Mohammed, a fifteen-year-old Ugandan child, seemed concerned when he met Kelsey Brown. His rheumatic heart disease had progressed to its latter stages; the disease kills an estimated 400,000 people per year worldwide. The original plan to treat the ailment with cardiac surgery was postponed by one day.
Mohammed had to sleep while sitting erect since the fluid from his heart had built up in his lungs to the point that he could no longer breathe. Brown, a cardiology fellow at Washington’s Children’s National Hospital, speculated that he was nervous about the upcoming operation.
Nevertheless, Mohammed reassured her that he was not afraid of the process. A sharp contrast to American children whose apprehension of the treatment is understandable—”He was most concerned that his heart would not be repaired and his surgery would be canceled,” Brown said.
A new AI-powered tool can detect rheumatic heart disease in patients before surgery is necessary, thanks to researchers at Children’s National. This shows the medical potential of artificial intelligence.
In collaboration with experts from the Uganda Heart Institute, the group developed a system that will allow skilled nurses to screen and diagnose children at an affordable cost of less than $1 per year for penicillin treatment. Thousands of individuals could avoid surgery if they got care sooner rather than later.
According to him, “We are concentrating on a grossly neglected global health issue that causes major [death and disease].” In agreement was Craig A. Sable, who headed Children’s National’s global health program and co-wrote a paper about the AI tool in the Journal of the American Heart Association. During his most recent November visit, Sable—a physician with more than 40 years of expertise in Uganda—reported performing eight or nine cardiac procedures weekly.
But he did mention that 1,500 kids were on a nationwide waiting list for the operation. According to him, “Only a quarter will undergo the operation before passing away.”
Limitations In Early Detection:
Rheumatic heart disease is a preventable condition that primarily arises from the cumulative impact of recurrent episodes of rheumatic fever, an infection caused by group A streptococcus bacteria. The bacterial group, which is ranked as the fifth most lethal pathogen globally, is accountable for the annual occurrence of strep throat in an estimated 600 million individuals.
A rare subset of infected individuals develops rheumatic heart disease, an abnormal immune response that causes valve damage and heart failure. According to an estimation, between 5 to 10 percent of the population possesses a genetic predisposition to this immune response.
Environmental factors, including overcrowding and substandard living conditions, also contribute. Typically, the onset of the disease’s initial symptoms in children occurs between the ages of 8 and 10. An echocardiogram, which is an ultrasound of the heart, can detect the condition in its early stages.
The issue is that there are insufficient cardiologists in countries like Uganda to interpret the images, which renders widespread screening impracticable. Consequently, the disease frequently evades detection during its initial phases.
“The child is unaware that they possess it.” “The parents are unaware,” Brown explained. “Even if they visit the physician, the physician would not hear it.”
Rheumatic heart disease is a prevalent ailment in impoverished nations like Uganda, where its prevalence among children ranges from 2 to 3 percent. Numerous strains of the bacteria reside in these nations, and children frequently contract recurrent infections. Each new episode exacerbates the condition.
Even though the sickness isn’t a big deal now, Sable knows firsthand how serious it was in the US in the 1940s and 1950s. His granny died of complications from emphysema and rheumatic heart disease when he was a teenager.
“During the first half of the 20th century, it ranked as the leading cause of cardiac death in the United States,” said Sable. During World War II, almost 100,000 potential recruits were disqualified from serving because they had rheumatic fever or rheumatic heart disease.
Is AI Capable Of “Working Like A Doctor”?
The application of AI in healthcare has been on a meteoric rise since the FDA approved the first independent AI-based diagnostic tool, the IDx-DR, to screen for diabetic retinopathy. This serious eye condition can develop in patients with diabetes. Nearly 700 medical gadgets utilizing AI and ML have been authorized by the FDA after little over five years.
Equipment for the Children’s National team would be lightweight and economical, weighing in at $2,000 to $5,000 instead of the $250,000 and 500 pounds that hospital echocardiography machines weigh. Us2.ai, a company based in Singapore, is now working on software that will use the team’s algorithm.
This algorithm underwent training using over a thousand echocardiograms. The algorithm is capable of analyzing images within seconds and categorizing them as “normal” or “considered rheumatic heart disease.” The novel tool demonstrated an estimated 90 percent accuracy rate when applied to echocardiograms of over 500 patients during testing.
“If AI can be used to improve the accuracy of images, its value is extraordinary,” said Chris Longenecker, director of the University of Washington’s Global Cardiovascular Health Programme and an individual who did not participate in the study. Ultrasound will become an exceptionally potent instrument, according to him, because artificial intelligence will enable its portability to the most remote regions of low-income nations.
AI may also assist physicians in gaining a deeper understanding of the critical traits associated with rheumatic heart disease, according to Longenecker.
According to Pooneh Roshanitabrizi, a staff scientist at Children’s National, and Marius Linguraru, principal investigator at the hospital’s Sheikh Zayed Institute for Paediatric Surgical Innovation, “AI has the potential to boost the capabilities of physicians.” The two colleagues also created an algorithm for echocardiogram interpretation.
According to Children’s National researchers, the novel screening system is undergoing evaluation and refinement in Uganda, drawing insights from as many as 200,000 pediatric cases. Each participant in the program undergoes a screening by the prevailing standard, wherein trained nurses independently interpret the echocardiograms.
The subsequent step involves utilizing the overwhelming majority of the data obtained from the 200,000 cases to train and enhance the algorithm. If the novel screening instrument is endorsed by regulatory bodies in Uganda and the United States, every nurse would be furnished with an AI-enabled tablet and a probe resembling a computer mouse in size. The probe would produce images of the heart.
Patients who were identified as potentially having rheumatic heart disease would undergo a cardiologist-led follow-up examination. Telemedicine could be utilized to conduct cardiologist visits, eliminating the need for patients to travel to and from remote locations.
Director of pediatric congenital cardiology at Johns Hopkins Children’s Centre Shelby Kutty, who was not involved in the project remarked, “Work in pediatric heart disease has been relatively limited.” Kutty continued that, in general, this is highly encouraging work, and he believes the field is proceeding as planned.
