UltraSight, a specialist in AI-guided cardiac imaging, has shared new clinical data from a prospective, peer-reviewed clinical study evaluating the UltraSight Echosystem paired with Philips Lumify handheld ultrasound
Published in the European Heart Journal – Digital Health, the study discovered that students could acquire diagnostic-quality cardiac ultrasound images and that AI-guided interpretation could help identify left ventricular dysfunction.
In the primary study of 496 patients, students conducted cardiac ultrasound scans after four hours of training using Philips Lumify paired with the UltraSight Echosystem. Scans took a median time of four minutes, with adequate images obtained in 95% of cases where AI interpretation was used and 97.4% of cases where there was expert analysis.
The study examined a two-step workflow combining AI-guided image acquisition, AI interpretation, and physician overread of abnormal or uninterpretable scans. With this approach, 15.1% of scans required expert review while maintaining high diagnostic performance for identifying left ventricular systolic dysfunction (LVEF <40%).
The workflow achieved a negative predictive value of 99.8% and specificity of 95.4%, demonstrating the potential for AI-guided imaging to support expanded access to cardiac ultrasound while preserving physician oversight.
Andrew Goldsmith, MD, MBA, medical director of UltraSight, said, “Healthcare systems globally are facing increasing demand for cardiac imaging alongside persistent workforce and workflow challenges. This study demonstrates how AI-guided imaging with the addition of interpretation can help expand access to cardiac ultrasound by enabling more healthcare professionals to acquire diagnostic-quality images while preserving physician oversight and interpretation. The future is the ability to tell the whole story and not just acquisition or interpretation.”
Jared G. Bird, MD, cardiologist, echocardiographer, and co-author of the study at Mayo Clinic, said, “The study demonstrated the ability to acquire diagnostic-quality cardiac ultrasound images and assess left ventricular function using AI guidance within the study workflow. These findings contribute to the growing body of evidence evaluating how AI-guided imaging may support broader access to cardiac ultrasound.”
The full study, including a complete list of authors, disclosures and funding, is available in European Heart Journal – Digital Health: https://academic.oup.com/ehjdh/article/7/5/ztag066/8677231
