BrightHeart View Classifier

BrightHeart View Classifier is a cloud-based software device using machine learning to automatically detect and classify standard fetal heart views in 2D ultrasound images and video clips. It is intended for use by qualified healthcare professionals (sonographers, MFMs, OB/GYNs, fetal surgeons) as an adjunct to fetal anatomic ultrasound examination acquisition and interpretation during the second or third trimester of pregnancy.

Both the subject device and predicate (Sonix Health) are Software as a Medical Device (SaMD) applications using trained neural networks to identify standard cardiac views from ultrasound images and video clips. The key difference is that BrightHeart View Classifier is trained on fetal ultrasounds, while Sonix Health uses adult echocardiograms. BrightHeart includes a Predetermined Change Control Plan; Sonix Health does not. Sonix Health includes measurement functionality not present in BrightHeart View Classifier.

IEC 62304:2016 (Medical device software — Software life cycle processes); FDA Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices; FDA Guidance on Content of Premarket Submissions for Management of Cybersecurity in Medical Devices.

Both devices share the same fundamental purpose and technological approach: using trained neural networks to classify standard cardiac views in ultrasound imaging as an adjunct to clinician decision-making. Although BrightHeart targets fetal ultrasound while Sonix Health targets adult echocardiography, this difference in application domain does not raise different safety or effectiveness questions because both devices perform the same core function—automated view classification. The Sonix measurement functionality (absent in BrightHeart) removes a potential difference. Performance testing demonstrated BrightHeart achieves sensitivity of 0.939 and specificity of 0.984 with consistent performance across demographic subgroups and ultrasound vendors tested.