Review of processing pathological vectorcardiographic records for the detection of heart disease

Review of Processing Pathological Vectorcardiographic Records for the Detection of Heart Disease Jaroslav Vondrak * and Marek Penhaker Faculty of Electrical Engineering and Computer Science, VSB-Technical University of Ostrava, Ostrava, Czech Republic Vectorcardiography (VCG) is another useful method that provides us with useful spatial information about the electrical activity of the heart. The use of vectorcardiography in clinical practice is not common nowadays, mainly due to the well-established 12-lead ECG system. However, VCG leads can be derived from standard 12-lead ECG systems using mathematical transformations. These derived or directly measured VCG records have proven to be a useful tool for diagnosing various heart diseases such as myocardial infarction, ventricular hypertrophy, myocardial scars, long QT syndrome, etc., where standard ECG does not achieve reliable accuracy within automated detection. With the development of computer technology in recent years, vectorcardiography is beginning to come to the forefront again. In this review we highlight the analysis of VCG records within the extraction of functional parameters for the detection of heart disease. We focus on methods of processing VCG functionalities and their use in given pathologies. Improving or combining current or developing new advanced signal processing methods can contribute to better and earlier detection of heart disease. We also focus on the most commonly used methods to derive a VCG from 12-lead ECG. Keywords: vectorcardiography, heart disease, VCG features, transformation methods, electrocardiography 1 INTRODUCTION Measuring the electrical activity of the heart using electrocardiography and vectorcardiography is well established, as these methods have been used for more than a hundred years Burch (1985),Burch and DePasquale (1990). The cells that create the myocardium are joined together by gap junctions, which have very low resistance to a normal healthy heart. As a result, activity in one cell easily spreads to neighboring cells. However, this muscle cannot be controlled by the will. Durrer et al. (1970) published a study which shows that activation wavefronts progress relatively evenly, from endocardium to epicardium and from apex to base. Electrical activation of the heart begins in the sinus node (SA), and it spreads along the atrial walls. Then depolarization reach the atrioventricular (AV) node. Propagation of AV junction is very slow and results in delays during activation, which is a desirable, because it allows completion of ventricular filling. Once the activation reaches the chambers, the excitement continues along the Purkinje fibers. Furthermore, depolarization waves occur from left to right of the septum. Then, the depolarization spread through the left and right ventricular wall. Because the left ventricular wall is thicker, depolarization of the left ventricle continues even after depolarization of a large portion of the right ventricle. The left ventricle is depolarized mainly in parallel through the left anterior and posterior fascicles and the left the lateral basal part is the last to be activated. Ventricular repolarization begins on the outside of the ventricles and “spreads”inward. Although the epicardium Edited by: Sebastian Clauss, Ludwig Maximilian University of Munich, Germany Reviewed by: Lennart Bergfeldt, University of Gothenburg, Sweden Mathias Baumert, University of Adelaide, Australia *Correspondence: Jaroslav Vondrak [email protected] Specialty section: This article was submitted to Cardiac Electrophysiology, a section of the journal Frontiers in Physiology Received: 17 January 2022 Accepted: 04 March 2022 Published: 21 March 2022 Citation: Vondrak J and Penhaker M (2022) Review of Processing Pathological Vectorcardiographic Records for the Detection of Heart Disease. Front. Physiol. 13:856590. doi: 10.3389/fphys.2022.856590 Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565901 REVIEW published: 21 March 2022 doi: 10.3389/fphys.2022.856590 is depolarized last, its action potential duration is short and it is the first to recover. Although single cell recovery does not propagate to neighboring cells, it can be noted that recovery generally moves from the epicardium to the endocardium. Inward repolarization generates a signal with the same sign as outward depolarization. Due to the diffuse form of repolarization, the amplitude of the signal is much smaller than the amplitude of the depolarization wave and lasts longer Malmivuo and Plonsey (1995). These changes in depolarization and repolarization are then measured from the patient’s body surface, most often in the form of a 12-lead ECG. The method of vectorcardiography dates back to 1887, when in the first article concerning the human electrocardiogram, Augustus D. Waller pointed out the dipolar nature of the cardiac electric generator. Thus, it is possible to describe an electric generator of the heart with reasonable accuracy by an equivalent dipole. This dipole can be described as an electric heart vector (EHV) and it is possible to display it in vector form Waller (1887). In 1920s Mann first introduced the concept of a “loop” representing a continuous series of vectors for depicting electrical depolarization and repolarization magnitudes. Mann derived these loops manually from three Einthoven leads Burch (1985). From 1936 to 1940 the technique of vector representation of the electric field of the heart was actively developed in Germany Howard (1946). For direct measurement of orthogonal leads, Schellong et al. (1937) introduced the first uncorrected orthogonal system in 1937. Their work was further developed by other authors who defined new lead systems Kimura (1939),Duchosal and Sulzer (1949),Grishman et al. (1951),Milnor et al. (1953). These lead systems differ by placing the electrodes on the hull and are represented by signals that are orthogonal to each other. However, they do not take into account the different torso geometry or intrinsic tissue inhomogeneity. The first corrected lead system was derived by Frank based on a mathematical model Frank (1956). This system, which uses seven measuring electrodes, is today one of the most widely used vectorcardiographic leads. Other published but less commonly used vectorcardiographic leads include McFee and Parungao (1961), SVEC III Schmitt and Simonson (1955), and hybrid lead systems Dellborg et al. (1995). Like the ECG, VCG is a diagnostic method which is considered as a very useful method for measuring the electrical activity of the heart. It is more sensitive than a standard 12-lead ECG and provides the cardiologist with important additional information such as a clearer indication of the phase relationships between leads Rubel et al. (1991), Levkov (1987). Today, the QRS-T angle is the most commonly analyzed of the VCG, while current ECG markers of repolarization abnormalities mainly include ST depression, T FIGURE 1 | 1) The basic principle of vectorcardiography is illustrated on ideal uniform lead fields, which are perpendicular to each other and are in a bipolar configuration (set by parallel electrodes on opposite sides of the torso) Malmivuo and Plonsey (1995); 2) Placement of measuring electrodes on the patient body using Frank lead system Hasan and Abbott (2016). Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565902 Vondrak and Penhaker Review of Processing Vectorcardiographic Records wave inversion, and QT prolongation Dilaveris et al. (2001), Voulgari and Tentolouris (2009). The VCG is projected into three mutually perpendicular planes: sagittal, transversal and frontal, see Figure 1 (left), where placement of the electrodes are shown in Figure 1 (right). The individual planes are most often located as: frontal plane between X and Y leads, transversal plane between X and Z leads and sagittal plane between Y and Z leads. Cardiac activity is then described by three loops, which represent the individual phases of the cardiac cycle. The first loop corresponds to wave P, the second loop which is the largest corresponds to QRS complex and the third loop corresponds to wave T. The loops can be seen in three 2-D projections or in one 3-D image in a demonstrative physiological record from the PTB database, see Figure 2. The records contained in this database are sampled by sampling frequency of 1 kHz with a 16-bit resolution in the range of ±16,384 mV. Based on the recommendation Merri et al. (1990), the sampling frequency for measuring electrical activity of the heart should be at least 128 Hz. However, setting parameters in the PTB database is suitable for subsequent and detailed analysis of the electrical activity of the heart. However, along with the required signal, interfering components that need to be removed from the records can also be measured. When designing filters, it is necessary to take into account the frequency range of the desired signal and the frequency characteristics of individual filters. The benefits of a 12-lead ECG for diagnostic evaluation of electrical activity of the heart has been a standard in clinical practice over a hundred of years. However, there are certain cases where the vectorcardiogram is superior to the electrocardiogram. In some publications, a higher sensitivity of VCG compared to conventional ECG has been reported in the diagnosis of atrial enlargement and right ventricular hypertrophy. It has been proposed to re-evaluate the frequency of 12-lead ECG usage to increase vectorcardiography measurements in clinical practice van Bemmel et al. (1992),Chou (1986). In recent years, VCG has become a method that is processed by modern signal processing procedures, mainly due to the possibility of obtaining and subsequent analysis of spatial features. Studies have shown that the vectorcardiogram is very useful in some specific situations, such as assessing intraventricular conduction disorders combined with inactive areas, identification of sudden cardiac death, identifying and locating ventricular preexcitation, differential diagnosis of patterns different from normal deviation from electrical axis, assessment particular aspects of Bruges’syndrome and estimating the severity of some cardiac enlargements Sur et al. (2013),Perez Riera et al. (2007). Also, more accurate results were obtained in the analysis of QRS in three-dimensional projection, such as improved patient selection for cardiac resynchronization therapy (CRT), detection of myocardial injury Correa et al. (2010) or extraction of VCG FIGURE 2 | Demonstration display of individual VCG planes for randomly selected physiological record: (A) Transverse plane of X and Y leads, (B) Sagittal plane of X and Z leads, (C) Frontal plane of Y and Z leads, (D) 3-D image of X, Y and Z leads. The record s0503 rem from the PhysioNet PTB database was used as a randomly selected physiological record. The individual planes are related to the basic principle of VCG measurement on ideal uniform lead fields from Figure 2. Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565903 Vondrak and Penhaker Review of Processing Vectorcardiographic Records features from QRS complex Correa et al. (2012) for ischemia detection. Another advantage of vectorcardiography over standard ECG is in detection of long QT syndrome Diamant et al. (2013,2010),Cortez et al. (2017a). Another use of VCG, due to its higher sensitivity, was in the analysis of inducted cardiac memory, which was mainly known from ECG recordings. Wecke et al. (2013) analyzed the VCG records of patients after previous ablation of accessory pathways (AP). They were aware of the phenomenon of induced cardiac memory, which was present on ECG records in different patient ratios. They assumed that VCG, which is more sensitive than ECG, would show cardiac memory independent of AP location after ablation. They found out from the analyzed records that after ablation there was a correlation between the directions of the overexcited maximum QRS vector and the post-ablation maximum T-vector, indicating the presence of cardiac memory. From their findings, it was confirmed that the information we know from the ECG can also be found in the VCG with the possibility of a new perspective. Vectorcardiography, which represents a slightly different approach, is used less commonly in practice and VCG leads are often derived Frank (1956),Iwaniec et al. (2018),Sedaghat et al. (2016),Sun et al. (2017),Treskes et al. (2015),Correa et al. (2012, 2013). This examination method can be considered as a useful tool in the study of many heart diseases and can provide additional information to the conventional ECG in the form of additional spatial information Vozda and Cerny (2015). However, VCG is not usually recorded in clinical practice but orthogonal leads can be derived from a conventional 12-lead ECG Belloch et al. (2007).The importance of vectorcardiography has been published in numerous publications, but VCG records are not available in most cases. Therefore, an alternative form of deriving VCG from a commonly measured 12-lead ECG was proposed. Derived VCG is useful for estimating some meaningful features that represent high diagnostic information such as QRS-T angle or Total cosine R to T (TCRT). These and other features can be estimated from derived VCG with sufficient accuracy Karsikas et al. (2009),Cortez and Schlegel (2010),Cortez et al. (2014). In recent years, vectorcardiographic recordings have been increasingly used for the analysis and detection of heart disease. If directly measured records using Frank lead system are not used, they are often transformed by various methods from a 12-lead ECG. In the following chapter, this paper provides an overview of the most commonly used transformation methods. In addition, it reviews the processing and extraction of important diagnostic parameters, the most promising techniques and current challenges for the detection of various heart diseases. 2 REVIEW STRATEGY This work presents a comprehensive overview of the multidisciplinary area of vectorcardiographic record processing and methods of possible transformation from 12-lead ECG to obtain derived VCG leads. The individual methods and steps for the implementation of this overview are presented here. These are mainly: selection of suitable databases, selection of search terms, and evaluation of results. This review was conducted using full papers, including publications in journals, conference papers, books, and academic papers. The search was performed without a time limit to provide a historical background. The search for relevant works was carried out in English. 2.1 Database Selection Four basic databases were selected for the classification of suitable literature. These include the Scopus database and the Web of Science, which are among the largest databases, including peerreviewed citation sources. PubMed and ProQuest databases, which focus on the selection of medical and medical literature, were also used. The combination of medical and technical literature should be the basis for comprehensive information, both medical and especially technical in the field of vectorcardiographic processing. 2.2 Indexed Terms In this section, we list the individual indexed terms that have been used for this overview. In the case of processing vectorcardiographic records, the terminology is inconsistent. Therefore, several combinations of indexed terms have been used to include the widest possible range of articles that are relevant to this review. A summary of the individual indexed terms used can be seen in Table 1. 3 TRANSFORMATION METHODS The first consideration on the transformation of individual lead systems was presented by Burger et al. (1952).Kornreich et al. (1974) described that the 12-lead ECG and Frank lead system were very similar in terms of their information content and therefore their mutual transformation was possible. This resulted in the first attempts to transform lead systems based on the transformation from VCG to 12-lead ECGs by Dower Dower (1968).Wolf et al. (1976) pointed out that knowledge from measuring of orthogonal lead systems can contribute to better diagnosis. He therefore considered the possibility of a two-way transformation and later he derived transformation matrices for the bidirectional transformation of conventional 12-lead ECG to VCG. Many of the transformation methods were derived only for the QRS complex of the heart cycle. There are also transformation matrices that are focused on other parts of the heart cycle. For example, Guillem et al. (2006) designed a modified transformation matrix optimized for the P wave, thereby achieving improved transformation for this ECG segment. Guillem et al. (2008) addressed the issue of accuracy of P wave derivation further in. The object of this study was to test the accuracy of the inverse Dower transformation in comparison with the P-wave optimized transform. The problem of P-wave accuracy in a transformed VCG was also addressed by Carlson et al. (2005). In their study, they used a total of 41 patient records, of which 20 records were diagnosed with atrial fibrillation. After the transformation using the inverse Dower transformation, the waveform of P was preserved. When comparing the directly measured VCG and the derived VCG, the morphological parameters of the P wave were consistent in the respective groups, and better conservation was observed in the healthy groups. Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565904 Vondrak and Penhaker Review of Processing Vectorcardiographic Records Transformation methods for the derivation of ECG and VCG leads are an integral part of obtaining further beneficial information from the measurement of electrical activity of the heart. Most of the published articles use databases with already measured ECG and VCG records. However, simultaneously measured orthogonal data with a 12-lead ECG is not always available. Therefore, this chapter is devoted to the most commonly used transformation methods used in publications and scientific works. 3.1 Quasi—Orthogonal Transformation Any ECG lead system can be converted into vectorcardiographic loops. However, these derived VCG loops will not be orthogonal. Certain leads from a 12-lead ECG show a high correlation with orthogonal leads. These leads are called as quasi-orthogonal. Such derived lead systems correspond approximately to uncorrected VCG leads. Correction can be achieved using geometry based on the torso model derived by Frank (1954). Bjerle and Arvedson (1986) derived some of the first quasiorthogonal leads that can be expressed as (Eq. 1): X1,06 ·V6 Y1,88 ·VF 1,25 ·aVF Z−0,532 ·V2+0,043 ·V6 (1) Kors et al. (1986) analyzed ECG lead systems and selected those leads that showed the highest correlation with orthogonal leads. Derived orthogonal leads can be expressed as Eq. 2: XV6 YII Z−0,5·V2 (2) Kors et al. (1990) compared two quasi-orthogonal lead systems (Eq. 1) and (Eq. 2) and VCG derived by regression and Inverse Dower Transformation (IDT). Based on the measurement of the mean absolute deviation and the evaluation of cardiologists, he concluded that the transformation matrix obtained by regression and IDT achieved better results than both quasi orthogonal leads. 3.2 Transformation Methods Based on Linear Approach The standard 12-lead ECG does not provide much information about the sagittal plane, so it is necessary to use all the information contained in the ECG to derive the VCG. It is the VCG that provides us information of cardiac activity in all three planes. The simplest conversion to a VCG is to use quasiorthogonal conversions, where one ECG lead corresponds to one VCG lead. A more reliable variant is the use of linear transformation methods, where each ECG lead (I, II, V1, V2, ... , V6) contributes to a certain extent to a specific orthogonal lead. The individual weight coefficients then form the resulting transformation matrix M. This approach was first introduced by Burger et al. (1952) in the analysis of the description of the electrical action of the heart by one dipole. Transformation matrix coefficients were derived using a regression approach and their accuracy tested in 169 patients. However, the authors did not publish the transformation coefficients. For the conversion, the IDT has become one of the most widely used transformations. The coefficients can therefore be applied to individual ECG leads in the form of (Eq. 3). X0,156 ·I−0,01 ·II −0,172 ·V1−0,074 ·V2 +0,122 ·V3+0,231 ·V4+0,239 ·V5+0,194 ·V6(3) The Y and Z leads can be derived similarly. The difference between a quasi-orthogonal systems and a linear transformation can be noticed here. Individual VCG leads in a quasi-orthogonal system correspond to exactly one ECG lead, while in the case of a linear transformation matrix the VCG leads is formed by the weight coefficients of the individual ECG leads. The mathematical transformation is then realized as a multiplication of two matrices (Eq. 4). VM·E(4) where M is the transformation matrix, E is matrix whose rows are formed by independent ECG leads and V is matrix whose rows correspond to the derived VCG. The coefficients of the transformation matrices are also derived on the basis of the torso model described by Frank (1954) or by regression methods based on data measured in a representative group of patients. These coefficients then differ from one approach to another, describing in particular the morphology of the average patient or torso model. 