AI Tool Detects Heart Failure with 85% Accuracy

Researchers from Weill Cornell Medicine and NewYork-Presbyterian have developed a multimodal AI model to identify advanced heart failure ^[1]. The tool analyzes cardiac ultrasound videos alongside electronic health records to assess patient health ^[1][2]. It predicts peak oxygen consumption, a key metric for heart function, with 85% accuracy ^[1]. Traditional methods for measuring oxygen consumption often require specialized exercise tests that are difficult for many patients ^[1]. This new AI approach provides a less invasive alternative by using data already available in medical systems ^[2]. The study highlights how machine learning can streamline diagnosis for critical cardiovascular conditions ^[1].