Diagnostic Conundrum

Kohli et al . tested the echocardiographic criteria of Chin et al ., Jenni et al ., and Stöllberger and Finsterer in patients with definitive LVNC. There was poor correlation among the three echocardiographic definitions, with only 30% of patients fulfilling all three criteria. Furthermore, Kohli et al . found that 8% of controls (four black subjects and one white subject) met at least one of the three criteria to diagnose LVNC. The review highlights the concern that the current echocardiographic criteria may be overly sensitive, particularly in black individuals, resulting in overdiagnosis of LVNC.

This concern is further reified by a Mayo Clinic series in which 25% of patients with echocardiographic diagnoses of LVNC did not meet diagnostic criteria after careful retrospective review by two experts from the same echocardiography laboratory. This raises a serious concern of overdiagnosis (false positives) of LVNC in the practices of echocardiographic laboratories, with a negative impact of diagnosis on patients’ lives. Additional support for the diagnostic conundrum is the pathologic definition of LVNC as described by Burke et al . and Freedom et al . Freedom et al . described the unique anatomic features of hearts explanted from patients with LVNC. The hearts revealed a luminal meshwork of thin and thick endocardial bands, tendons, filaments, and trabeculations that intermixed to form a discrete trabecular layer of the LV wall to its apex. The thick and coarse trabeculations of the trabecular layer were separated by deep recesses. They did not use ratios to identify LVNC. Burke et al . described 14 explanted hearts with LVNC (nine of 14 sudden cardiac deaths). The pathologic definition included poorly developed papillary muscles and a noncompacted inner LV myocardial layer that comprised > 50% of the LV thickness. An important finding to note is that in four of 14 cases (27%), the noncompacted/compacted myocardial ratio was < 2. The statistic that 27% of patients with pathoanatomic diagnoses of LVNC did not meet an arbitrary cutoff ratio of noncompacted to compacted myocardium > 2 reinforces the limitation of diagnosing LVNC solely on the basis of measurements of ratios.

We are at the precipice of a slippery slope of labeling people without a disease as having a serious disease entity. This is déjà vu, because a similar chasm was scaled with asymmetric septal hypertrophy (ASH) and hypertrophic cardiomyopathy (HCM). The recognition that HCM could not be defined by diagnostic criteria using only measurements and an arbitrary ratio resulted in the withering of ASH as pathognomonic of HCM.

Churchill’s quotation should resonate in our approach to obtain a deeper understanding of the LVNC spectrum. Lessons learned from a historical review of the measurements and ratios used to validate a diagnosis of HCM may prevent us from repeating the same errors. We present how the lesson of the experience with ASH and HCM could apply to LVNC.

Historical Lessons: Asymmetric Septal Hypertrophy

Four decades ago, in 1974, Henry and colleagues published their finding that ASH was a sensitive and specific disease marker of idiopathic hypertrophic subaortic stenosis. ASH was characterized by a ventricular septum that was ≥ 1.3 times as thick as the posterior LV free wall. They concluded that ASH was a pathognomonic anatomic abnormality of idiopathic hypertrophic subaortic stenosis. The observation that ASH was possible even with a normal thickness of the septum resulted in many normal individuals’ being incorrectly diagnosed with idiopathic hypertrophic subaortic stenosis (e.g., ventricular septum 12 mm, posterior wall 9 mm). With the introduction of two-dimensional echocardiography, the disease-identifying ratio went from ASH to disproportionate upper septal thickness. However, soon disproportionate upper septal thickness also was noted in systemic hypertension, pulmonary hypertension, valvular heart disease, and coarctation of the aorta.

The lack of a true criterion standard, such as a genetic marker, to identify LVNC and its heterogeneous phenotypic expressions has inhibited our ability to determine a valid, reproducible definition. Furthermore, our inability to define whether a hypertrabeculated phenotype M _LVNC/LVHT will transition to a cardiomyopathy complicates the issue. Thus, diagnosing the difference between the variant characterized by prominent apical trabeculations, M _LVNC/LVHT , and M _LVNC+D or M _LVNC+H poses a new challenge to the medical community. A diagnostic modality ensuring a valid diagnosis of LVNC as reliably as a genetic marker has remained elusive. Measurement of not only morphologic features but also of advanced echocardiographic parameters, such as strain, strain rate, displacement, rotation, torsion, and deformation, currently has not achieved this dream.

The main lesson learned from a temporal review of HCM that lends insight to the spectrum of LVNC is that disease should not be defined by measurements and ratios in isolation.

Historically, the entity LVNC was described and studied in patients with abnormal systolic function. In 1990, Chin et al . described their echocardiographic findings in eight patients with isolated LVNC. Five of the eight patients had depressed LV systolic function. Jenni et al . in 2001 reviewed 34 adult patients with LVNC. Seven hearts were available for anatomic review through heart transplantation (three patients) or autopsy (four patients). LV ejection fractions ranged from 29% to 38% in these seven patients. A recent study in France reviewed the clinical and echocardiographic features of 105 patients with LVNC. Of these patients, 84% had reduced LV ejection fractions. These three studies reveal that the ratio-based criteria to diagnose LVNC were derived from diseased myopathic ventricles. We must not repeat the historical mistakes made with ASH and rely on an arbitrary ratio to diagnose HCM. Analysis must be multivariate to prevent under- and overdiagnosis. The diagnostic tools must have clinical application that is sensitive and specific to allow a distinction between normal anatomic variants, M _LVNC/LVHT , and disease-state M _LVNC+D or M _LVNC+H .

Beyond Measurements and Ratios

The present diagnostic tools result in uncertainty, as the noncompacted/compacted myocardial ratio has limitations, and the extent of noncompacted LV areas cannot distinguish with certainty disease from a variation of normal anatomy. Available studies have identified conflicting results: that the degree of LVNC is not related to the severity of LV systolic dysfunction and that the degree of LVNC is related to the degree of LV dysfunction. One study observed myocardial fibrosis in noncompacted and compacted myocardial segments with similar prevalence, supporting the theory that noncompacted myocardium is a phenotypic marker of an underlying diffuse cardiomyopathy involving both noncompacted and compacted myocardium. These findings need further investigation and validation.

Where do we go from here? Recent research suggests that maximal systolic compacta thickness < 8 mm is specific for LVNC, allowing differentiation from normal hearts and myocardial thickening due to pressure overload, and may be used as an additional diagnostic criteria. Peters et al . found that LV twist is diminished in subjects with LVNC and normal rotation. Rigid body rotation occurred in 53% of patients (32 of 60) with LVNC. Captur et al . identified that fractal analysis of trabeculations may aid in distinguishing healthy volunteers from patients with LVNC and that fractal dimension of trabeculations is greater in healthy blacks than whites. Despite these incremental advances, we are still left without a definitive way to diagnose LVNC. Current diagnostic criteria alone appear inadequate to make or refute the diagnosis of LVNC definitively.