In Ulaanbaatar, Mongolia, the international eye-care nonprofit Orbis International and its technology partner Siloam Vision have begun what they describe as the first use anywhere of assistive AI to help clinicians screen premature infants for retinopathy of prematurity, the leading cause of childhood blindness worldwide. The program, run in partnership with Mongolia’s National Center for Maternal and Child Health, had conducted more than 270 exams on more than 170 newborns as of its most recent reporting, with results reviewed within seconds rather than the days it previously took to get a specialist’s read.
What Retinopathy of Prematurity Is and Why Speed Matters
Retinopathy of prematurity, or ROP, occurs when abnormal blood vessels grow in the retina of babies born very early, and it can progress from mild to blindness-causing within days if untreated. The World Health Organization and eye-health researchers estimate that roughly 32,000 premature babies worldwide become permanently blind from ROP every year, the large majority in countries where there are too few pediatric ophthalmologists trained to screen every at-risk newborn in the narrow window available before the disease becomes irreversible. Mongolia, like many countries with widely dispersed populations and few specialists concentrated in the capital, has historically struggled to get every premature baby screened in time regardless of where they were born.
How the Screening Program Works
Doctors and nurses in remote Mongolian provinces, who are not eye specialists, have been trained to use a specialized camera that attaches to a standard cell phone to capture retinal images of premature infants. Those images are analyzed by Siloam Vision’s proprietary AI system, which flags signs of ROP, and the results are also shared through the iTelegen telemedicine platform so a specialist can review flagged cases remotely. Because analysis happens in seconds rather than requiring images to be physically or digitally sent to a specialist and waiting for a response, babies who need urgent treatment can be identified and referred while the disease is still treatable, instead of after the point where intervention would help.
The People Behind the Numbers
Dr. Chimgee Chuluunkhuu, Orbis’s Mongolia country director, and Dr. Tsengelmaa Chuluunbat, head of pediatric ophthalmology at Mongolia’s National Center for Maternal and Child Health, have both been involved in building out the screening network across the country’s provinces. Karyn Jonas, Siloam Vision’s chief development officer, has said the intent behind the technology is to support, not replace, the judgment of ophthalmologists, framing the AI as a way to extend a scarce specialist’s reach into places a specialist cannot physically be. Among the infants already screened are a pair of twins, Ariunbaatar and Ariunnandin, whose mother Otgonchimeg was among the parents interviewed about the program; both were screened and cleared without needing referral for treatment.
The Technology’s Regulatory Standing
Siloam Vision’s ROP detection system has received FDA breakthrough device designation in the United States, a status the agency grants to technologies that show potential for more effective diagnosis or treatment of life-threatening or irreversibly debilitating conditions and that qualifies them for expedited review. That designation gives the Mongolia deployment more than local significance: it signals that the underlying AI system is on a path toward broader regulatory clearance and potential use in other countries facing similar specialist shortages, including an expansion Orbis and Siloam have said they plan for Dhaka, Bangladesh.
Why Global Health Specialists Are Watching Closely
Ophthalmologists working in global health have generally described point-of-care AI screening tools as one of the more promising near-term uses of AI in low-resource settings, precisely because the underlying bottleneck, a shortage of trained specialists, is a problem AI can meaningfully help address without requiring years of additional medical training to close. At the same time, some public-health researchers caution that screening technology alone does not solve the harder problem of ensuring babies flagged as high-risk can actually access laser treatment or other interventions once identified, particularly in rural areas where the nearest treatment center may still be a significant distance away even after a diagnosis is made.
What’s Next
Orbis and Siloam Vision have said their next priority is scaling the Mongolia program to cover more of the country’s remote provinces while preparing for the planned Bangladesh expansion, treating Ulaanbaatar as a proof of concept for a model that could eventually reach other countries where premature birth rates are high and pediatric ophthalmologists are scarce. The organizations frame the broader goal in stark terms: reducing the roughly 32,000 preventable cases of infant blindness the WHO estimates occur each year by getting AI-assisted screening tools into the hands of general clinicians in places specialists cannot reach in time.
