Fetal EchoScan (v1.1)

BrightHeart's Fetal EchoScan (v1.1) is a cloud-based, machine learning-driven software designed to enhance the accuracy of healthcare professionals in interpreting second-trimester fetal cardiac ultrasound examinations by identifying suspicious findings. Extensive standalone and clinical studies confirm that this adjunct tool significantly improves diagnostic performance, leveraging advanced algorithms to assist in the early detection of potential cardiac anomalies.

BrightHeart's Fetal EchoScan (v1.1) is a cloud-based, machine learning-driven software designed as a Class II adjunct tool to assist healthcare professionals in interpreting second-trimester fetal cardiac ultrasound examinations. Intended for pregnant women aged 18 or older, between 18 and 24 weeks of gestation, it identifies suspicious radiographic findings such as overriding arteries, septal defects, abnormal outflow tract relationships, enlarged cardiothoracic ratios, cardiac axis deviation, and size discrepancies in ventricles and valves. This software is not a replacement for physician judgment or other diagnostic testing and is explicitly not for use in multiple pregnancies, cases of heterotaxy, or postnatal exams, having been determined substantially equivalent to its predecessor, v1.0. Extensive performance data supports the safety and effectiveness of Fetal EchoScan v1.1, including comprehensive software verification, validation, and cybersecurity testing. Standalone testing on 877 clinically acquired fetal ultrasound exams demonstrated high sensitivity and specificity in detecting the specified suspicious findings. Furthermore, a multi-reader, multi-case clinical study showed that physician interpretation accuracy significantly improved with software aid, evidenced by notable increases in ROC AUC, sensitivity, and specificity. The device's performance was consistent across various subgroups and is validated for use with ultrasound devices from major manufacturers including Fujifilm, GE, Philips, Samsung, and Toshiba.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the machine learning model specifically identify and prioritize the listed suspicious findings, and what are its false positive/negative rates for each individual anomaly?; What are the exact values for the 'high sensitivity and specificity' achieved in the standalone testing, and how do these metrics compare to the diagnostic accuracy of experienced sonographers without the aid of the software?; Given that the software is an adjunct tool, what specific training or workflow adjustments are recommended for clinicians to optimally integrate Fetal EchoScan v1.1 into their existing diagnostic protocols?; Were there any specific types of cardiac anomalies or image qualities where the software's performance was notably weaker or stronger, and what implications does this have for its clinical application?; What are the long-term plans for software updates, including potential expansion to other gestational ages, detection of additional cardiac conditions, or integration with a wider range of ultrasound hardware?