Cardisiography: A Non-Invasive Heart Screening Test

Coronary heart disease can come as a complete surprise and occur suddenly. Cardisiography was designed to lower the risk and make faster intervention possible. As a non-invasive heart screening test, cardisiography offers the possibility of early detection for heart diseases.

In this MEDICA.de interview, Meik Baumeister explains what prompted the development of Cardisiography and describes how it promotes early detection.

How did you come up with the idea behind Cardisiography?

Meik Baumeister: I had a heart attack at age 30 despite having had no prior risk factors. I was incredibly lucky at the time. Afterwards, I wondered how I could suddenly have coronary heart disease without having had any previous symptoms. Quite often, you don't notice any signs until it's too late and the heart attack is imminent.

Next, I had to regularly visit a cardiologist, who performed an ECG and a stress test once a year. Yet my research showed that this procedure can't definitively determine whether you have atherosclerosis in your heart arteries. An ECG or stress test only reliably detects coronary heart disease in 25 – 40 percent of patients. I learned that there is no reliable way to detect this disease until it’s too late.

That's when I met Prof. Dr. Gero Tenderich, a cardiac surgeon who has worked extensively on vectorcardiography. He had long believed that resting vectorcardiography should be able to detect ischemia in a non-invasive manner. This method records the magnitude and direction of the electrical forces that are generated by the heart in a multidimensional space.

Vectorcardiography has been around for 70 years, but it has always remained a theory due to the complexity that cannot be solved by humans. We then used artificial intelligence to crack the complexity of vectorcardiography, which marks the moment the Cardisio GmbH was born in the summer of 2015.

Cardisiography is a non-invasive heart screening test. Results are analyzed using AI. What are the AI ​​algorithms able to capture?

Baumeister: The Cardisio screening test essentially works like an ECG. We measure the electrical signals of the heart for 4 minutes. After the 4 minutes, the recording is automatically sent to our data center in Frankfurt. It assesses 290 parameters per heartbeat. These are different calculations based on the ECG signal, including lengths, intervals, amplitudes, widths, angles, etc. The results are then checked for significant changes using a combination of different AI methods to determine the Cardisio Index. The Cardisio Index specifies the risk for coronary disease in the colors red, yellow or green. It also indicates the probability of risk in percentage.

How is Cardisiography able to detect heart disease even if patients experience no symptoms?

Baumeister: The electrical field is the most sensitive area that can be measured in the body. Thanks to our highly sensitive measurement technique we apply with the ECG, we can detect the smallest changes in the spread of cardiac excitation. If there is a reduction in perfusion, meaning there is less blood flow to specific parts of the heart, the response of the myocardial cells is delayed. This creates an asymmetry in the spread of cardiac excitation. We analyze and evaluate this asymmetry.

What are sensitivity and specificity?

Baumeister: We first focused on achieving a relatively high level of sensitivity. That is to say, if the test indicates that the patient is healthy, there is a 97 percent probability that this is indeed the case. Specificity is at the opposite pole: higher sensitivity automatically means lower specificity. Specificity refers to the extent of accuracy or probability that a test will correctly identify that a person is sick.

Our screening test is currently at a specificity of 85 percent. This means that out of 100 people who are identified by Cardisiography as being sick, 15 are actually not sick and tested "false positive". You can determine overall accuracy by combining the false positive rate, sensitivity, and specificity. Cardisiography is 91 percent accurate in this setting, which is a very high percentage by today's standards. Cardisiography accuracy will continually improve in the future as we enhance and optimize our AI models with new data and new information.

The collected data is correlated together in the Cardisio HealthNet. What happens to the data?

Baumeister: Once the data gets to the data center, it is fed into an algorithm - our proprietary artificial intelligence. We have 25 connected neural networks that subsequently assess 290 parameters per heartbeat. This is followed by various statistics to conclude significance and correlations. The neural network uses the results to determine the values ​​for the Cardisio Index.

The neural networks were trained with data from patients who were examined using Cardisiography and cardiac catheterization. Afterward the results were compared to review whether the Cardisio test results concurred with the cardiac catheterization results. Data from matching results were subsequently used to train the neural networks.

What are the characteristics that differentiate Cardisiography from standard medical devices used to diagnose heart disease?

Baumeister: First we have the test’s accuracy. Right now, there is nothing comparable out there, aside from cardiac catheterization, which is an invasive and expensive procedure. The second benefit is the speed, which is very fast as testing only takes 4 minutes or less. What’s more, Cardisiography is a very easy-to-use and cost-effective method.

You are a startup from Frankfurt that has already expanded internationally - what are your future goals?

Baumeister: Our primary goal is to make Cardisiography part of the cardiology guidelines. We also aim for a comprehensive use in an effort to help as many people as possible to detect a risk of coronary heart disease early and prevent future complications and heart attacks. Early detection can set the stage for preventive measures and effective treatment. There are simple ways to limit or stop the progression of this disease. That’s why our goal is to make Cardisiography accessible to all people and enable the monitoring of heart disease.