Left ventricular (LV) remodeling is a serious process that may lead to fatal heart failure and can be caused by many cardiovascular diseases, such as ischemic cardiomyopathy, dilated cardiomyopathy, myocarditis, valvular disease, and others. Several pathophysiologic processes have been demonstrated to take part in this complex process. For instance, cardiomyocyte apoptosis is increased in remodeled ventricles in patients with ischemic and dilated cardiomyopathy, with as much as 2% of cardiomyocyte nuclei being apoptotic at the same time. Myocyte hypertrophy and increased interstitial collagen are also found.
The main features of a remodeled ventricle are a large volume, poor contractility (which leads to a decreased ejection fraction [EF]), and an altered geometry.
Larger volumes correlate with patient outcomes, and current therapeutic efforts aim to decrease LV volumes to prevent adverse events. Two-dimensional echocardiographic volume measurements have shown good correlation with the incidence of outcomes, but three-dimensional (3D) echocardiography provides more accurate analyses of LV volumes and EFs and correlates better with magnetic resonance imaging values, and its measurements have less intraobserver and interobserver variability. Moreover, it has been shown to identify parameters associated with the way LV function declines as the left ventricle dilates, such as ventricular mechanical dyssynchrony.
LV systolic dysfunction, quantified as a decreased EF, is also better assessed by 3D echocardiography rather than two-dimensional echocardiography. Other markers of LV systolic dysfunction, such as longitudinal and circumferential strain rate, have also shown associations with adverse outcomes.
As for LV geometry, remodeling leads to changes in the normal elliptical LV chamber configuration, which becomes more spherical. Circumferential LV strain rate deterioration correlates with LV remodeling, whereas longitudinal strain rate does not, suggesting that circumferential contractility is crucial in preserving LV shape.
Multiple attempts to obtain new parameters suitable for quantifying these shape modifications have been carried out. The LV sphericity index was one of the first developed parameters. It compares the volume of a sphere whose diameter is equivalent to the length of the LV long axis with the actual LV volume. Higher sphericity indexes correlate with higher 10-year mortality in patients after myocardial infarction. The sphericity index is associated with global LV shape, however, and does not take into account regional differences; these regional differences are especially important at the apex, because regional shape changes at that location usually precede global LV dilation. Di Donato et al. proposed a new “apical conicity index,” which calculates the ratio between “apical axis” (defined as the diameter of the hypothetical sphere that best fits the apical curvature) and short-axis length and positively correlates with mitral valve function after an anterior myocardial infarction: LV chambers that dilate globally after a myocardial infarction (i.e., those left ventricles having high sphericity indexes) tend to have more severe mitral valve regurgitation than LV chambers in which dilation is regional at the apex (having a high apical conicity index but without increased sphericity). This means that it is dilation of the basal (where the mitral annulus rests) and mean (where the chordae tendineae are mainly anchored) segments that play the leading role in functional ischemic mitral regurgitation.
Three-dimensional echocardiography has also arisen as a robust approach to obtain more precise and useful measures of LV geometry. Recently, several studies have tried to assess the process of LV remodeling in various populations using many different measurements, including a “remodeling index” (the ratio between LV mass and LV end-diastolic volume).
In this issue of JASE , Salgo et al. present a new and challenging 3D echocardiographic geometric approach for evaluating LV remodeling. They used customized new software to obtain 3D LV endocardial meshes at end-systole and end-diastole in 106 consecutive patients (47 with dilated cardiomyopathy and 59 healthy controls). The endocardial meshes were also segmented according to the American Heart Association’s 17-segment classification. A total of 840 equally distributed nodes were then allocated to cover the LV endocardial surface, and the software calculated seven different curvature values according to the relationship between each node and its surroundings. These values were averaged for all nodes within each segment, and also for all LV nodes, to obtain regional and global curvature values, respectively. Curvature values were computed as the reciprocal of the radii of the circles that tangentially fitted the curved regions around the nodes. No geometric assumptions were needed for the computational process.
A similar computational approach has been used with magnetic resonance imaging to assess LV remodeling, but this is the first time a comparable geometric approach has been developed for 3D echocardiography.
Salgo et al. show that in healthy controls, curvature values were higher than those in patients with dilated cardiomyopathy and that the most abnormal regions were the apex and the interventricular septum. End-systolic curvature values were more affected than end-diastolic ones, illustrating that “normal” LV contraction makes the cavity curved as systole progresses, whereas in patients with dilated cardiomyopathy, LV contraction cannot change the shape of the LV cavity appropriately, leading to a decreased ejection volume and, therefore, an impaired EF. In fact, apical and septal curvature values correlated well with LV EF.
This curvature analysis seems promising, but many points must be clarified. First, many curvature values were calculated by Salgo et al. , but no evidence regarding which ones are more useful is available. The investigators used a mean between the maximum and minimum curvature values at each node to conduct correlation analysis with LV EF, probably because this measure had the largest significant intergroup differences among the curvature measurements, but this decision was not based on multivariate analysis, which may help decide which curvature values are the most important and which are superfluous.
Mean global curvature values represent the reciprocal of the radius of a sphere that best defines the mean LV curvature. Using the data published, the end-diastolic radius would be 2.56 cm for healthy controls and 3.22 cm for patients with dilated cardiomyopathy. Hence, the end-diastolic volume of the sphere that represents healthy controls would be approximately 70 mL, while the end-diastolic volume for patients with dilated cardiomyopathy would be approximately 140 mL.
A limitation of this study may be the lack of a multivariate analysis, which would precisely describe whether curvature values are independently associated with LV EF or other LV functional parameters, once LV volume, sphericity index, and other well-known LV dysfunction parameters are included in the analysis. It may be that apart from being original and innovative, 3D curvature analysis would not significantly contribute to the prognostic value given by the aforementioned and easy-to-measure variables.
Another point to be clarified is that curvature analysis was carried out only in patients with dilated cardiomyopathy (probably as the paradigm of the LV remodeling process), and it should be validated in this and other populations. Moreover, the accuracy and reproducibility of this geometric analysis have not been tested yet. Comparisons with magnetic resonance imaging may be useful.
Several conditions and therapies may benefit from curvature analysis during the follow-up of patients. For instance, it may be used to optimize cardiac resynchronization therapy by helping decide the most appropriate location for the LV lead according to the degree of regional remodeling of each LV segment or for the follow-up of patients to quantify LV reverse remodeling. It also may have a prognostic value in ischemic cardiomyopathy; different regional shape changes may occur after a myocardial infarction, and 3D curvature analysis could be a good way to assess their prognostic value and to precisely quantify the remodeling. This technique might also be applicable to valvular heart disease, as the assessment of LV shape and its evolution with time could be useful for therapeutic decisions, perhaps being able to predict future LV dilation on the basis of previous shape changes and thereby useful in optimizing the timing of surgery.
Finally, once this technique is validated and the independent value of curvature measures over “classical” measures of LV size, shape, and function is confirmed, prospective studies need to be started to assess the clinical utility of this new curvature analysis. As has been demonstrated with LV volume, sphericity index and other variables, changes in curvature values should show prognostic and therapeutic value to justify including them in our daily routine.