Electrocardiogram analysis for early detection of cardiovascular diseases in COVID-19 patients

Accelerate and improve quality of non-invasive 12-lead ECG recording and data analysis to support early detection of Cardiovascular disease in COVID-19 patients

Patients with COVID-19 frequently have an impact of the disease on their heart performance. Furthermore, if a COVID-19 patient enters the hospital with a pre-existing heart disorder, this is important information for the medical staff. However, this is not always known to both the patient or physicians. These cardiac disorders are reflected as abnormal ECG waveforms. Especially the part of the heart cycle which originates from the electrical activity from the large heart chambers, i.e. the activation and relaxation of the cardiac myocardium. In ECG terms these are labelled the QRS and T-wave morphology. For physicians however especially the T-wave part of the heart cycle is very difficult to interpret from the classic graphical ECG output. The measured electrical potential difference on the skin are sometimes very small. It is also difficult to relate the relaxation signals (T-wave) to the cardiac anatomy due to the slowness of the electrical relaxation of the heart chambers and thus the effective direction of the electrical current within the heart is only determined in the frontal plane of the patient. For the COVID-19 patient the use of alternative diagnostic technology such as MRI, CT and Echocardiography is not preferred given the impact on staff time. Both for recording time and cleaning the equipment after each recording.

The aim of the study was to reduce the burden on cardiac testing and detect cardiac complications during Intensive Care Unit admission by using CineECG derived from only the standard 12 lead ECG. In the project a novel technology for analysing the recorded ECG data was applied. The novel CineECG is creating digital reconstruction of the electrical activation and relaxation cycle in the heart. CineECG uses new algorithms (labelled MeanTSI) to relate the activation and recovery pathway to the cardiac anatomy. The aim of this COCVD study was to investigate if the digital health CineECG approach is able to separate normal from abnormal ECGs. In this project ECGs of COVID-19 patients were used, which have been collected and made available through the Capacity Covid Registry. This registry is an initiative of the Dutch Cardiovascular Alliance and is collecting data from Dutch and European Clinics and Universities.

In the project the ECGs of 100 normal controls were used to obtain the normal mTSI paths values for the QRS, ST segment and T-wave. The results of this analysis have been used to upgrade the basic CineECG algorithm. Effectively this means that 3 parameters related to CineECG (such as direction of the T wave within the heart anatomy) were added to the algorithm. The additional set of parameters derived from CineECG values were then used to classify the COVID-19 ECGs as either as normal or abnormal of 107 patients being treated for COVID-19 in the University Medical Centre Utrecht. The CineECG was able to classify 98% of the normal ECG correctly and 94% of the abnormal ECG in comparison to expert ECG classifications. The project results indicate the ability of the CineECG to relate the ECG to the cardiac anatomy effectively supports the detection of abnormal ECGs. The CineECG might be a novel ECG screening tool to detect potential cardiac involvement of the COVID-19 disease for non-ECG experts. 

COVID-19 is a new phenomenon which has both a respiratory and a cardiac impact regarding the health of the patient. Rapid detection of the cardiovascular abnormalities in a COVID-19 patient will support treatment decisions. In this project a new rapid method will be developed to detect these cardiovascular characteristics in COVID-19 patients by only using non-invasive 12-lead ECG data.
Technology Readiness Level (TRL)
3 - 6
Time period
5 months