3.3 Kors Regression Transformation The transformation matrix introduced by Kors et al. (1990) was derived by regression technique for a group of patients from the CSE database. The transformation matrix coefficients, see Table 2, were derived by minimizing the mean error between the measured VCG and the transformed VCG. In this way, Kors derived several transformation matrices for different complex segments. He also stated that the differences between individual transformation matrices are small. Given that the QRS complex is most often analyzed, the resulting matrix is based only on the regression of the QRS complex. Several publications have studied the transformation method introduced by Kors. The authors in Cortez and Schlegel (2010) and Cortez et al. (2014) discussed which of the available transformation methods provides the QRS-T spatial angle value closest to the values obtained from Frank lead system. They used two available transformation matrices: the Kors regression transform and the TABLE 1 | Indexed terms and their combinations. Index terms 1. Vectorcardiography OR Vector Cardiography OR Vector Electrocardiography 2. VCG OR ECG 3. Transformation methods OR Derivation methods OR Linear transformation methods OR Quasi-orthogonal transformation OR Frank lead system 4. VCG features OR ECG features 5. P-loop OR QRS-loop OR T-loop 6. Medical signal processing OR Biomedical signal processing Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565905 Vondrak and Penhaker Review of Processing Vectorcardiographic Records IDT. The authors concluded that the resulting values from the Kors regression method did not differ significantly from the values from Frank lead system. In their further publications, the authors in Cortez et al. (2015) have focused on the analysis of the spatial angle of QRS-T in patients with hypertrophic cardiomyopathy. Similarly, Man et al. (2011) analyzed the QRS-T spatial angle from the derived VCG data using the Kors regression method and IDT. Of the two transformation methods used, the Kors regression method achieved better results. The analysis was performed between the transformed VCGs and the VCGs measured by the Frank lead system. 3.4 Inverse Dower Transformation (IDT) In 1980, Dower et al. (1980) presented the possibility of deriving 12-lead ECGs from three leads measured by the Frank lead system. They created the transformation coefficients, see Table 3, which provide better correlation for precordial electrodes with respect to the voltage and P and T waves in leads V1 and V2 Dower (1968);Dower et al. (1980);Dower and Machado (1979). Using this knowledge, Edenbrandt and Pahlm (1988) derived a pseudo-inverse matrix that can be used for transformation from 8 independent ECG leads to VCG. The resulting pseudoinverse matrix is shown in Table 4. The Inverse Dower Transformation was used by Panagiotou et al. (2013) and Dima et al. (2013) to transform a 12-lead ECG into a VCG for subsequent feature analysis to detect myocardial scar. Similarly, Sun et al. (2017) used IDT to analyze the VCG loop in patients with myocardial ischemia. Dawson et al. (2009) compared the Dower transformation matrix with the affinity transform in relation to the transformation from 12 to 8 lead ECG to 3 lead VCG and back. Based on the evaluated results, they conclude that in both myocardial infarction (MI) and healthy (HC) patients, the affinity transformation achieves better results in transformation from 3-lead VCG to 12-lead ECG than the Dower transform. 3.5 P Least Square Value (PLSV) and Q Least Square Value (QLSV) Transformations Transformation matrices derived from the regression approach mainly focus only on the QRS complex of the heart cycle. The accuracy of P and T waves is usually considered sufficient in transformations and the differences in transformation matrices are minimal. VCG loops can be used to detect arrhythmias with high accuracy, emphasis is also placed on the accuracy of P and T wave transformation Kors et al. (1990). Guillem et al. (2006) presented a transformation matrix derived using the regression method, which is optimized for P wave. They named the transformation matrix as PLSV, see Table 5. In addition to the matrix targeting the P wave, they also derived a matrix optimized only for the QRS complex, transformation coefficients are shown in Table 6. Both of these matrices were derived from a total of 124 patients. Using the least squares method, a regression model was found for each patient and the resulting matrix is given as the mean value of the transformation matrices for all patients. Guillem et al. (2006) also compared the Kors and PLSV matrices for atrial fibrillation records, and the PLSV transform yielded significantly better results in this regard. 3.6 Transformation From Mason-Likar (ML) ECG Leads During stress tests, measuring a 12-lead ECG is not appropriate due to limb movement. Therefore, Mason and Likar have published their recommendations on how to limit the movement of electrodes when measuring ECG in stress tests Mason and Likar (1966). Interfering components that are created by movement are eliminated by moving the measuring electrodes to the chest. The resulting differences in signals should always be taken into account Papouchado et al. (1987). For these reasons, standard linear transformation methods cannot be used. Guldenring et al. (2012) designed a new transformation matrix, see Table 7, for the transformation of ECG measured using Mason-Likar leads. 3.7 Singular Value Decomposition of 12-Lead ECG Another possibility of deriving vector cardiographic leads is by reducing the dimension of data using Singular Value Decomposition (SVD). This is an orthogonal matrix reduction of the data dimension defined by Golub and Van Loan (1996). The principle of this transformation is as follows (Eq. 5): ΣUTMV (5) where columns of U are referred to as the left singular vectors, columns of V are referred to as the right singular vectors and M is 8xnmatrix of individual ECG lead of nsamples Acar et al. (1999). TABLE 2 | Transformation coefficients of Kors regression method. Lead I II V1 V2 V3 V4 V5 V6 X 0.38 −0.07 −0.13 0.05 −0.01 0.14 0.06 0.54 Y−0.07 0.93 0.06 −0.02 −0.05 0.06 −0.17 0.13 Z 0.11 −0.23 −0.43 −0.06 −0.14 −0.20 −0.11 0.31 TABLE 3 | Leading vectors for deriving a 12-lead electrocardiogram from the Frank XYZ signal. Lead I II III aVR aVL aVF V1 V2 V3 V4 V5 V6 X 0.632 0.235 −0.397 −0.434 0.515 −0.081 −0.515 0.044 0.882 1.213 2.125 0.831 Y−0.235 1.066 1.301 −0.415 −0.768 1.184 0.157 0.164 0.098 0.127 0.127 0.076 Z 0.059 −0.132 −0.191 0.037 0.125 −0.162 −0.917 −1.387 −1.277 −0.604 −0.086 0.230 Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565906 Vondrak and Penhaker Review of Processing Vectorcardiographic Records The orthogonal leads obtained in this way do not correspond directly to the Frank VCG, and the indications obtained from these leads do not correspond to the indications derivedfromtheVCGleads,asstatedbyBelloch et al. (2007). However, this method found use Acar and Koymen (1999) where Hasan et al. (2012a,b) analyzed the morphology of QRST loops using SVD. 3.8 Summary of Transformation Methods Transformation methods were derived to obtain orthogonal lead leads from a 12-lead ECG. There are many linear transformation methods that are used in various branches of VCG processing. However, each method has its advantages and disadvantages, especially in the processing of pathological records, where different pathologies affect different parts of the heart cycle. In such a case, knowledge of the effect of pathology on the ECG would be required to select the correct transformation method. If a transformation method is selected that is not suitable for a particular part of the ECG recordings, diagnostic information may be lost by signal distortion. The following Table 8 summarizes the key features of each linear transformation method. The Accuracy column shows the accuracy of the transformation method according to the evaluation parameter of correlation in individual leads. This is one of the most frequently used parameters in the evaluation of transformation methods. The stated quantitative values are defined as average values of all leads from publications Jaros et al. (2019),Vozda and Cerny (2015),Kors et al. (1990). However, these values are only indicative, as the accuracy of the derivation depends on several factors. One of the most frequently used transformation method is the IDT followed by the Kors regression method. There are more and more studies that point to the fact that Kors regression method achieves higher accuracy than other transformations Cortez and Schlegel (2010),Cortez et al. (2014),Man et al. (2011),Kors et al. (1990). The QLSV and PLSV methods are mainly focused on a certain part of the ECG (QLSV—QRS complex, PLSV—P wave). The quasi orthogonal method is derived by approximation to VCG leads and is not suitable for processing due to its high error rate. If we assume a pathology that will affect more parts of the ECG, a combination of two or more transformation methods couldbeusedtoobtainamoreaccuratelyderivedVCG.A graphic comparison of the individual transformation methods can be seen in Figure 3.Fromthefirstpointofview,itcanbe seen that the Kors regression method copies the shape of the directly measured curve most accurately. However, various statistical tests are needed to verify the accuracy of the TABLE 4 | Transformation matrix for Inverse Dower transformation (IDT). Lead I II V1 V2 V3 V4 V5 V6 X 0.156 −0.010 −0.172 −0.074 0.122 0.231 0.239 0.194 Y−0.227 0.887 0.057 −0.019 −0.106 −0.022 0.041 0.048 Z 0.022 0.102 −0.229 −0.310 −0.246 −0.063 0.055 0.108 