Experience with the pandemic has shown that the disease can pose a severe threat to the lives of patients. The main danger of the disease is acute respiratory syndrome and lung injury. However, patients may experience damage to other organs and systems: the cardiovascular system, the immune system, the liver and the kidneys. Myocardial injury occurred in at least 10% of unselected COVID-19 cases and up to 41% in critically ill patients or those with comorbidities [1].
In the survivors, the majority showed long-term symptoms, now often referred to as long COVID-19 [2, 3]. One of the critical long-term clinical consequences of COVID-19 seems to be myocardial injury [4–6].
Signs and symptoms of possible myocardial injury after COVID-19 may include severe fatigue, palpitations, chest pain, shortness of breath, postural orthostatic tachycardia syndrome (POTS) due to neurologic disturbances, post-exertional fatigue, and higher troponin levels [7–10].
In addition, COVID-19 appears to cause severe myocarditis. It can affect the myocardium and pericardium, causing severe fatigue without other apparent symptoms [7]. Diagnosis of myocarditis is relatively inaccurate because both tests and diagnostic protocols lack accuracy. Some reports showed that symptoms persisted for an average of 47 days before being diagnosed by cardiac magnetic resonance (CMR) imaging [11].
Therefore, it is critical to identify critical factors for assessing COVID-19 severity, predicting treatment outcomes, and optimizing individual treatment strategies [12, 13]. It is known that 49 variables can provide valuable prognostic information about mortality and disease severity in patients with COVID-19 [12].
Numerous studies have confirmed that cardiac [14] and other biomarkers may reflect cardiovascular injury and inflammation in COVID-19 and are strongly associated with poor prognosis and mortality [15, 16]. In addition, some electrocardiographic [17] and echocardiographic alterations [18] appear to have prognostic implications for patients with COVID-19.
Several prognostic models have been developed to assess disease severity in patients with COVID-19 and predict mortality [19–24].
Such classification models have usually been developed using various machine learning (ML) algorithms. For example, one neural network model has demonstrated 93% accuracy in predicting mortality based on patients' physiological status, symptoms, and demographic information [20].
A multivariable logistic regression model and an online risk calculator based on 10 clinical indicators were proposed to predict critical illness development among hospitalized patients with COVID-19 [21]. A support vector machine (SVM) model based on 11 routine clinical parameters was developed to assess the severity of COVID-19 patients [22].
An interpretable mortality prediction model for COVID-19 patients was proposed by Yan et al., where the XGBoost ML algorithm was used to select predictors. The interpretable decision tree and the decision rule for 3 biomarkers that predict the survival of individual patients with more than 90% accuracy were obtained [23].
It should be noted that in one of the ML models for predicting the severity of COVID disease, among the 33 analyzed signs and indicators, there was the cardiac functional grading (according to New York Heart Association functional classification) [24]. However, this cardiac indicator was excluded from the model because of its weak positive correlation with the severity of COVID-19.
In this context, the advanced analysis of ECG is highly demanded.This is especially true for patients with a normal or slightly changed electrocardiogram, i.e. if the analysis did not reveal any“major” category according to the, for example, Minnesota coding system. The only way to increase the diagnostic value of ECG examination is to develop proper information technology (IT) — a combination of up-to-date methods and equipment bound into a chain that provides collection, storage, pre-processing, interpretation, conclusion and dissemination of information [25].
At the same time, the advancement of diagnostic methods, especially instrumental ones (i.e., methods of functional diagnostics), primarily entails a constant increase of their "distributive capacity" — the ability to detect subtler and subtler changes in the function examined by one method or another. Such opportunities emerge due to progress in technical measurement tools of a specific function and even more due to the development of informational technologies. In other words, due to the creation of new metrics — numerical parameters using which one can assess the aspects of the functioning of various human organs and systems that were inaccessible before.
As a result, new ways of improving the diagnostic accuracy of a particular method within its traditional application scenarios are discovered. Additionally, familiar methods find unconventional uses in new areas.
Everything mentioned above fully applies to the new informational technologies for cardiac electrical activity assessment developed at V.M. Glushkov Institute of Cybernetics of the National Academy of Sciences of Ukraine.
The main goal set by the developers in this context was to make any electrocardiography informative. Routine ECG analysis is based on specific ECG syndromes or phenomena defined within one of the existing visual ECG analysis algorithms. However, in most cases, no ECG syndrome can be identified during the analysis of an individual electrocardiogram, at least not one that reflects cardiac pathology, i.e., belongs to the "major" category according to the Minnesota coding system, for example. During the routine analysis, one is forced to assign a single class to all these electrocardiograms — electrocardiograms with no primary ECG syndrome identified. However, the question arises: are all these electrocardiograms the same in terms of their relative "distance" to the "ideal" electrocardiogram of a healthy human? They are not. Depending on the myocardial condition, this "distance" can be further or closer. Moreover, there is a reasonable hypothesis that this "distance" reflects the likelihood of serious cardiovascular events. This is where routine analysis of an electrocardiogram is uninformative.
That is why the Universal Scoring system method and software for ECG scaling that can provide the quantitative evaluation of the slightest changes in ECG signal were developed [25, 26]. This approach is based on, first of all, measuring the maximum number of ECG parameters and heart rate variability and, secondly, on positioning each parameter value on a scale between the absolute norm and extreme pathology. The suggested approach follows a popular Z-scoring ideology, where quantitative, usually point-based assessment of test results is determined using a unique scale containing data about intra-group test results variation. To calculate the Z-score mean, the test value of the group and its standard deviation are needed [27].
This study aimed to investigate the value of a new electrocardiographic metric for detecting subtle myocardial injury in patients during COVID-19 treatment. And also to test the hypothesis about the prognostic value of myocardial injury on the treatment outcome.