Multisystem inflammatory syndrome in children, or MIS-C, is a complex syndrome associated with SARS-CoV-2, the virus that causes COVID-19. Children have presented with diarrhea, vomiting and severe cardiovascular problems in addition to respiratory problems. Other symptoms include conjunctivitis, skin rash, swollen hands or feet, cracked lips and a red tongue — signs typically associated with classic Kawasaki disease, an inflammatory disease that can cause coronary artery aneurysms. Also confounding this clinical picture, some children experience neurologic symptoms such as headache, unusual sleepiness, disorientation and confusion. Needless to say, clinicians are significantly challenged when it comes to diagnosing and managing these patients, and predicting which children need hospitalization and which will become critically ill.
“Normally the clinical management of patients with this novel syndrome is very difficult — it affects multiple systems and there is a lot of overlap with different diseases and signs and symptoms that are still evolving,” says Shelby Kutty, the Helen B. Taussig professor and director of pediatric and congenital cardiology at Johns Hopkins. “Treatment at the onset and the epidemiology of this disease are still very new, presenting the need for salient prediction models for diagnosis and figuring out which patients are likely to develop new problems.”
Kutty and lead principal investigator Cedric Manlhiot, in partnership with the International Kawasaki Disease Registry (IKDR), aim to produce such predictive models for a syndrome that doesn’t act in a predictive way. How? Kutty is chair and Manlhiot is director of Johns Hopkins’ Cardiovascular Analytic Intelligence Initiative (CV-Ai²), which uses artificial intelligence and machine learning tools to turn real-time clinical data into prediction models that can help forecast patient outcomes.
“The goal of CV-Ai² is to find better ways to utilize data and design solutions for important clinical problems, which in turn are directly evaluated at the physician and patient level,” Manlhiot says of the initiative created by the Blalock-Taussig-Thomas Pediatric and Congenital Heart Center.
After submitting a proposal to the National Institutes of Health (NIH), the co-investigators were awarded a four-year, $4.8 million Rapid Acceleration of Diagnostics-Radical (RADx-rad) initiative grant. The first goal, explains Manlhiot, is to develop algorithms that clinicians can use in managing patients with MIS-C. During the first two years, patient data collected by the IKDR, a consortium of 19 hospitals participating in the research, will be used to help create the algorithms employing artificial intelligence-based models for diagnosis, treatment and outcome prediction. In the third and fourth years, the performance and clinical utility of these models will be validated in a predictive decision support system, adding real-time epidemiological surveillance data.
“Ultimately, by year four, we’ll establish an interface that can be deployed in Epic and similar electronic medical records with clinical data that can aid diagnosis and treatment,” says Kutty.
Because MIS-C has some similarities with Kawasaki disease in its presentation, the strategic partnership with the IKDR will provide an established data collection platform and make use of substantial clinical and research expertise. Along with Kutty and Manlhiot, Johns Hopkins investigators include pediatric cardiologists Allen Everett, Lasya Gaur and Benjamin Barnes.
Pointing to interest from some 40 other hospitals in the United States as well as hospitals in South America, France, Italy, the United Kingdom, Taiwan and India, Kutty concludes that this work will grow the IKDR consortium and improve care and outcomes globally for children with MIS-C.
“Based on what we have gathered so far, this should lead to a meaningful prediction for patients all over the world,” says Kutty.
Redonda Miller, Johns Hopkins Hospital president, calls the awarding of this NIH grant a much needed “game changer” for children with SARS-CoV-2 related disease.