TABLE 5 | PLSV transformation matrix. Lead I II V1 V2 V3 V4 V5 V6 X 0.370 −0.154 −0.266 0.027 0.065 0.131 0.203 0.220 Y−0.131 0.717 0.088 −0.088 0.003 0.042 0.048 0.067 Z 0.184 −0.114 −0.319 −0.198 −0.167 −0.099 −0.009 0.060 TABLE 6 | QLSV transformation matrix. Lead I II V1 V2 V3 V4 V5 V6 X 0.199 −0.018 −0.147 −0.058 0.037 0.139 0.232 0.226 Y−0.164 0.503 0.023 −0.085 −0.003 0.033 0.060 0.104 Z 0.085 −0.130 −0.184 −0.163 −0.193 −0.119 −0.023 0.043 TABLE 7 | Mason–Likar transformation matrix. Lead I II V1 V2 V3 V4 V5 V6 X 0.5169 −0.0722 −0.0753 0.0162 0.0384 0.0545 0.1384 0.4606 Y−0.2406 0.6344 0.1707 −0.0833 0.1182 0.0237 −0.1649 0.2100 Z−0.0715 −0.1962 −0.4987 −0.0319 −0.2362 −0.0507 −0.2007 0.4122 Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565907 Vondrak and Penhaker Review of Processing Vectorcardiographic Records individual methods. Nevertheless, in the case of greater use of Frank’s lead system in clinical practice to obtain the original VCG, it would not be necessary to address the possibilities of transformation and their accuracy in the preservation of diagnostic information. The following chapter describes the possibilities of detecting individual heart diseases from VCG records. Because VCG is not commonly obtained in clinical practice, transformation methods are being approached. These transformation methods are also listed in the individual publications when used. 4 HEART DISEASE ANALYSIS FROM VECTORCARDIOGRAPHY Since the 19th century, it has been known that the heart muscle emits quasi-periodic electrical signals during its activity. TABLE 8 | Overview of transformation methods. Transformation method Derivation of transformation methods Primary use Accuracy Kors regression transformation Minimizing the mean error between the measured VCG and the transformed VCG All types of ECG >98% Inverse Dower transformation (IDT) Pseudo-inverse matrix to a system based on a torso model Pathology affecting the QRS section >97.2% PLSV transformation Derivation by least squares method P wave of ECG >96.8% QLSV transformation Derivation by least squares method QRS complex of ECG >97% Quasi-orthogonal transformation Approximation to VCG leads from ECG leads All types of ECG >90% Mason-Likar (ML) Designed using the regression method Exercise and movement ECG >95% FIGURE 3 | An exemplary comparison of transformation methods, where the blue curve is measured by the Frank lead system and the red curves are the transformed one. Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565908 Vondrak and Penhaker Review of Processing Vectorcardiographic Records Although it is very important, there are no standardized methods for detecting pathologies from electrical heart activity measured in three mutually perpendicular planes. This is mainly due to the poor use of VCG in clinical practice and the scientific work that does not provide the necessary evaluation from cardiologists. Greater cooperation between authors and relevant doctors and mutual openness to new procedures could address this shortcoming. Using various methods, it is possible to find a pattern of healthy cardiac rhythms and different patterns for each type of hearth disease. Published contributions and studies evaluate patient records using various methods. It is important to use the right criteria for an objective comparison of the new methods. This section presents the different methods and techniques used for the subsequent evaluation of the most frequently analyzed pathologies. 4.1 Acute Cardiac Ischemia Acute cardiac ischemia can be characterized as an imbalance between myocardial oxygen supply and demand. The reduction in blood circulation is progressive and by delivering insufficient blood to the myocardium are myocardial cells deprived of the blood combinations necessary for their survival Tomey and Gidwani (2016).The authors approached to the detection of these life-threatening conditions by various methods of processing VCG records, where the most common cases were mainly myocardial ischemia and myocardial infarction. 4.1.1 Myocardial Ischemia In recent years, several publications have promoted different methods for detecting and classifying electrical cardiac activity in patients with myocardial ischemia Persson et al. (2006), Martínez et al. (2006),Arini et al. (2008),Toledo and Wagner (2007). Several publications have dealt with the topic of preventing these life-threatening conditions by early detection, where authors approach different methods of VCG analysis. The detection of myocardial ischemia was dealt by Sun et al. (2017), where they focused on the development of a screening system using a vectorcardiogram. They used a total of 132 patient ECG records from two databases and introduced a total of 14 new characteristics to detect arrhythmia. They used derived VCGs using IDT to analyze VCG features. Accuracy of the proposed method was verified by calculating Accuracy (Acc) (Eq. 6), Sensitivity (Sens) (Eq. 7)andSpecificity (Spec) (Eq. 8). The same calculation procedure was also used by Dima et al. (2013),Sun et al. (2017),Correa et al. (2013),Yang et al. (2012),Correa et al. (2016),Yang et al. (2013),Tripathy and Dandapat (2017),Huang et al. (2011). Sens TP TP +FN (6) Spec TN TN +FP (7) Acc TP +TN FN +FP +TP +TN (8) where True Positive (TP)/False Negative (FN) are records that have a Ischemic Heart Disease (IHD) and are correctly/ incorrectly identified, while False Positive (FP)/True Negative (TN) are records that do not have a IHD and are incorrectly/ correctly classified. The accuracy of the proposed method reached 98.07% sensitivity 98.63% and specificity 99.04%. They state that the reliability of the proposed method is guaranteed by the analysis of pathophysiology and application of QRS, ST and T criteria. Another processing of QRS area connected with ST segment was presented by Kawahito et al. (2003). They analyzed the difference in the QRS vector, which reflects changes in the shape of the QRS complex, and the size of the ST vector, which represents the diversion of the ST segment from the isoelectric level. They states that monitoring the ST vector size and the QRS vector difference by vectorcardiography may be useful for identifying myocardial ischemia during carotid endarterectomy. Correa et al. (2012) analyzed vectorcardiographic curves for the detection of ischemia and used a total of 80 ischemic and 52 healthy records. They studied five parameters, where the best results were achieved with a QRS Volume with a sensitivity of 64.5%, a specificity of 74.6% and an Area Under Receiver Operating Characteristic Curve (AUC) = 0.77. Later, they expanded their work, where their objective was evaluating the vectorcardiographic difference between both groups Correa et al. (2013). Synthesized orthogonal leads were obtained by Kors transformation. They analyzed seven QRS loop parameters, and from their analysis the best results achieves a QRS volume, which achieved sensitivity 64.5% and specificity 74.6%. In conclusion, they emphasize the fact that VCGandECGparametershavesignificant differences between healthy and ischemic subjects. The QRS analysis was extended by Sun et al. (2017) by evaluation of ST and T segment. They analyzed a total of 17 parameters from the modified VCG using the IDT. Out of the tested number of 132 records they achieved an accuracy of 98.07%, a sensitivity of 98.63% and a specificity of 99.04%. Dehnavi et al. (2011) used the Principal Component Analysis (PCA) and Independent Component Analysis (ICA) methods to reduce extracted VCG features. They used a set of 60 ischemic and 10 physiological records and extracted a total of 22 VCG parameters. After the reduction, they acquired five features that served as parameters for input to the neural network classifier (NNC). Their designed model has achieved accuracy 86% in distinguishing HC from MI. They also used the classical method of IHD detection from ECG parameters (ST segment and T-wave morphology) and compared to VCG parameters, detection from ECG achieves lower accuracy (73%). Based on the results obtained in this study authors indicate that VCG has higher accuracy and sensitivity in automated detection IHD than ECG. Schuepbach et al. (2008) used cardiogoniometric leads (VCG C ) and classifies healthy (HC) and ischemic (IHD) patients. Coronary angiography (CA) was used as a reference method. They describe VCG as a set of parameters including: time scale, size and direction of vectors, ratios of R/T vectors and ST/T segments, and the percentage location of QRS and T loops in Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 8565909 Vondrak and Penhaker Review of Processing Vectorcardiographic Records information is addressed, would be solved by more frequent use of Frank lead system in clinical practice. As additional information for the detection of heart disease are used VCG features. Studies have shown that VCG features extracted from pathological records such as myocardial infarction, myocardial ischemia, myocardial scar or atrial fibrillation can reliably distinguish test groups from healthy controls. Current studies have confirmed that the VCG features analysis especially of the QRS complex in particular provides reliable detection results Correa et al. (2010); Yang et al. (2012);Tripathy and Dandapat (2017);Kan and Yang (2012). An overview of selected important comparative studies dealing with VCG features are summarized in Table 9.The relevant table describes the scientific work relevant in particular to this review. Central to the table, is the column VCG features, which shows the total number of VCG features for individual works and their most important features. For publications with a large number of features, only the number is given. Paramount is to evaluate the diagnostic significance of individual features, which may be different for different pathologies. Most studies only deal with the individualization of VCG features to a particular dataset. Thus, their results do not have to correspond to another database. It would be useful to compare the effectiveness of different features between the databases because each dataset is obtained by different measurement parameters. This fact is also correlated to the small number of available VCG datasets, whereonlyafewauthorsusetheirownmeasuredrecordsfor their study. The lack of records could be reduced by providing records by the authors or creating new datasets using modern devices. There are VCG features that are used in most publications and achieve high detection accuracy in terms of statistical evaluation. However, these features are not standardized for certain arrhythmias despite their possible diagnostic benefit, caused mainly by not consulting the results with cardiologists. The derivation and extraction of additional VCG features could further contribute to obtaining additional information necessary for the early diagnosis of heart disease. 