Atherosclerosis is one of the most severe vascular pathologies, in which plaques appear on the inner walls of the arteries and become a serious danger to human life. Atherosclerosis is most easily diagnosed using ultrasound of the brachiocephalic arteries.
The danger of the formation of atherosclerotic plaques in the brachiocephalic arteries lies in the fact that the risk of developing strokes increases as a result of impaired blood circulation in the brain. The detection of plaques is a medically important procedure; however, it is costly. A machine learning model helps to determine the most accurate action to refer patients to screening for atherosclerosis when the indications for it are not obvious, for example, when obesity is a risk factor, using a machine learning model.
Webiomed has a built-in ‘Model for assessing the probability of the presence of atherosclerotic plaques in brachycephalic arteries in obese patients without CVD’ based on formalized data extracted from electronic health records (EHR).
The model is based on 16 input features and has the following quality metrics: ACCURACY: 0.96 ROC AUC: 0.97.
The digital algorithm predicts the probability of BCA atherosclerosis in patients with obesity and the need to perform additional examination methods to verify BCA atherosclerosis with an accuracy of 84%. The model can be used at the initial medical appointment to support the medical decision to screen for atherosclerosis in BCA.
“t is well known that atherosclerosis is a trigger for the development of most cardiovascular diseases. Using a predictive machine learning model for the development of atherosclerosis of the BCA using common cardiovascular risk factors will predict the likelihood of atherosclerosis of the BCA and the need to perform an ultrasound of the BCA during a medical examination of the patient more accurately, - stated Denis Gavrilov, Chief Medical Officer of Webiomed.