Author + information
- Steffen E. Petersen, MD, DPhil,
- Joseph B. Selvanayagam, DPhil, FRACP,
- Frank Wiesmann, MD,
- Matthew D. Robson, PhD,
- Jane M. Francis, DCRR, DNM,
- Robert H. Anderson, MD, FRCPath,
- Hugh Watkins, MD, PhD, FRCP and
- Stefan Neubauer, MD, FRCP⁎ ()
- ↵⁎University of Oxford Centre for Clinical Magnetic Resonance Research, University Department of Cardiovascular Medicine, John Radcliffe Hospital, Oxford OX3 9DU, United Kingdom
We thank Drs. Stöllberger and Finsterer for their interest in our report (1). Many of the questions they raise reflect concern with how we selected our study population. Their suggestion of enrolling only symptomatic patients, however, with subsequent access to pathoanatomical specimens in order to establish diagnostic criteria would heavily bias findings to the severe end of the spectrum.
Concerns raised by Drs. Stöllberger and Finsterer do not recognize that the likelihood of any individual with a ratio of thicknesses between the trabecular and compact myocardium >2.3:1 suffering from pathological noncompaction, when expressed as the posterior probability, depends on the pretest probability or prevalence. A patient with a known family history of noncompaction and a 1 in 2 chance of having inherited the condition in autosomal dominant fashion would have a 97% chance of suffering from pathological noncompaction. In contrast, given the estimated prevalence of significant noncompaction of 1 in every 2,000 members of the normal population (2), the posterior probability would only be 26%. This Bayesian calculation (3) exemplifies the importance of taking into account the pretest probability, especially in diseases with relatively low prevalence.
We included one patient with left ventricular noncompaction on the basis of marked electrocardiographic changes, a family history of left ventricular noncompaction, and a distinct bilaminar appearance of the trabeculated and compacted layers of the myocardium. Based on imaging findings alone, this patient would not have been diagnosed with left ventricular noncompaction. Interestingly, this patient very recently developed symptomatic cardiac failure. This patient, now with clinical evidence of disease, but in the face of an initially nondiagnostic ratio between the trabecular and compacted layers, must now be assumed partially to express the disease, whereas a healthy volunteer with a low pretest probability would be interpreted as having nonpathological noncompaction.
We agree that the pathophysiological mechanisms leading to left ventricular noncompaction are not fully understood. Our data on the segmental distribution of trabeculated myocardium, nonetheless, support the hypothesis of an early arrest during the embryological process of compaction.
Our experience is that trabeculations can be more readily visualized in the relaxed heart. Magnetic resonance imaging (MRI) allows easy delineation of the epicardial border in both systole and diastole, in contrast to echocardiography, with poor epicardial delineation in diastole. State-of-the-art sequences, such as steady-state free precession, are not prone to flow artifacts, as once seen with the older sequences employing gradient echo techniques. The new sequences permit clear distinction between the trabeculations and the compacted myocardium. This makes this imaging technique ideal to distinguish noncompaction from papillary muscles, false tendons, and aberrant bands, because it has the accuracy to permit analysis of the entirety of the ventricular mass. We are surprised that Drs. Stöllberger and Finsterer raise this point as it is widely acknowledged that MRI, unlike echocardiography, does not suffer from this caveat.
Because of the constraints of space, we can only answer some of the further enquiries concerning the patients studied. An apical thrombus was detected by contrast-enhanced MRI in the patient with a history of systemic embolism. We enrolled all first-degree members of our families with left ventricular noncompaction who agreed to participate in the study and to undergo screening. None of our patients, including the seventh patient, had a family history of neuromuscular disorders. Our patients with left ventricular noncompaction and a positive family history showed an autosomal dominant pattern of transmission.
In conclusion, our study demonstrated the high diagnostic accuracy of cardiovascular MRI for left ventricular noncompaction. Owing to the importance of pretest probabilities, we disagree with the view that the diagnosis of left ventricular noncompaction should rely exclusively on morphological abnormalities. Thus, assessments using either echocardiography or cardiovascular MRI can at best contribute to the clinical diagnosis.
- American College of Cardiology Foundation