6 CONCLUSION Vectorcardiography, which in recent years is more often analyzed mainly in the research field, and which represents a different approach in the representation of electrical activity of the heart, can help us to obtain information that can help in the early TABLE 9 | An overview of important comparative studies. Author Year Purpose VCG features Data collection Transformation method Kawahito et al. (2003) 2003 QRS, ST-T parameters study 2 features QRS vector difference, ST vector magnitude VCG measured by authors none Huang et al. (2011) 2011 MI detection 64 features, QRS, T vector magnitudes R-T peak angle PhysioNet PTB none Yang et al. (2012) 2012 MI detection 48 features Q, R, T—vector magnitude, R, T—vector angle, Angle between R and T-vector PhysioNet PTB none Correa et al. (2012) 2012 QRS, ST-T parameters study 8 features QRS—Volume, Planar Area, Ratio between Area and Perimeter, Perimeter, ST Vector Magnitude, ST segment Level, T-wave amplitude PhysioNet PTB Kors regress Correa et al. (2013) 2013 Cardiac ischemia detection 8 features PhysioNet PTB Kors regress Panagiotou et al. (2013) 2013 Myocardial scar detection 27 features R-width, T-width magnitude, R-peak, T-peak PhysioNet PTB, Cardiology Department (UHS-NHS) Dower’s inverse Dima et al. (2013) 2013 Myocardial scar detection 25 features PhysioNet PTB, Cardiology Department (UHS-NHS) Dower’s inverse Treskes et al. (2015) 2015 Myocardial ischemia detection 2 features ST vector, Ventricular gradient vector ECG measured by authors Kors regress Aranda et al. (2015) 2015 MI detection 98 features PhysioNet PTB Dower’s inverse Correa et al. (2016) 2016 MI detection 9 features QRS—Volume, Planar Area, Perimeter, Vector difference in ST segment and T-wave, ST-T Vector Magnitude Difference PhysioNet PTB none Sedaghat et al. (2016) 2016 QRS parameters Study 5 features QRS loop roundness, planarity, thickness, rotational angle, dihedral angle ECG and VCG measured by authors Dower’s inverse Sun et al. (2017) 2017 Myocardial ischemia screening 17 features PhysioNet PTB, STAFF III Dower’s inverse Tripathy and Dandapat (2017) 2017 MI detection 15 features PhysioNet PTB none Sharma and Sunkaria (2018) 2018 MI detection 10 features PhysioNet PTB none Gemmell et al. (2020) 2020 Myocardial scar detection 6 features Derived whole-torso computational model with simulations none Hafshejani et al. (2021) 2021 MI detection 48 features PhysioNet PTB none Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 85659016 Vondrak and Penhaker Review of Processing Vectorcardiographic Records diagnosis of heart disease. The main problem is that this method is not measured in clinical practice and the authors approach this method using transformation methods. The most important condition for the transformations is that diagnostic information must not be lost. Only an experienced cardiologist can evaluate the effect of transformation. If we only need a certain part of the heart revolution for processing, it is more appropriate to use specialized methods, which are focused on that part. The most accurate current methods are the IDT and the Kors regression transform but they lag slightly behind the accuracy of the P and T waves. This gap in the imperfection of linear transformation methods needs to be further addressed, due to the consequent better accuracy of transformation and preservation of diagnostic information. The new transformation method could refine VCG findings and provide physicians additional information for the early treatment of heart disease. Complications of applying transformation methods could be reduced by more frequent use of Frank lead system in clinical practice. In most cases processing of VCG is performed on already measured datasets and is performed so-called offline. Future intentions may lead to the creation of devices that will process data using already proven techniques in real time or on newly created data or datasets. To detect different pathologies, the authors use different VCG features applied to derived or directly measured VCGs. Certain features are often pointed out that achieve reliable pathological detection. The authors analyze various parts of the cardiac revolution, such as the QRS complex, where they achieve relatively high detection accuracy (sensitivity and specificity is often >90%) or the S-T region with sensitivity and specificity around 80%. However, different processing methods for each access are often used. To complete the knowledge, it would be appropriate to focus on a certain type of processing. Few authors also consult their results with cardiologists to confirm their importance. They merely point out differences in morphological properties to assess the new method. The assessment of certain features by cardiologists could help standardize the features for certain pathologies. Combinations of VCG features for different loops of cardiac cycle achieve promising results but this approach is rarely used. By combining the QRS complex and the ST segment features, greater differences can be achieved between groups HC and the pathology. It would be advisable to consider using multiple VCG parts (QRS loop, ST segment, T wave) for analysis rather than just one. Combining the above suggestions and discussing the results with cardiologists could help with visualization and gain new insights into the electrical activity of the heart. It would also be beneficial to increase the frequency of use of Frank lead system in clinical practice because as has been shown by more accessible and improving computer technology, the VCG method is suitable for the analysis of various heart diseases. AUTHOR CONTRIBUTIONS JV: Conceptualization, Investigation, Methodology, and Writing—review and editing. MP: Funding acquisition, Supervision, and Project administration. FUNDING The work and the contributions were supported by the project SV450021/2101, SP 2021/112 “Biomedical Engineering systems XVII.” REFERENCES Acar, B., and Koymen, H. (1999). SVD-based On-Line Exercise ECG Signal Orthogonalization. IEEE Trans. Biomed. Eng. 46, 311–321. doi:10.1109/10. 748984 Acar, B., Yi, G., Hnatkova, K., and Malik, M. (1999). Spatial, Temporal and Wavefront Direction Characteristics of 12-lead T-Wave Morphology. Med. Biol. Eng. Comput. 37, 574–584. doi:10.1007/bf02513351 Aranda, A., Bonizzi, P., Karel, J., and Peeters, R. (2015). Performance of Dower’s Inverse Transform and Frank lead System for Identification of Myocardial Infarction. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. 2015, 4495–4810. doi:10. 1109/EMBC.2015.7319393 Arini, P., Baglivo, F., Martínez, J., and Laguna, P. (2008). “Ventricular Repolarization Dispersion during Ischemia Course Measured by Temporal and Spatial Electrocardiographic Parameters,”in 2008 Computers in Cardiology, Bologna, Italy, 14-17 Sept. 2008 (IEEE), 323–326. doi:10.1109/ cic.2008.4749043 Bartall, H., Desser, K. B., Benchimol, A., and Massey, B. J. (1978). Assessment of Echocardiographic Left Atrialenlargement in Patients with Atrial Fibrillation. An Electrovectorcardiographic Study. J. Electrocardiol. 11, 269–272. doi:10. 1016/s0022-0736(78)80127-7 Belloch, J., Guillem, M., Climent, A., Millet, J., Husser, D., and Bollman, A. (2007). “Comparison of Different Methods for the Derivation of the Vectorcardiogram from the ECG and Morphology Descriptors,”in 2007 Computers in Cardiology, Durham, NC, USA, 30 Sept.-3 Oct. 2007 (IEEE), 435–438. doi:10.1109/cic.2007.4745515 Bizarro, R. O., O’Brien, P. C., Titus, J. L., and Smith, R. E. (1978). Vectorcardiographic Identification of Myocardial Scar: A Discriminative Study with Automatically Processed Vectorcardiographic Information. J. Electrocardiol. 11, 273–276. doi:10.1016/s0022-0736(78)80128-9 Bjerle, P., and Arvedson, O. (1986). Comparison of Frank Vectorcardiogram with Two Different Vectorcardiograms Derived from Conventional ECG-Leads. Proc. Eng. Found. 11, 13–26. Burch, G. E., and DePasquale, N. P. (1990). A History of Electrocardiography. San Francisco, CA: Norman Publishing. Burch, G. E. (1985). The History of Vectorcardiography. Med. Hist. 29, 103–131. doi:10.1017/s002572730007054x Burger, H. C., Van Milaan, J. B., and Den Boer, W. (1952). Comparison of Different Systems of Vectorcardiography. Heart 14, 401–405. doi:10.1136/hrt.14.3.401 Camm, A. J., Kirchhof, P., Lip, G. Y., Schotten, U., Savelieva, I., Ernst, S., et al. (2010). Guidelines for the Management of Atrial Fibrillation: The Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). Eur. Heart J. 31, 2369–2429. doi:10.1093/europace/euq350 Carlson, J., Havmöller, R., Herreros, A., Platonov, P., Johansson, R., and Olsson, B. (2005). Can Orthogonal lead Indicators of Propensity to Atrial Fibrillation Be Accurately Assessed from the 12-lead ECG? EP Europace 7, S39–S48. doi:10. 1016/j.eupc.2005.04.012 Chen, C.-H., and Kawai, C. (1978). Anterior QRS Loop in Hypertrophic Cardiomyopathy. J. Electrocardiol. 11, 129–136. doi:10.1016/s0022-0736(78)80102-2 Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 85659017 Vondrak and Penhaker Review of Processing Vectorcardiographic Records Chen, Y., and Yang, H. (2013). Wavelet Packet Analysis of Disease-Altered Recurrence Dynamics in the Long-Term Spatiotemporal Vectorcardiogram (VCG) Signals. 2013 35th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. (Embc) (Ieee), 2595–2598. doi:10.1109/EMBC.2013.6610071 Chou, T.-C. (1986). When Is the Vectorcardiogram superior to the Scalar Electrocardiogram? J. Am. Coll. Cardiol. 8, 791–799. doi:10.1016/s07351097(86)80419-3 Correa, R., Arini, P. D., Correa, L. S., Valentinuzzi, M., and Laciar, E. (2016). Identification of Patients with Myocardial Infarction. Methods Inf. Med. 55, 242–249. doi:10.3414/ME15-01-0101 Correa, R., Arini, P. D., Valentinuzzi, M. E., and Laciar, E. (2013). Novel Set of Vectorcardiographic Parameters for the Identification of Ischemic Patients. Med. Eng. Phys. 35, 16–22. doi:10.1016/j.medengphy.2012.03.005 Correa, R., Arini, P. D., Valentinuzzi, M., and Laciar, E. (2012). “Study of QRSLoop Parameters and Conventional ST-T Indexes for Identification of Ischemic and Healthy Subjects,”in 2012 Computing in Cardiology, Krakow, Poland, 9-12 Sept. 2012 (IEEE), 649–652. Correa, R., Laciar, E., Arini, P., and Jané, R. (2010). Analysis of QRS Loop in the Vectorcardiogram of Patients with Chagas’Disease. 