IBM and the pharmaceutical company AstraZeneca have developed a machine learning framework that notices the oncoming signs of acute coronary syndrome.
Acute coronary syndrome (ACS) is a broad term that describes situations that occur when blood flow in the heart is suddenly disturbed. ACS can lead to myocardial infarction. According to the WHO, 17.9 million people died from cardiovascular diseases in 2016, now it is the main cause of death worldwide.
For their study, IBM and the pharmacological giant AstraZeneca used data from nearly 27,000 adult patients from 38 urban and rural hospitals in China: age, gender, medical history, laboratory test results and procedures, and about 40 other parameters.
Scientists have designed a neural network that takes into account four factors simultaneously: whether the patient had major adverse cardiovascular events (MACE) before ACS; whether he has received anti-platelet drugs to prevent the formation of blood clots in the coronary arteries; if he was given beta blockers for lowering blood pressure; and whether they were prescribed statins, a class of cholesterol-lowering drugs.
The developers then applied the k-means statistical method to divide the patients into seven groups. This provided useful information: in one group, which included people with a high MACE, but a low level of the disease, diabetes was the main predictor of ACS. In the other – with patients in serious condition – age and systolic pressure, according to VentureBeat.
So scientists were able to get an analysis of the picture of the condition of patients with ACS, which also contains recommendations for treatment and a prognosis for recovery.
Nevertheless, the developers emphasize that it is still too early to use this information for prescribing treatment without further research, despite the full potential of this method. Clinical trials should be carried out first.
Predicting the decline in cognitive functions for five years ahead using AI has been learned in Canada. The algorithm examines MRI signatures, genetic characteristics, and the patient’s clinical data set and determines whether cognitive impairment can lead to the development of Alzheimer’s disease.