Prevalence and clinical relevance of the morphological substrate of ventricular arrhythmias in patients without known cardiac conditions detected by cardiovascular MR

Abstract

Objective: Cardiac MR (CMR) identifies the substrate of ventricular arrhythmia (VA) in cardiomyopathies and coronary heart disease. However, little is known about the value of CMR in patients with VA without previously known cardiac disorders.

Methods: 76 patients with VA (Lown ≥2) without known cardiac disease after regular diagnostic work-up were studied with CMR, and findings were correlated with electrocardiogram (ECG) and electrophysiological stimulation (EPS). Structural abnormalities matching the VA origin as defined by ECG and/or EPS, or a CMR-detected cardiac condition known to cause arrhythmia were defined as VA substrate. CMR findings were defined as clinically relevant, if resulting in a new diagnosis, change of treatment or additional diagnostic procedure.

Results: 44/76 patients demonstrated pathological CMR findings. In 24/76 patients, the pathology was detected by CMR and not by echocardiography. CMR-based diagnoses of cardiac disease were established in 20/76 patients, and all were morphological substrates for VA. In seven patients, the location of the CMR finding (scar) directly matched the VA origin. CMR findings resulted in a change of treatment in 21 patients and/or additional diagnostics in 8 patients.

Conclusion: Undetected cardiac conditions are frequent causes of VA. This is the first study demonstrating the value of CMR for detection of morphological substrate and/or underlying cardiac disorders in VA patients without known cardiac disease.

Advances in knowledge: The high incidence of clinically relevant CMR findings which were not detected during initial diagnostic work-up strongly supports the use of CMR to screen VA patients for underlying heart disease.

Figure 1.

An overview of the study. CD, cardiac disease; CHD, coronary heart disease; CMR, cardiac MR; VA, ventricular arrhythmia.

Figure 2.

Electrocardiogram (ECG) recording of ventricular tachycardia (VT). ECG recording on admission: VT with right bundle branch morphology of inferior/septal origin (negative axis in leads II, III + aVF; positive axis in leads I + aVL).

Figure 3.

Cardiac MR (CMR) of scattered post-myocarditis scar. There is scattered scarring throughout the entire left ventricle involving the subepicardial and the midmyocardial layers (arrows). Note, slow flow/gadolinium pooling is present along the right ventricle wall and may mimic pathological contrast enhancement (arrowhead). The subendocardium is spared, a typical finding consistent with post-myocarditis scarring. The most prominent scar was present in the midventricular inferoseptal/inferior segment. Although most likely attributable to viral myocarditis, this form of scarring may also be found in non-infectious myocarditis (i.e. cardiac sarcoidosis) and warrants further work-up if the clinical history is inconclusive.

Figure 4.

Recording of electrophysiological stimulation. After application of programmed extrastimuli, a pleomorphic-onset ventricular tachycardia followed by ventricular tachycardia originating from the inferoseptal axis was induced.

Figure 5.

Typical findings detected by cardiac MR (CMR). (a) A patient with recurrent sustained ventricular tachycardia (VT). During electrophysiological stimulation (EPS), two fast VTs with right bundle branch block (RBBB) morphology, inferior axis, were inducible. Matching perimyocarditis scar is predominantly located in the subepicardial myocardial wall, involving the subepicardial fat, extending to the pericardium (arrows). (b) A patient presenting with unclear tachycardia revealed in the electrocardiogram. During EPS, a VT with the same morphology was inducible in the left ventricle (RBBB, superior axis and marked left-axis deviation). A CMR-detected post-myocardial infarction scar in the inferior basal wall. The scar was subendocardial and in part transmural (arrows). Because of the small size of the scar, wall motion was not significantly impaired and relevant thinning of the wall had not occurred. Note, myocardial clefts are occasionally present in the same location and may mimic ischaemic scars. These clefts can easily be differentiated from scars by consulting the corresponding cine images, which demonstrate the corresponding muscular cleft.