A massive clinical problem, long underestimated
Congenital heart defects affect approximately 1% of births. When detected in utero, neonatal mortality and morbidity can be significantly reduced: organization of delivery in a specialized center, immediate care of the newborn, anticipation of complications. On the other hand, when they escape prenatal diagnosis, the consequences are often serious.
Morphological ultrasound of the second trimester requires the visualization, on a moving fetus, of cardiac structures of a few millimeters. This examination is carried out by gynecologists, midwives or radiologists who are not specialized cardiologists, but from whom extreme precision is nevertheless required.
“Many malformations are visible, but difficult to identify,” summarizes Cécile Dupont. Despite training and protocols, the human factor remains determining — and fallible.
An AI that fits into the care pathway, without disrupting it
Brightheart made the strategic choice to secure the existing system rather than reinvent it. The solution integrates with the second trimester ultrasound, without adding an exam or changing the medical schedule.
The tool takes the form of software connected to the ultrasound machine, with a tablet interface. The AI analyzes ultrasound videos in real time, verifies the completeness of the sections and provides a morphological indication of the fetal heart.
The AI does not interpret in place of the doctor, but acts as a layer of reliability.
Standardize a heterogeneous medical procedure
The first deployments demonstrated an immediate operational benefit. In 30 to 40% of cases, certain examinations prove to be incomplete a posteriori, requiring patients to be recalled. The tool makes it possible to detect these shortages in real time, improving medical quality and the efficiency of centers.
“The centers tell us that they can no longer do without it,” notes the manager.
Medical AI, above all a question of trust
The interview highlights a central point: the adoption of AI in health relies less on algorithmic performance than on trust. Fear of replacement, fear of opaque technology, complexity of integration remain major obstacles.
Brightheart responds with scientific proof: publications, database of more than 90,000 annotated examinations, design in constant interaction with clinicians. The tool is designed as a clinical assistant, not as a substitute.
Regulation: the foundation of clinical AI
FDA in the United States, MDR regulation in Europe: the regulatory constraint was integrated from the first phases of development. This anticipation allowed rapid marketing in the United States and prepares for an entry into Europe planned for 2026.
A trajectory built in stages
Created three years ago from an idea brought by two pediatric and antenatal cardiologists, Brightheart has progressed in a sequenced manner: data, algorithms, clinical validation, field integration, then commercial structuring.
The team, currently made up of a dozen people spread between France and the United States, is largely R&D and regulatory oriented.
Fundraising as a signal of maturity
This structuring convinced investors. Brightheart announces a lifting of eleven million eurosled by Odyssey Ventures And GoCapitalwith the participation of investor doctors and funds specializing in cardiology.
The funding aims to accelerate commercialization in the United States, prepare for entry into Europe and expand the functionality of the platform.
Towards an ultrasound infrastructure brick
Brightheart positions itself as a technological building block that will integrate into the ecosystems of ultrasound machine manufacturers and medical software publishers.
IPO, acquisition or industrial partnerships: the scenarios remain open, with a clear ambition: to standardize AI as an invisible but critical layer of obstetric ultrasound.
