Mount Sinai affiliated obstetric clinics in Manhattan are bringing an FDA cleared artificial intelligence tool into routine prenatal care, using it to screen fetal ultrasounds for structural heart problems, hospital leaders say. The software runs quietly in the background during second trimester anatomy scans, flags moving cine clips that look suspicious for major congenital heart defects, and prompts earlier referral to fetal cardiology. Clinicians involved report that in testing, the tool has already boosted detection performance and cut down the time doctors spend reviewing scans.
Carnegie Imaging for Women, a Mount Sinai affiliated OB/GYN imaging group with three Manhattan locations, is the first center in New York City to adopt the BrightHeart platform, according to Mount Sinai. “AI assistance in prenatal diagnosis offers not only improved detection, but has the potential to offer significant improvement in workflow and efficiency benefits,” Dr. Jennifer Lam-Rachlin, the study’s corresponding author, said in the release.
Study shows detection gains
The rollout follows a multicenter validation published in Obstetrics & Gynecology that tested the AI across hundreds of second trimester exams and compared clinicians’ reads with and without the software. In that peer reviewed analysis, the algorithm rendered a decision in nearly 99% of exams and delivered high sensitivity for findings associated with severe congenital heart defects. An earlier 200 exam analysis presented at the Society for Maternal Fetal Medicine meeting likewise showed about a 15% increase in sensitivity and roughly an 18% reduction in average reading time when clinicians used the tool, according to SMFM.
How the tool works and approval
BrightHeart, a Paris based med tech company, says its platform analyzes grayscale cine clips from standard four chamber and outflow tract views to flag eight morphologic markers that can indicate major congenital heart defects. The company has secured multiple FDA 510(k) clearances and a Predetermined Change Control Plan, regulatory steps BrightHeart says will let it iterate the AI under pre approved updates, as detailed in BusinessWire.
Why earlier diagnosis matters
Earlier prenatal diagnosis can change what happens after delivery. A 2011 multicenter analysis found that infants whose congenital heart defects were diagnosed after birth were about 16.5 times more likely to require emergency transport and generated far higher transport costs, roughly $5,144 per postnatal case versus about $389 per prenatally diagnosed infant, the authors reported. The paper argued that improving prenatal detection, whether through additional training or better tools, could produce substantial savings by avoiding urgent transports and their associated costs, and might recoup investments in improved screening.
Funding and limits
The Obstetrics & Gynecology article discloses that BrightHeart funded the research and that several authors reported consulting or compensation ties to the company, details the paper lists in its disclosures. Both the authors and outside experts cited in the paper call for further validation across different ultrasound machines, sonographer experience levels, and patient populations before the technology is treated as a diagnostic replacement rather than a support tool.
What’s next in New York
Mount Sinai says it will fold the software into everyday workflows at partner imaging sites while BrightHeart prepares a limited U.S. rollout, with the stated goal of bringing “close to expert level” review to centers that do not have fetal cardiology specialists, according to the hospital’s release. Both hospital and company statements stress that the tool is meant to support clinician interpretation, and that formal referral pathways to specialists remain central to care.
As reported by Crain's New York Business, Dr. Lam-Rachlin suggested hospitals could reduce malpractice exposure by catching more defects before birth, a claim that would depend on documentation and local standards of care. For now, researchers and clinicians say the device should be viewed as decision support that may help level screening quality, not as a legal shield.
