUID, which includes SIDS. SIDS is the leading cause of death of infants 1 month to 1-year-old in the US and other developed countries.
John Kahan, the former Microsoft vice president and chief data analytics officer who co-founded The Aaron Matthew SIDS Research Foundation; Jan-Marino Ramirez, PhD, director of the Center for Integrative Brain Research at Seattle Children’s; and Juan Lavista, chief scientist and lab director at Microsoft’s AI for Good Research Lab, hosted the 6th annual SIDS Summit this spring. Attendees came from Amsterdam, Australia, Italy, Sri Lanka, Switzerland, New Zealand, Norway, the UK, and throughout the US. They presented findings across five pillars: epidemiology, pathology, physiology, genetics, and education.
Among the unique aspects of the summit were presentations given by data scientists from Microsoft on machine learning-aided studies of genetic samples from infants, alongside talks from research physicians from some of the best universities in the world.
The Aaron Matthew SIDS Research Foundation funds the first-ever SIDS genomic database, maintained at Seattle Children’s Hospital, which was also discussed at the summit. With the help of data scientists at Microsoft, who leverage AI and machine learning for the effort, the research hospital has completed data analysis on 145 samples of children lost to sudden or unexplained deaths. A study of 200 more samples is ongoing. Samples may be studied by medical researchers anywhere in the world.
“The machine learning capabilities used by some of the most successful companies in the world now allow researchers to analyze health data across disciplines, and what we’ve been able to learn about the potential causes of infant deaths is nothing short of incredible,” says Ramirez in a release. “It has led to several breakthrough discoveries and spurred new theories, not just on the causes of SIDS, but on the potential to screen infants for other fatal diseases that pose threats later in life.”
Among the breakthroughs from the research collaboration:
- Researchers now estimate that 22% of SUIDs in the US can be directly attributed to maternal smoking during pregnancy and that SIDS rates can be reduced through education programs about the risks of smoking during pregnancy.
- SUID is not a single risk factor. Infants who succumb within a few days of birth have a different genetic profile and different risk factors than those who pass away later. This finding opens doors to potential treatment and diagnostic options in the future, depending on specific genetic and environmental risk factors.
Among the new theories shared at the SIDS summit, Ramirez proposed that, based on his team’s findings, in which signs of cardiac problems were found in the cohort of infants who passed away from SIDS, early genetic testing might not only reduce SIDS but also the risk of sudden cardiac deaths later in life. Such deaths claim between 180,000 to 300,000 lives annually in the US.
Ramirez and his team continue to collaborate with data scientists from Microsoft, headquartered in the region. They also work with Seattle’s Allen Institute for Brain Science, a leader in single-cell identification. The Ramirez lab is particularly interested in identifying the neurons implicated in arousal that might be affected in SIDS.
They work with the pathology department at the University of Washington, which has a unique infrastructure and expertise in postmortem characterization, which will allow the Ramirez lab to obtain brain samples more effectively from children who died of SIDS to further characterize the pathological changes associated with it. And they interact with the genome center at the University of Washington to implement advanced genetic and single-cell characterization.
“None of us had any idea about where this would take us,” Kahan says in a release. “The power of diverse resources across data science and AI, epidemiology, physiology, pathology, and genetics is opening new avenues of research. We are confident that this work will continue to translate into prevention, better diagnostics, and potential therapies to prevent infants from succumbing to SIDS.”