A disproportionate contribution of papillary muscles and trabeculations to total left ventricular mass makes choice of cardiovascular magnetic resonance analysis technique critical in Fabry disease

In this study, we demonstrate that the choice of including or excluding papillary muscles and trabeculations in LV mass measures can result in large differences in the quantification of LV function and mass that is exacerbated in a condition classically associated with cardiac hypertrophy. We show that in patients with Fabry disease, in which sphingolipids are deposited in myocardial tissue, the contribution of papillary and trabecular muscle to the total LV mass is markedly increased compared to healthy controls. The inclusion of papillary muscles in myocardial volume improves accuracy of volumetric measures of LV stroke volume when compared with the internal reference measure taken from velocity encoded flow data at the aortic outflow. These findings may have significant implications for assessment and targeted therapy of patients with Fabry disease, as well as other conditions associated with cardiac hypertrophy, and should prompt a move toward more consistent standards in reporting volumetric data.

Our data demonstrate significant differences between the two analysis approaches for both the Fabry and control groups across a range of parameters. While the impact of analysis approach was modest in the control subjects, the magnitude of these differences was much greater in the Fabry cohort (between 10-30% - see Figure 2). The comparison between the flow-derived LVSV provides strong evidence that M _inc P&T is a more accurate approach in the setting of Fabry-related hypertrophy (Figure 4), and even though for control subjects the comparison was non-significant, it is likely that using the M _ex P&T approach would introduce a bias toward higher LVEDV, LVESV and LVSV with increasing hypertrophy. In a population setting, this would have the effect of over reporting in these three parameters. We also know, from the Bland Altman plots, that M_incP&T is reproducible and reliable.

Recent data using a computer-aided contouring approach in a large sample (n = 1494) drawn from the Framingham study also supports the use of a M _inc P&T approach, demonstrating that the inclusion of P&T in the calculation of LVM results in a measureable difference in normal subjects [12]. The historic motivation for using M _ex P&T contouring was primarily pragmatic – motivated by the relative speed and reproducibility of this technique in normal subjects using manual contouring. Early computer-based contouring routines were not capable of accurate delineation of trabecular contours, hence the use of this technique has been limited by pragmatic considerations [13]. However, in the setting of hypertrophy, the M _ex P&T approach has difficulties in delineating the interface between tightly packed trabeculae and compacted myocardium without careful attention to detail and close reference to the moving cine images [6]. This is especially difficult towards the apex [12]. This technique is therefore not suited to automation in the presence of hypertrophy due to the increase in compacted trabeculations. Although we use a manual approach in this paper, the increasing availability of computer-aided analysis capable of accurate M _inc P&T contours means that the practical advantage of M _ex P&T is no longer compelling.

In the context of cardiac hypertrophy, the effect of type of volumetric analysis approach on the estimation of cardiac mass is of even greater importance than the functional measures. In the Fabry group, papillary muscle and trabecular mass contributed to an average 20% of the total cardiac mass, over one and a half times more than that of the control group (Figure 3). Thus, comparing to the normal range calculated using a papillary exclusion method (as recommended by different organisations) will not fully account for this, and will therefore underestimate the degree of hypertrophy and cardiac involvement for a patient with Fabry disease. The distortion of this important measure may have implications for patients’ disease severity classification and therapy management, which at our institutions is based on organ involvement and the Mainz Severity Scale Index (MSSI) scoring system [14]. Currently, the cardiovascular score in the MSSI includes an assessment of “cardiac muscle thickness” based on electrocardiogram and echocardiography features. In this study, the inclusion of P&T in LVM measurements increased the number of Fabry patients exceeding the upper limit of normal from 6/20 to 11/20 after normalization by BSA [15]. Therefore, an additional five Fabry patients were reclassified as having an abnormally high cardiac mass using the M _inc P&T approach. This effect is driven by the increased contribution papillary muscle hypertrophy makes to overall cardiac mass in Fabry disease, versus non-deposition forms of LV hypertrophy, and has important implications for disease severity classification and therapy in this disease.

Only males were analysed in this study to avoid any influence of gender on hypertrophy in general. We acknowledge that this is a limitation considering Fabry disease affects both genders. It is unclear whether the results can be translated to females, given the gender differences in Fabry cardiomyopathy regarding LVH [16].

The findings of our study have considerable importance in the setting of Fabry disease due to the availability of enzyme replacement therapy (ERT) and the targeting of this therapy to those with identified end-organ involvement. ERT has been available since 2001, and data over the last decade suggests that ERT has the potential to change the prognosis of Fabry disease [17-21]. Despite this, there remain numerous uncertainties regarding these agents – these relate to the optimal stage for treatment initiation, considerable heterogeneity of response, and the magnitude of the long-term clinical benefits. The current criteria for initiation of ERT are variable by location, but in Australia include left ventricular hypertrophy as measured by echocardiography or cardiac MRI, or significant arrhythmia. Making use of the M _inc P&T approach in this cohort may have altered the eligibility for ERT for five of this study’s patients. Our data strongly suggest that accurate measurement of LVM is best achieved using a method that accurately defines the endocardial border so as to include all significant trabeculations and papillary muscles. This analysis deliberately focused on examining subjects with Fabry disease versus controls; further studies are required to explore if the results apply to other types of cardiac hypertrophy.