2010 Annu. Int. Conf. IEEE Eng. Med. Biol., 2561–2564. doi:10.1109/IEMBS.2010.5626863 Cortez, D., Bos, J. M., and Ackerman, M. J. (2017a). Vectorcardiography Identifies Patients with Electrocardiographically Concealed Long QT Syndrome. Heart Rhythm 14, 894–899. doi:10.1016/j.hrthm.2017.03.003 Cortez, D. L., and Schlegel, T. T. (2010). When Deriving the Spatial QRS-T Angle from the 12-lead Electrocardiogram, Which Transform Is More Frank: Regression or Inverse Dower? J. Electrocardiol. 43, 302–309. doi:10.1016/j. jelectrocard.2010.03.010 Cortez, D., Schlegel, T. T., Ackerman, M. J., and Bos, J. M. (2017b). ECG-derived Spatial QRS-T Angle Is Strongly Associated with Hypertrophic Cardiomyopathy. J. Electrocardiol. 50, 195–202. doi:10.1016/j.jelectrocard.2016.10.001 Cortez, D., Sharma, N., Cavanaugh, J., Tuozo, F., Derk, G., Lundberg, E., et al. (2015). The Spatial QRS-T Angle Outperforms the Italian and Seattle ECGBased Criteria for Detection of Hypertrophic Cardiomyopathy in Pediatric Patients. J. Electrocardiol. 48, 826–833. doi:10.1016/j.jelectrocard.2015.07.016 Cortez, D., Sharma, N., Devers, C., Devers, E., and Schlegel, T. T. (2014). Visual Transform Applications for Estimating the Spatial QRS–T Angle from the Conventional 12-lead ECG: Kors Is Still Most Frank. J. Electrocardiol. 47, 12–19. doi:10.1016/j.jelectrocard.2013.09.003 Cruces, P. D., and Arini, P. D. (2016). A Novel Method for Cardiac Vector Velocity Measurement: Evaluation in Myocardial Infarction. Biomed. Signal Process. Control. 28, 58–62. doi:10.1016/j.bspc.2016.04.003 Dawson, D., Yang, H., Malshe, M., Bukkapatnam, S. T., Benjamin, B., and Komanduri, R. (2009). Linear Affine Transformations between 3-lead (Frank XYZ Leads) Vectorcardiogram and 12-lead Electrocardiogram Signals. J. Electrocardiol. 42, 622–630. doi:10.1016/j.jelectrocard.2009.05.007 Dehnavi, A. R. M., Farahabadi, I., Rabbani, H., Farahabadi, A., Mahjoob, M. P., and Dehnavi, N. R. (2011). Detection and Classification of Cardiac Ischemia Using Vectorcardiogram Signal via Neural Network. J. Res. Med. Sci. official J. Isfahan Univ. Med. Sci. 16, 136. Dellborg, M., Malmberg, K., Ryden, L., Svensson, A.-m., and Swedberg, K. (1995). Dynamic On-Line Vectorcardiography Improves and Simplifies In-Hospital Ischemia Monitoring of Patients with Unstable Angina. J. Am. Coll. Cardiol. 26, 1501–1507. doi:10.1016/0735-1097(95)00361-4 Diamant, U.-B., Jensen, S. M., Winbo, A., Stattin, E.-L., and Rydberg, A. (2013). Vectorcardiographic Recordings of the QT Interval in a Pediatric Long QT Syndrome Population. Pediatr. Cardiol. 34, 245–249. doi:10.1007/s00246-0120425-2 Diamant, U.-B., Winbo, A., Stattin, E.-L., Rydberg, A., Kesek, M., and Jensen, S. M. (2010). Two Automatic QT Algorithms Compared with Manual Measurement in Identification of Long QT Syndrome. J. Electrocardiol. 43, 25–30. doi:10. 1016/j.jelectrocard.2009.09.008 Dilaveris, P., Gialafos, E., Pantazis, A., Synetos, A., Triposkiadis, F., and Gialafos, J. (2001). The Spatial QRS-T Angle as a Marker of Ventricular Repolarisation in Hypertension. J. Hum. Hypertens. 15, 63–70. doi:10.1038/sj.jhh.1001129 Dima,S.-M.,Panagiotou,C.,Mazomenos,E.B.,Rosengarten,J.A.,Maharatna,K., Gialelis,J.V.,etal.(2013).OntheDetection of Myocadial Scar Based on ECG/VCG Analysis. IEEE Trans. Biomed. Eng. 60, 3399–3409. doi:10.1109/tbme.2013.2279998 Dower, G. E. (1968). A lead Synthesizer for the Frank System to Simulate the Standard 12-lead Electrocardiogram. J. Electrocardiol. 1, 101–116. doi:10.1016/ s0022-0736(68)80013-5 Dower, G. E., Machado, H. B., and Osborne, J. (1980). On Deriving the Electrocardiogram from Vectorcardiographic Leads. Clin. Cardiol. 3, 87–95. doi:10.1002/clc.1980.3.2.87 Dower, G. E., and Machado, H. B. (1979). XYZ Data Interpreted by a 12-lead Computer Program Using the Derived Electrocardiogram. J. Electrocardiol. 12, 249–261. doi:10.1016/s0022-0736(79)80058-8 Dubin, D. (2000). Rapid Interpretation of EKG’s: An Interactive Course. Tampa, FL: Cover Publishing Company. Duchosal, P., and Sulzer, R. (1949). La Vectorcardiographie (Bibliotheca Cardiologica Fasc 3). Durrer, D., Van Dam, R. T., Freud, G., Janse, M., Meijler, F., and Arzbaecher, R. (1970). Total Excitation of the Isolated Human Heart. Circulation 41, 899–912. doi:10.1161/01.cir.41.6.899 Edenbrand, L., Pahlm, O., Lyttkens, K., and Albrechtsson, U. (1990). Vectorcardiogram More Sensitive Than 12-lead ECG in the Detection of Inferior Myocardial Infarction. Clin. Physiol. 10, 551–559. doi:10.1111/j. 1475-097x.1990.tb00447.x Edenbrandt, L., and Pahlm, O. (1988). Vectorcardiogram Synthesized from a 12lead ECG: Superiority of the Inverse Dower Matrix. J. Electrocardiol. 21, 361–367. doi:10.1016/0022-0736(88)90113-6 El Haddad, M., Vervloet, D., Taeymans, Y., De Buyzere, M., Bové, T., Stroobandt, R., et al. (2016). Diagnostic Accuracy of a Novel Method for Detection of Acute Transmural Myocardial Ischemia Based upon a Self-Applicable 3-lead Configuration. J. Electrocardiol. 49, 192–201. doi:10.1016/j.jelectrocard.2015. 11.007 Eriksson, M., Hofman-Bang, C., Persson, H., and Tornvall, P. (2007). Limited Prognostic Value of Noninvasive Assessment of Reperfusion by Continuous Vectorcardiography in an Unselected Cohort of Patients with Acute STElevation Myocardial Infarction Treated with Thrombolysis. J. Electrocardiol. 40, 305–310. doi:10.1016/j.jelectrocard.2006.12.004 Fesmire, F. M., Eriksson, S. V., Stout, P. K., Wojcik, J. F., and Wharton, D. R. (2002). Use of Baseline ST-Vector Magnitude to Identify Electrocardiographic Injury in Patients with Suspected Acute Myocardial Infarction. Am. J. Emerg. Med. 20, 535–540. doi:10.1053/ajem.2002.34801 Filipova, S., Vachulova, A., Skultetyova, D., Hlivak, P., and Gaspar, L. (2017). Abstract P350: Abnormal VCG and Hi-Res Parameters of P and QRS Complex in Hypertensive Patients Without/with Paroxysmal Atrial Fibrillation. Hypertension 70, AP350. doi:10.1161/hyp.70.suppl_1.p350 Filos, D., Chouvarda, I., Tachmatzidis, D., Vassilikos, V., and Maglaveras, N. (2017). Beat-to-beat P-Wave Morphology as a Predictor of Paroxysmal Atrial Fibrillation. Computer Methods Programs Biomed. 151, 111–121. doi:10.1016/j. cmpb.2017.08.016 Frank, E. (1956). An Accurate, Clinically Practical System for Spatial Vectorcardiography. circulation 13, 737–749. doi:10.1161/01.cir.13.5.737 Frank, E. (1954). General Theory of Heart-Vector Projection. Circ. Res. 2, 258–270. doi:10.1161/01.res.2.3.258 Ge, D. (2008). “Detecting Myocardial Infraction Using VCG Leads,”in 2008 2nd International Conference on Bioinformatics and Biomedical Engineering, Shanghai, China, 16-18 May 2008 (IEEE), 2217–2220. doi:10.1109/icbbe. 2008.885 Gemmell, P. M., Gillette, K., Balaban, G., Rajani, R., Vigmond, E. J., Plank, G., et al. (2020). A Computational Investigation into Rate-Dependant Vectorcardiogram Changes Due to Specific Fibrosis Patterns in Non-ischæmic Dilated Cardiomyopathy. Comput. Biol. Med. 123, 103895. doi:10.1016/j. compbiomed.2020.103895 Goernig, M., Hoeffling, B., Lau, S., Figulla, H. R., and Haueisen, J. (2015). T Vector and Loop Characteristics Improve Detection of Myocardial Injury after Infarction. J. Med. Biol. Eng. 35, 381–386. doi:10.1007/s40846-015-0041-8 Goldenberg, I., and Moss, A. J. (2008). Long QT Syndrome. J. Am. Coll. Cardiol. 51, 2291–2300. doi:10.1016/j.jacc.2008.02.068 Golub, G. H., and Van Loan, C. F. (1996). Matrix Computations. Baltimore: Johns Hopkins studies in the mathematical sciences. Grishman, A., Borun, E. R., and Jaffe, H. L. (1951). Spatial Vectorcardiography: Technique for the Simultaneous Recording of the Frontal, Sagittal, and Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 85659018 Vondrak and Penhaker Review of Processing Vectorcardiographic Records Horizontal Projections. I. Am. Heart J. 41, 483–493. doi:10.1016/0002-8703(51) 90017-8 Guillem, M., Sahakian, A. V., and Swiryn, S. (2006). “Derivation of Orthogonal Leads from the 12-lead ECG. Accuracy of a Single Transform for the Derivation of Atrial and Ventricular Waves,”in 2006 Computers in Cardiology, Valencia, Spain, 17-20 Sept. 2006 (IEEE), 249–252. Guillem, M. S., Climent, A. M., Bollmann, A., Husser, D., Millet, J., and Castells, F. (2009). Limitations of Dower’s Inverse Transform for the Study of Atrial Loops during Atrial Fibrillation. Pacing Clin. Electrophysiol. 32, 972–980. doi:10.1111/ j.1540-8159.2009.02426.x Guillem, M. S., Sahakian, A. V., and Swiryn, S. (2008). Derivation of Orthogonal Leads from the 12-lead Electrocardiogram. Performance of an Atrial-Based Transform for the Derivation of P Loops. J. Electrocardiol. 41, 19–25. doi:10. 1016/j.jelectrocard.2007.07.003 Guldenring, D., Finlay, D. D., Strauss, D. G., Galeotti, L., Nugent, C. D., Donnelly, M. P., et al. (2012). Transformation of the Mason-Likar 12-lead Electrocardiogram to the Frank Vectorcardiogram. 2012 Annu. Int. Conf. IEEE Eng. Med. Biol. Soc., 677–680. doi:10.1109/EMBC.2012.6346022 Hafshejani, N. J., Mehridehnavi, A., Hajian, R., Boudagh, S., and Behjati, M. (2021). Identification of Myocardial Infarction Using Morphological Features of Electrocardiogram and Vectorcardiogram. IET Signal. Process. 15, 674–685. doi:10.1049/sil2.12072 Hasan, M. A., and Abbott, D. (2016). A Review of Beat-To-Beat Vectorcardiographic (VCG) Parameters for Analyzing Repolarization Variability in ECG Signals. Biomed. Engineering/Biomedizinische Technik 61, 3–17. doi:10.1515/bmt-2015-0005 Hasan, M. A., Abbott, D., and Baumert, M. (2012a). “Beat-to-beat Spatial and Temporal Analysis for QRS-T Morphology,”in 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, San Diego, CA, USA, 28 Aug.-1 Sept. 2012 (IEEE), 4193. –4195. doi:10.1109/EMBC. 2012.6346891 Hasan, M. A., Abbott, D., and Baumert, M. (2012b). Beat-to-beat Vectorcardiographic Analysis of Ventricular Depolarization and Repolarization in Myocardial Infarction. PLoS One 7, e49489. doi:10.1371/ journal.pone.0049489 Hernandez, A. A., Bonizzi, P., Karel, J., and Peeters, R. (2018). “Myocardial Ischemia Diagnosis Using a Reduced Lead System,”in 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Honolulu, HI, USA, 18-21 July 2018 (IEEE), 5302–5305. doi:10.1109/EMBC.2018.8513511 Howard, F. H. (1946). A Method for the Construction of the Vectorcardiogram from the Einthoven Electrocardiogram. Am. Heart J. 31, 191–219. doi:10.1016/ 0002-8703(46)90610-2 Huang, C.