A.I. BEATS DOCTORS IN SPOTTING CARDIAC RISKS

An artificial intelligence program (AI) outperformed doctors in spotting subtle signs in the hearts of COVID-19 patients that placed the patients at higher risk of suffering long-term complications or death.
Physicians measure a heart’s strength by running a cardiac ultrasound test that gauges the volume of blood a heart is pumping from its left ventricle.
However, the volume may appear normal even if the heart is straining to pump out enough blood. The volume also can be artificially increased by high blood pressure.
Ultimately, the test’s validity depends on the skill and intuition of the doctor interpreting the results.
But now an AI program by Ultromics [sic], spun out of Oxford University, bested doctors in a trial.
The program locked onto the heart’s ultrasound image, mapped key points, and monitored how the heart’s ventricles moved as they contracted and relaxed. The software used changes in the ventricles’ shape to calculate exactly how strong or weak the heart was.
That precise data, combined with information about the patient’s age and any history of lung disease, allowed the program to predict the likelihood that a COVID patient would die in the hospital much more reliably than simply measuring the amount of blood a ventricle was squirting out.
The software predicted which COVID cases would become severe more accurately than a panel of cardiologists, among whom interpretations varied.
The program is now at work in hospitals in the U.S. and U.K.
TRENDPOST: Artificial intelligence is rapidly becoming your “second doctor,” with physicians relying on it to make or confirm diagnoses as well as to tailor treatments precisely to an individual patient’s biological makeup.