Electrocardiographic Imaging for Cardiac Arrhythmias and Resynchronization Therapy

Helder Pereira; Steven Niederer; Christopher A. Rinaldi


Europace. 2020;22(10):1447-1462. 

In This Article

Limitations of Electrocardiographic Imaging

Despite being a novel diagnostic tool that is capable of providing essential information about various heart rhythm disorders non-invasively, ECGi has some limitations that should be addressed. The procedure itself is fairly complicated; it requires complex recording with hundreds of leads for a single patient, as well as dedicated and experienced personnel to obtain and interpret the ECGi findings. Regarding the number of electrodes, 125 seems to be the most suitable for capturing sensitive body-surface regions by an analysis of constructed effort gain plots. Several designed ECGi methods exceed this number, with the principle that a high number of electrodes will meet clinical requirements and improve the resolution of activation-time imaging.[71] The application of close to 300 electrodes on the patient's chest can be both time-consuming and inconvenient. While this problem has been improved by the implementation of a pre-constructed multi-electrode vest (Figure 1), it still requires added time for CT or MRI scans. Although the vest makes its application in a clinical setting easier, it too has limitations. Since it has to be fitted to different torso measurements, limited adaptability can pose a problem in the case of underweight or obese patients, or in patients with chest imperfections or deformities. The equipment itself has been reported to be uncomfortable and have limited tolerability by some patients, particularly in situations where longer recording sessions are required.[72] Moreover, the adhesive used to fix the electrodes might irritate the skin. Further technical developments seek to facilitate easier and faster application of the equipment.[36]

The cost of the procedure is another concern. Compared to the current methods of electrophysiological assessment, ECGi can be fairly expensive for widespread use.

Imaging of epicardial potentials from the body surface requires the solution of a Laplace equation in the tissues located between the epicardium and skin, which is known as the inverse problem of electrocardiography.[73,74] Most of the models used simplify this problem by treating the body as a uniform isotropic volume conductor, like that used in the torso-tank experiment. Inverse electrocardiography has also been applied in synthetic myocardial ischaemia localization[75] and a macroscopic electrical heart model has been constructed using MRI-based biomechanical information to assess R- and T-wave conduction.[76] However, this differs from the actual in vivo environment, where the heart is surrounded by the lungs, fat, bone and muscle tissue, each of which has its own specific conductivity. These homogeneities in conductivity cause measurement inaccuracy, resulting in the incorrect localization of electrical activity, which has been calculated to average 10mm but can reach as high as 50mm in some cases.[77]

One of the major discrepancies among different types of ECGi lies in the solution of the inverse problem. Although we have briefly mentioned Tikhonov regularization, this is only one of many approaches suggested by several specialists. In recent work, Karoui et al.[78] compared 15 different algorithms to resolve the inverse problem using ex vivo and in silico data and found that each of them had benefits but varied according to left-ventricular, right ventricular, and bi-ventricular pacing.

Electrocardiographic imaging provides abundant and detailed information about the electrical potentials of the heart's epicardial surface. However, the endocardial and intramural potentials differ to various degrees from those of the epicardium.[2] For this reason, the electrical activity of structures deeper than the epicardium is far too weak and remains unrecognized by body surface ECGi.[4] Even if signals are recorded, their localization is uncertain because the intramural potentials have different conduction paths and do not correlate with those of the epicardium. Furthermore, direct mapping of the interventricular septum, which can often be a source of ectopic impulses, is also unavailable for ECGi.[36]

The atrial wall consists of only a thin layer of myocardium, which produces very weak electrical potentials. This can sometimes pose a problem for ECGi interpretation, particularly in cases of strong ventricular activity and large QRS complexes, which then overlap P waves and make them impossible to analyse.[4] In addition, a low signal-to-noise ratio may also be a limiting factor, because atrial potentials can easily be corrupted by artefacts and interference, making ECGi useless in the diagnosis of some ATs.[25]

In the search to construct better ECGi activation maps from unipolar ECGs for AT, it was found that the use of intrinsic deflection time as an AT marker is imprecise and leads to the appearance of false gradients. This occurs mainly because all myocardial unipolar activation signals are fragmented, or the acquisition process leads to a spatial or temporal low-pass filtering effect. To counter this limitation, the estimation of delays between neighbouring points may be a potential option to create more precise activation maps.[79]

An important drawback of ECGi, in most manufacturers, is the need to perform a CT scan to acquire a 3D image of the heart and the precise positioning of the electrodes. This exposes patients to radiation, with all of its negative effects, although the radiation dose is estimated to be minimal (approximately 148 mGy·cm).[6] In addition, the requirement to perform a CT scan as part of the procedure limits the use of ECGi to laboratories equipped with a CT scanner. Another limitation of ECGi is the false assumption that both the heart and torso are static elements. The geometry calculations are usually based on a fixed CT image that is captured during heart diastole and breath-hold.[80] However, during the recording of electrical potentials, these structures are inevitably in continuous motion, due to the cardiac cycle and breathing. This results in complex conduction relations that are constantly changing, thus providing room for error in inverse solution calculation.[73]

Considering that ECGi is a relatively new concept in cardiology, it is not surprising that only a few studies have examined its use in humans, as described above. Most of these studies were performed in just a few patients, and large-scale, multicentre, randomized studies will still be required to confirm the encouraging results.[39]