-S., Ko, L.-W., Lu, S.-W., Chen, S.-A., and Lin, C.-T. (2011). “A Vectorcardiogram-Based Classification System for the Detection of Myocardial Infarction,”in 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, Boston, MA, USA, 30 Aug.-3 Sept. 2011 (IEEE), 973–976. doi:10.1109/IEMBS.2011.6090220 Huebner, T., Schuepbach, W. M., Seeck, A., Sanz, E., Meier, B., Voss, A., et al. (2010). Cardiogoniometric Parameters for Detection of Coronary Artery Disease at Rest as a Function of Stenosis Localization and Distribution. Med. Biol. Eng. Comput. 48, 435–446. doi:10.1007/s11517-010-0594-1 Hurd, H., 2nd, Starling, M. R., Crawford, M. H., Dlabal, P., and O’rourke, R. (1981). Comparative Accuracy of Electrocardiographic and Vectorcardiographic Criteria for Inferior Myocardial Infarction. Circulation 63, 1025–1029. doi:10.1161/01.cir.63.5.1025 Iwaniec, J., Iwaniec, M., and Kalukiewicz, A. (2018). Application of Vectorcardiography and Recurrence-Based Methods to Analysis of ECG Signals. MATEC Web of Conferences 241, 01015. doi:10.1051/matecconf/ 201824101015 Jaros, R., Martinek, R., and Danys, L. (2019). Comparison of Different Electrocardiography with Vectorcardiography Transformations. Sensors 19, 3072. doi:10.3390/s19143072 Jim, M.-H., Siu, C.-W., Chan, A. O.-O., Chan, R. H.-W., Lee, S. W.-L., and Lau, C.- P. (2006). Prognostic Implications of PR-segment Depression in Inferior Leads in Acute Inferior Myocardial Infarction. Clin. Cardiol. Int. Indexed PeerReviewed J. Adv. Treat. Cardiovasc. Dis. 29, 363–368. doi:10.1002/clc. 4960290809 Jimenez, E., El-Bokl, A., and Cortez, D. (2021). Vectorcardiography as a Prognostic Tool in Hypertrophic Cardiomyopathy. J. Electrocardiol. 68, 80–84. doi:10. 1016/j.jelectrocard.2021.08.004 Kan, C., and Yang, H. (2012). “Dynamic Spatiotemporal Warping for the Detection and Location of Myocardial Infarctions,”in 2012 IEEE International Conference on Automation Science and Engineering (CASE), Seoul, Korea (South), 20-24 Aug. 2012 (IEEE), 1046–1051. doi:10.1109/coase.2012.6386354 Karisik, F., Schmidt, M., and Baumert, M. (2019). Beat-to-Beat Analysis of Vectorcardiogram by Inhomogeneous Template Adaptation. 2019 41st Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. (Embc) (Ieee),83–86. doi:10. 1109/EMBC.2019.8857436 Karsikas, M., Noponen, K., Tulppo, M., Huikuri, H., and Seppänen, T. (2009). “Beat-to-beat Variation of Three-Dimensional QRS-T Angle Measures during Exercise Test,”in 2009 36th Annual Computers in Cardiology Conference (CinC), Park City, UT, USA, 13-16 Sept. 2009 (IEEE), 125–128. Kataoka, Y., and Tomoike, H. (2021). Spatial Feature Extraction of Vectorcardiography via Minimum Volume Ellipsoid Enclosure in Classifying Left Ventricular Hypertrophy. 2021 43rd Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. (Embc) (Ieee), 625–628. doi:10.1109/embc46164.2021. 9630932 Kawahito, S., Kitahata, H., Tanaka, K., Nozaki, J., and Oshita, S. (2003). Dynamic QRS-Complex and ST-Segment Monitoring by Continuous Vectorcardiography during Carotid Endarterectomy. Br. J. Anaesth. 90, 142–147. doi:10.1093/bja/aeg031 Kawano, S., Hiraoka, M., and Sawanobori, T. (1988). Electrocardiographic Features of P Waves from Patients with Transient Atrial Fibrillation. Jpn. Heart J. 29, 57–67. doi:10.1536/ihj.29.57 Keshtkar, A., Seyedarabi, H., Sheikhzadeh, P., and Rasta, S. H. (2013). Discriminant Analysis between Myocardial Infarction Patients and Healthy Subjects Using Wavelet Transformed Signal Averaged Electrocardiogram and Probabilistic Neural Network. J. Med. signals sensors 3, 225. Kimura, N. (1939). Study on Heart Function by Vectorcardiography of ThreeDimensional Projection. Jap. Circ. J. 5, 177. Korhonen, P., Väisänen, J., Nousiainen, J., Huhtala, H., Yli-Mäyry, S., and Raatikainen, P. (2013). Detection of Concurrent Atrial Ischemia with Continuous Monitoring of Dynamic PR-segment Changes in Patients with Acute Myocardial Infarction. J. Electrocardiol. 46, 324–330. doi:10.1016/j. jelectrocard.2013.03.013 Kornreich, F., Block, P., and Brismee, D. (1974). The Missing Waveform Information in the Orthogonal Electrocardiogram (Frank Leads) III. Computer Diagnosis of Angina Pectoris from”Maximal”QRS Surface Waveform Information at Rest. Circulation 49, 1212–1222. doi:10.1161/01.cir.49.6.1212 Kors, J. A., Talmon, J. L., and Van BemmeL, J. H. (1986). Multilead ECG Analysis. Comput. Biomed. Res. 19, 28–46. doi:10.1016/0010-4809(86)90004-2 Kors, J., Van Herpen, G., Sittig, A., and Van Bemmel, J. (1990). Reconstruction of the Frank Vectorcardiogram from Standard Electrocardiographic Leads: Diagnostic Comparison of Different Methods. Eur. Heart J. 11, 1083–1092. doi:10.1093/oxfordjournals.eurheartj.a059647 Kulbertus, H., and de Leval-Rutten, F. (1974). Vectorcardiographic Study of QRS in Patients with Transvenous Pacemakers and Myocardial Infarction. J. Electrocardiol. 7, 27–33. doi:10.1016/s0022-0736(74)80005-1 Laufberger, V. (1982). Octant Vectorcardiography and Automatic Diagnosis of Coronary Artery Disease. Physiologia bohemoslovaca 31, 485–495. Levkov, C. L. (1987). Orthogonal Electrocardiogram Derived from the Limb and Chest Electrodes of the Conventional 12-lead System. Med. Biol. Eng. Comput. 25, 155. doi:10.1007/bf02442844 Lloyd, J. D. (2009). Heart Disease and Stroke Statistics-2009 Update: A Report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee Malmivuo, J., and Plonsey, R. (1995). Bioelectromagnetism: Principles and Applications of Bioelectric and Biomagnetic Fields. USA: Oxford University Press. Man, S., Algra, A. M., Schreurs, C. A., Borleffs, C. J. W., Scherptong, R. W., Van Erven, L., et al. (2011). Influence of the Vectorcardiogram Synthesis Matrix on the Power of the Electrocardiogram-Derived Spatial QRS-T Angle to Predict Arrhythmias in Patients with Ischemic Heart Disease and Systolic Left Ventricular Dysfunction. J. Electrocardiol. 44, 410–415. doi:10.1016/j. jelectrocard.2011.04.007 Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 85659019 Vondrak and Penhaker Review of Processing Vectorcardiographic Records Man, S., Rahmattulla, C., Maan, A. C., Holman, E., Bax, J. J., van der Wall, E. E., et al. (2012). Role of the Vectorcardiogram-Derived Spatial QRS-T Angle in Diagnosing Left Ventricular Hypertrophy. J. Electrocardiol. 45, 154–160. doi:10. 1016/j.jelectrocard.2011.10.001 Maron, B. J. (1997). Hypertrophic Cardiomyopathy. The Lancet 350, 127–133. doi:10.1016/s0140-6736(97)01282-8 Martínez, J. P., Olmos, S., Wagner, G., and Laguna, P. (2006). Characterization of Repolarization Alternans during Ischemia: Time-Course and Spatial Analysis. IEEE Trans. Biomed. Eng. 53, 701–711. doi:10.1109/tbme.2006.870233 Mason, R. E., and Likar, I. (1966). A New System of Multiple-lead Exercise Electrocardiography. Am. Heart J. 71, 196–205. doi:10.1016/0002-8703(66)90182-7 Matveev, M., Krasteva, V., Naydenov, S., and Donova, T. (2007). Possibilities of Signal-Averaged Orthogonal and Vector Electrocardiography for Locating and Evaluating the Size of Acute Myocardial Infarction. J. Electrocardiol. 40, S62–S63. doi:10.1016/j.jelectrocard.2007.03.059 McFee, R., and Parungao, A. (1961). An Orthogonal lead System for Clinical Electrocardiography. Am. Heart J. 62, 93–100. doi:10.1016/0002-8703(61) 90488-4 Merri, M., Farden, D. C., Mottley, J. G., and Titlebaum, E. L. (1990). Sampling Frequency of the Electrocardiogram for Spectral Analysis of the Heart Rate Variability. IEEE Trans. Biomed. Eng. 37, 99–106. doi:10.1109/10.43621 Milnor, W. R., Talbot, S. A., and Newman, E. V. (1953). A Study of the Relationship between Unipolar Leads and Spatial Vectorcardiograms, Using the Panoramic Vectorcardiograph. Circulation 7, 545–557. doi:10.1161/01.cir.7.4.545 Murata, K., Kurihara, H., Hosoda, S., Ikeda, M., and Seki, M. (1964). Frank lead Vectorcardiogram in Left Ventricular Hypertrophy. Jpn. Heart J. 5, 543–548. doi:10.1536/ihj.5.543 Myerburg, R. J., and Junttila, M. J. (2012). Sudden Cardiac Death Caused by Coronary Heart Disease. Circulation 125, 1043–1052. doi:10.1161/ circulationaha.111.023846 Ng, J., Sahakian, A. V., Fisher, W. G., and Swiryn, S. (2004). Surface ECG Vector Characteristics of Organized and Disorganized Atrial Activity during Atrial Fibrillation. J. Electrocardiol. 37, 91–97. doi:10.1016/j.jelectrocard.2004.08.031 Nguyên, U. C., Claridge, S., Vernooy, K., Engels, E. B., Razavi, R., Rinaldi, C. A., et al. (2018). Relationship between Vectorcardiographic QRSarea, Myocardial Scar Quantification, and Response to Cardiac Resynchronization Therapy. J. Electrocardiol. 51, 457–463. doi:10.1016/j.jelectrocard.2018.01.009 Okafor, O., Umar, F., Zegard, A., van Dam, P., Walton, J., Stegemann, B., et al. (2020). Effect of QRS Area Reduction and Myocardial Scar on the Hemodynamic Response to Cardiac Resynchronization Therapy. Heart Rhythm 17, 2046–2055. doi:10.1016/j.hrthm.2020.07.025 Panagiotou, C., Dima, S.-M., Mazomenos, E. B., Rosengarten, J., Maharatna, K., Gialelis, J., et al. (2013). Detection of Myocardial Scar from the VCG Using a Supervised Learning Approach. 2013 35th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. (Embc), 7326–7329. doi:10.1109/EMBC.2013.6611250 Papouchado, M., Walker, P., James, M., and Clarke, L. (1987). Fundamental Differences between the Standard 12–lead Electrocardiograph and the Modified (Mason—Likar) Exercise lead System. Eur. Heart J. 8, 725–733. doi:10.1093/eurheartj/8.7.725 Perez Riera, A. R., Uchida, A. H., Filho, C. F., Meneghini, A., Ferreira, C., Schapacknik, E., et al. (2007). Significance of Vectorcardiogram in the Cardiological Diagnosis of the 21st century. Clin. Cardiol. Int. Indexed PeerReviewed J. Adv. Treat. Cardiovasc. Dis. 30, 319–323. doi:10.1002/clc.14 Persson, E., Pettersson, J., Ringborn, M., Sörnmo, L., Warren, S. G., Wagner, G. S., et al. (2006). Comparison of ST-Segment Deviation to Scintigraphically Quantified Myocardial Ischemia during Acute Coronary Occlusion Induced by Percutaneous Transluminal Coronary Angioplasty. Am. J. Cardiol. 