A tool for automated echocardiography
If you or your doctor suspects that there might be something wrong with your heart the go to thing is to do a heart ultrasound (TTE). This safe and non-invasive procedure is widely used in cardiology setting and globally there are more than 250 000 of them performed daily. However, TTE is time consuming. It takes from 30 to 90 minutes for a cardiologist or a technician to complete. Those measurements is what tells a cardiologist if everything is fine. It is because of these manual measurements that we have waiting queues of more than 5 weeks to get admitted for this examination.
Artificial intelligence at its best
The pain is simple – too much manual work of rather basic measurements. The solution is to use deep learning algorithms that are trained to recognize different heart pictures and perform different measurements on them so that a medical professional can skip doing them manually and instead focus on the patient.
Scale of the problem
Cardiovascular diseases (CVD) is the number 1 cause of death globally;
Direct healthcare and loss of productivity costs are set to reach US $ 1,044 billion by 2030;
CVD will continue to increase: World Health Organization (WHO) recognizes a developing obesity pandemic, which is a major CVD risk factor. Therefore, healthcare providers, especially in cardiology, are bound to become overwhelmed.
We developed a software for the automated heart ultrasound imaging workflow that allows to reduce the average examination time by ~85%, i.e. from 30 to 5 min, and increase the overall accuracy. This increases patient admission 6 times and increases revenue more than 2M (by one cardiologist).