❤️ AI Can Predict Heart Failure 5 Years Before The First Symptoms

The system was developed by scientists from Oxford. It analyzes CT scans of the heart and looks for changes in the fatty tissue around it. These micro-signs may contain early indications of inflammation that are invisible to the human eye.

The model was trained and tested on data from more than 70,000 patients. First, the system learned to identify these specific changes, and then to match these markers with whether or not a person developed heart failure.

This way, the algorithm learned to recognize early "patterns" of risk and calculate the probability of heart failure for patients who have not yet been diagnosed, with 86% accuracy.

➡️ The results were especially significant for the most vulnerable group. High-risk patients were found to be 20 times more likely to develop heart failure than those in the low-risk group. The approximate chance is 1 in 4 over five years.

@hiaimediaen