97, 295–300. doi:10.1016/j.amjcard.2005.08.044 Rad, M. M., Wijntjens, G. W., Engels, E. B., Blaauw, Y., Luermans, J. G., Pison, L., et al. (2016). Vectorcardiographic QRS Area Identifies Delayed Left Ventricular Lateral wall Activation Determined by Electroanatomic Mapping in Candidates for Cardiac Resynchronization Therapy. Heart Rhythm 13, 217–225. doi:10. 1016/j.hrthm.2015.07.033 Romero, D., Ringborn, M., Laguna, P., Pahlm, O., and Pueyo, E. (2010). “A Vectorial Approach for Evaluation of Depolarization Changes during Acute Myocardial Ischemia,”in 2010 Computing in Cardiology, Belfast, UK, 26-29 Sept. 2010 (IEEE), 265–268. Rubel, P., Benhadid, I., and Fayn, J. (1991). Quantitative Assessment of Eight Different Methods for Synthesizing Frank VCGs from Simultaneously Recorded Standard ECG Leads. J. Electrocardiol. 24, 197–202. doi:10.1016/ s0022-0736(10)80045-7 Sadhukhan, D., Datta, J., Pal, S., and Mitra, M. (2017). “Automated Identification of Myocardial Infarction Using a Single Vectorcardiographic Feature,”in International Conference on Modelling and Simulation (Springer), 641–651. Schellong, F., Heller, S., and Schwingel, E. (1937). Das Vektordiagramm, eine Untersuchungsmethode des Herzens. Ztschr. F. Kreislaufforsch. 29, 497. Schmitt,O.H.,andSimonson,E.(1955).SymposiumonElectrocardiography and Vectorcardiography: The Present Status of Vectorcardiography. AMA Arch. Intern. Med. 96, 574–590. doi:10.1001/archinte.1955. 00250160016002 Schreurs, C. A., Algra, A. M., Man, S.-C., Cannegieter, S. C., van der Wall, E. E., Schalij, M. J., et al. (2010). The Spatial QRS-T Angle in the Frank Vectorcardiogram: Accuracy of Estimates Derived from the 12-lead Electrocardiogram. J. Electrocardiol. 43, 294–301. doi:10.1016/j.jelectrocard. 2010.03.009 Schuepbach, W. M., Emese, B., Loretan, P., Mallet, A., Duru, F., Sanz, E., et al. (2008). Non-invasive Diagnosis of Coronary Artery Disease Using Cardiogoniometry Performed at Rest. Swiss Med. weekly 138, 230–238. doi:10.2008/15/smw-12040 Sedaghat, G., Ghafoori, E., Waks, J. W., Kabir, M. M., Shvilkin, A., Josephson, M. E., et al. (2016). Quantitative Assessment of Vectorcardiographic Loop Morphology. J. Electrocardiol. 49, 154–163. doi:10.1016/j.jelectrocard.2015. 12.014 Seeck, A., Garde, A., Schuepbach, M., Giraldo, B., Sanz, E., Huebner, T., et al. (2009). “Diagnosis of Ischemic Heart Disease with Cardiogoniometry—Linear Discriminant Analysis versus Support Vector Machines,”in 4th European Conference of the International Federation for Medical and Biological Engineering (Springer), 389–392. doi:10.1007/978-3-540-89208-3_92 Selvester, R., Sanmarco, M., Solomon, J., and Wagner, G. (1982). “Methods for Determining Infarct Size, ECG: QRS Change,”in Myocardial Infarction: Measurement And Intervention. Editor G. S. Wagner (The Hague, Boston, London: Martinus Nijhoff). Selvester, R., Solomon, J., and Sapoznikov, D. (1979). Computer Simulation of the Electrocardiogram. Computer Tech. Cardiol. 9, 417–453. Selvester, R., Wagner, J., and Rubin, H. (1971). “Quantitation of Myocardial Infarct Size and Location by Electrocardiogram and Vectorcardiogram,”in Quantitation in Cardiology (Springer), 31–44. doi:10.1007/978-94-0102927-8_4 Sharma, L. D., and Sunkaria, R. K. (2018). Inferior Myocardial Infarction Detection Using Stationary Wavelet Transform and Machine Learning Approach. Signal. Image Video Process. 12, 199–206. doi:10.1007/s11760-017-1146-z Shvilkin, A., Bojovic, B., Vajdic, B., Gussak, I., Ho, K. K., Zimetbaum, P., et al. (2010). Vectorcardiographic and Electrocardiographic Criteria to Distinguish New and Old Left Bundle branch Block. Heart Rhythm 7, 1085–1092. doi:10. 1016/j.hrthm.2010.05.024 Stovicek, P., Stojan, M., Kasalický, J., Vondracek, V., Vojacek, J., Linhart, A., et al. (1993). Analysis of Ventricular Activation in Patients with Chronic Non-Q Wave Myocardial Infarction: Comparison with Left Ventricular Asynergy and Myocardial Perfusion Defects. Physiol. Res. 42, 109–117. Strauss, D. G., and Selvester, R. H. (2009). The QRS Complex-A Biomarker that” Images”the Heart: QRS Scores to Quantify Myocardial Scar in the Presence of normal and Abnormal Ventricular Conduction. J. Electrocardiol. 42, 85. doi:10. 1016/j.jelectrocard.2008.07.011 Sun, Z., Duan, X., Nie, X., Deng, Y., Jiao, B., and Wang, D. (2017). “A Screening System for Myocardial Ischemia Based on Pathophysiological Vectorcardiogram,”in 2017 International Conference on Sensing, Diagnostics, Prognostics, and Control (SDPC), Shanghai, China, 16-18 Aug. 2017 (IEEE), 339–344. doi:10.1109/sdpc.2017.71 Sur, S., Han, L., and Tereshchenko, L. G. (2013). Comparison of Sum Absolute QRST Integral, and Temporal Variability in Depolarization and Repolarization, Measured by Dynamic Vectorcardiography Approach, in Healthy Men and Women. PloS one 8, e57175. doi:10.1371/journal.pone. 0057175 Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 85659020 Vondrak and Penhaker Review of Processing Vectorcardiographic Records Tereshchenko, L. G., Han, L., Cheng, A., Marine, J. E., Spragg, D. D., Sinha, S., et al. (2010). Beat-to-beat Three-Dimensional ECG Variability Predicts Ventricular Arrhythmia in ICD Recipients. Heart Rhythm 7, 1606–1613. doi:10.1016/j. hrthm.2010.08.022 Toledo, E., and Wagner, G. (2007). HyperQ-new Horizons in Ischemia Detection. J. Electrocardiol. 40, S37. doi:10.1016/j.jelectrocard.2006.10.028 Tomey, M. I., and Gidwani, U. K. (2016). “Acute Cardiac Ischemia,”in Critical Care. Editors J. M. Oropello, S. M. Pastores, and V. Kvetan (New York, NY: McGraw-Hill Education). Treskes, R. W., Ter Haar, C. C., Man, S., De Jongh,M. C., Maan, A. C., Wolterbeek, R., et al. (2015). Performance of ST and Ventricular Gradient Difference Vectors in Electrocardiographic Detection of Acute Myocardial Ischemia. J. Electrocardiol. 48, 498–504. doi:10.1016/j.jelectrocard.2015.04.016 Tripathy, R., and Dandapat, S. (2017). Detection of Myocardial Infarction from Vectorcardiogram Using Relevance Vector Machine. Signal. Image Video Process. 11, 1139–1146. doi:10.1007/s11760-017-1068-9 van Bemmel, J. H., Kors, J. A., and van Herpen, G. (1992). Combination of Diagnostic Classifications from ECG and VCG Computer Interpretations. J. Electrocardiol. 25, 126–130. doi:10.1016/0022-0736(92)90078-e van Deursen, C. J., Vernooy, K., Dudink, E., Bergfeldt, L., Crijns, H. J., Prinzen, F. W., et al. (2015a). Vectorcardiographic QRS Area as a Novel Predictor of Response to Cardiac Resynchronization Therapy. J. Electrocardiol. 48, 45–52. doi:10.1016/j.jelectrocard.2014.10.003 van Deursen, C. J., Wecke, L., van Everdingen, W. M., Ståhlberg, M., Janssen, M. H., Braunschweig, F., et al. (2015b). Vectorcardiography for Optimization of Stimulation Intervals in Cardiac Resynchronization Therapy. J. Cardiovasc. translational Res. 8, 128–137. doi:10.1007/s12265-015-9615-7 van Hellemond, I. E., Bouwmeester, S., Olson, C. W., Hassell, M., Bøtker, H. E., Kaltoft, A. K., et al. (2013). Consideration of QRS Complex in Addition to ST Segment Abnormalities in the Estimation of the ‘risk Region’during Acute Inferior Myocardial Infarction. J. Electrocardiol. 46, 215–220. doi:10.1016/j. jelectrocard.2013.02.004 van Oosterom, A., Ihara, Z., Jacquemet, V., and Hoekema, R. (2007). Vectorcardiographic lead Systems for the Characterization of Atrial Fibrillation. J. Electrocardiol. 40, 343–e1. doi:10.1016/j.jelectrocard.2006.08.002 Villongco, C. T., Krummen, D. E., Stark, P., Omens, J. H., and McCulloch, A. D. (2014). Patient-specific Modeling of Ventricular Activation Pattern Using Surface ECG-Derived Vectorcardiogram in Bundle branch Block. Prog. Biophys. Mol. Biol. 115, 305–313. doi:10.1016/j.pbiomolbio.2014.06.011 Voulgari, C., and Tentolouris, N. (2009). Assessment of the Spatial QRS-T Angle by Vectorcardiography: Current Data and Perspectives. Curr. Cardiol. Rev. 5, 251. doi:10.2174/157340309789317850 Vozda, M., and Cerny, M. (2015). Methods for Derivation of Orthogonal Leads from 12-lead Electrocardiogram: A Review. Biomed. Signal Process. Control. 19, 23–34. doi:10.1016/j.bspc.2015.03.001 Waks, J. W., Soliman, E. Z., Henrikson, C. A., Sotoodehnia, N., Han, L., Agarwal, S. K., et al. (2015). Beat-to-beat Spatiotemporal Variability in the T Vector Is Associated with Sudden Cardiac Death in Participants without Left Ventricular Hypertrophy: The Atherosclerosis Risk in Communities (ARIC) Study. J. Am. Heart Assoc. 4, e001357. doi:10.1161/JAHA.114.001357 Waller, A. D. (1887). A Demonstration on Man of Electromotive Changes Accompanying the Heart’s Beat. J. Physiol. 8, 229–234. doi:10.1113/jphysiol. 1887.sp000257 Warner,R.,Hill,N.,Sheehe,P.,Mookherjee,S.,Fruehan,C.,andSmulyan,H. (1982). Improved Electrocardiographic Criteria for the Diagnosis of Inferior Myocardial Infarction. Circulation 66, 422–428. doi:10.1161/01. cir.66.2.422 Wecke, L., Poçi, D., Schwieler, J., Johansson, B., Edvardsson, N., Lundahl, G., et al. (2013). Vectorcardiography Shows Cardiac Memory and Repolarization Heterogeneity after Ablation of Accessory Pathways Not Apparent on ECG. Int. J. Cardiol. 166, 152–157. doi:10.1016/j.ijcard.2011.10.106 Wickline, S. A., and McNamara, J. J. (1978). Vectorcardiographic Quantification of Infarct Size in Baboons. Circulation 57, 910–920. doi:10.1161/01.cir.57.5.910 Wolf, H., Rautaharju, P., Unite, V., and Stewart, J. (1976). Evaluation of Synthesized Standard 12 Leads and Frank Vector Leads1. Electrocardiology 16, 87–97. doi:10.1159/000398375 Yamashita, S., Date, T., Yamane, T., Matsuo, S., Ito, K., Narui, R., et al. (2011). Noninvasive Vectorcardiographic Evaluation of Pulmonary Vein–Atrial Reconnection after Pulmonary Vein Isolation for Atrial Fibrillation. Heart rhythm 8, 1398–1403. doi:10.1016/j.hrthm.2011.04.023 Yang, H., Bukkapatnam, S. T., Le, T., and Komanduri, R. (2012). Identification of Myocardial Infarction (MI) Using Spatio-Temporal Heart Dynamics. Med. Eng. Phys. 34, 485–497. doi:10.1016/j.medengphy.2011.08.009 Yang, H., Kan, C., Liu, G., and Chen, Y. (2013). Spatiotemporal Differentiation of Myocardial Infarctions. IEEE Trans. Automation Sci. Eng. 10, 938–947. doi:10. 1109/tase.2013.2263497 Yang, H. (2010). Multiscale Recurrence Quantification Analysis of Spatial Cardiac Vectorcardiogram Signals. IEEE Trans. Biomed. Eng. 58, 339–347. doi:10.1109/ TBME.2010.2063704 Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Copyright © 2022 Vondrak and Penhaker. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Physiology | www.frontiersin.org March 2022 | Volume 13 | Article 85659021 Vondrak and Penhaker Review of Processing Vectorcardiographic Records