Although blood tests are widely used for its confirmation, it is found to be highly inaccurate. It involves measuring the natriuretic peptide (BNP) levels in the blood. The other alternative approach is the imaging tests that include CT scans, MRI, echocardiograms which are often expensive. The current AI assisted ECG method is found to detect ALVD with 85 percent accuracy and inline with mammography for breast cancer in terms of accuracy.
Mechanism of AI-assisted ECG for detection of ALVD
The research team ought to explore the potential of Artificial intelligence (AI) for the detection of ALVD. For this, they fed the neural network with data from Mayo Clinic corresponding to 6,25,326 patients. The fed the data in combination with the Electrocardiogram(EKG) and ECG. They tested the resulting algorithm on an independent dataset from 50,000 patients. The results are found to be exceptionally positive with an accuracy of 85 percent.
The method also detected patients that are prone to ALVD but did not display any visible symptoms. Later on, it was found that such patients are four times likely to develop ALVD in future when compared with negative results. The accuracy of the AI/EKG is found to be more or less similar to other tests such as a prostate-specific antigen for prostate cancer, cervical cytology for cervical cancer and mammography for breast cancer.
Since ECG’s are inexpensive, this new AI-ECG may offer affordable means in the future for early detection of heart failure. Paul Friedman MD, is the senior author of this study. The Research study is published in the journal Nature Medicine.