Abstract
We used color kinesis, a recent echocardiographic technique that provides regional information on the magnitude and timing of endocardial wall motion, to quantitatively assess regional right ventricular (RV) systolic and diastolic properties in 76 subjects who were divided into five groups, as follows: normal (n = 20), heart failure (n = 15), pressure/volume overload (n = 14), pressure overload (n = 12), and RV hypertrophy (n = 15). Quantitative segmental analysis of color kinesis images was used to obtain regional fractional area change (RFAC), which was displayed in the form of stacked histograms to determine patterns of endocardial wall motion. Time curves of integrated RFAC were used to objectively identify asynchrony of diastolic endocardial motion. When compared with normal subjects, patients with pressure overload or heart failure exhibited significantly decreased endocardial motion along the RV free wall. In the presence of mixed pressure/volume overload, the markedly increased ventricular septal motion compensated for decreased RV free wall motion. Diastolic endocardial wall motion was delayed in 17 of 72 segments (24%) in patients with RV pressure overload, and in 31 of 90 segments (34%) in patients with RV hypertrophy. Asynchrony of diastolic endocardial wall motion was greater in the latter group than in normal subjects (16% versus 10%: p < 0.01). Segmental analysis of color kinesis images allows quantitative assessment of regional RV systolic and diastolic properties.
Although the importance of alterations in right ventricular (RV) function in various cardiopulmonary disease states has been recently recognized (1-3), assessment of RV function remains challenging because of the complex geometry of the right ventricle, its asynchronous contraction pattern, and its mechanical interaction with the left ventricle (4, 5). These factors limit the validity of simple geometric assumptions required for function analysis, and thus limit the use of noninvasive imaging techniques, such as computed tomography (6) and radionuclide angiography (7), for assessing RV function.
Color kinesis is a recently developed echocardiographic technique, based on integrated backscatter analysis (8), which tracks and color-encodes endocardial motion throughout the cardiac cycle in real time. Thus, color kinesis provides regional information about the magnitude and timing of cardiac wall motion, and has been recently used for the quantitative assessment of regional left ventricular (LV) function, both in systole and in diastole (9-11). However, the clinical value of this technique for the objective assessment of regional RV function has yet to be determined.
Accordingly, the aim of this study was to use color kinesis to quantitatively evaluate regional RV systolic and diastolic properties in normal subjects and in patients with various pathologic cardiopulmonary states known to be associated with RV dysfunction.
Ninety consecutive subjects were studied prospectively. Exclusion criteria were: pericardial effusion, arrhythmias, conduction abnormalities (e.g., atrioventricular or bundle-branch blocks), heart rate (HR) below 55 beats/min or above 110 beats/min, and catheters or pacemakers positioned in the RV. Fourteen of the original 90 subjects were excluded (normal subjects, n = 3; heart failure, n = 4; pressure/ volume overload, n = 3; pressure overload, n = 2; RV hypertrophy, n = 2) because of poor image quality that resulted in inadequate RV endocardial border tracking. The remaining 76 subjects (35 males and 41 females; age: 57 ± 17 yr [range: 24 to 90 yr]) were divided into the following five groups according to: (1) medical history; and (2) predefined Doppler echocardiographic criteria described in Table 1:
| n | Age (yr) | Male/Female | Heart Rate (bpm) | RV Systolic Function | LV Fractional Shortening (%) | TR Peak Velocity (m/s) | Color Doppler TR Flow Area (cm2 ) | RV Free Wall Thickness (mm) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| (1) Normal subjects | 20 | 43 ± 15 | 11/9 | 63 ± 9 | Normal | > 30 | < 2.5 | ⩽ 2 | < 5 | |||||||||
| (2) Biventricular dysfunction | 15 | 62 ± 13 | 10/5 | 72 ± 13 | Decreased | < 25 | > 3 | — | ⩽ 5 | |||||||||
| (3) RV pressure/volume overload | 14 | 56 ± 16 | 4/10 | 78 ± 16 | — | > 30 | > 3 | ⩾ 12/ASD | — | |||||||||
| (4) RV pressure overload | 12 | 59 ± 12 | 3/9 | 81 ± 11 | — | > 30 | > 3 | < 8 | — | |||||||||
| (5) RV hypertrophy | 15 | 71 ± 12 | 7/8 | 69 ± 12 | Normal | > 30 | ⩽ 3† | ⩽ 4 | ⩾ 7 |
1. Normal group (n = 20). This included nonsmoking subjects with no history of cardiopulmonary disease and normal Doppler echo- cardiographic studies.
2. Heart failure group (n = 15). This included patients with biventricular dysfunction with associated secondary pulmonary hypertension (seven patients with idiopathic dilated cardiomyopathy and eight patients with dilated cardiomyopathy resulting from coronary artery disease).
3. Pressure volume overload group (n = 14). This included 10 patients with RV pressure and/or volume overload associated with severe tricuspid regurgitation (color Doppler flow area ≠ 12 cm2) resulting from pulmonary hypertension, and normal LV function. Seven of these 10 patients had chronic obstructive lung disease, one patient had primary pulmonary hypertension, and two patients had mild to moderate pulmonary stenosis. In addition, this group included four patients with predominant volume overload secondary to an ostium secundum atrial septal defect.
4. Pressure overload group (n = 12). This included patients with normal LV systolic function and predominant RV pressure overload resulting from pulmonary hypertension. In contrast to the pressure/ volume overload group, all of these patients had a tricuspid regurgitation flow area of less than 8 cm2 on color Doppler imaging. These 12 patients had chronic obstructive lung disease (n = 8), mitral stenosis (n = 2), idiopathic pulmonary fibrosis (n = 1), and sarcoidosis (n = 1).
5. RV hypertrophy group (n = 15). This included patients with RV hypertrophy and no chronic pulmonary hypertension. With the exception of one patient with amyloid heart disease, all patients in this group had biventricular hypertrophy resulting from systemic hypertension with normal systolic function.
Transthoracic echocardiographic examination included M-mode, two-dimensional imaging, pulsed- and continuous-wave Doppler interrogation, and color flow mapping of valvular orifices with either a 2.5 or 3.5 MHz transducer (SONOS 2500; Hewlett-Packard Co., Andover, MA).
Two-dimensional targeted M-mode imaging of the RV anterior wall was done in the parasternal long-axis view to measure the end-diastolic RV anterior wall thickness with the conventional leading-edge technique. The end-diastolic diameters of the mid-RV and mid-LV cavities were measured in the apical four-chamber view using electronic calipers (Figure 1A). Additionally, in patients with RV pressure and pressure/volume overload, LV minor-axis diameters, perpendicular (D1) and parallel (D2) to the septum, were measured in the parasternal short-axis view at the level of the midpapillary muscles at end-diastole and end-systole (Figures 1B and 1C). The D2/D1 ratio was used as an index of eccentricity (13). End-diastolic measurements were obtained at the R wave of the electrocardiogram, and the end-systolic frame was defined as that containing the smallest LV area in the short-axis view.
Fig. 1. Ventricular measurements made on two-dimensional echocardiographic images. In each patient, end-diastolic right and left ventricular diameters were measured in the apical four-chamber view (A), and the ratio between these values was calculated. In patients with mixed pressure/volume overload and in subjects with isolated pressure overload, left ventricular minor-axis diameters, perpendicular (D1) and parallel (D2) to the septum, were measured in the parasternal short-axis view at the level of the midpapillary muscles at end-diastole (B) and end-systole (C ). The index of eccentricity was obtained by calculating the D2/D1 ratio.
[More] [Minimize]The severity of tricuspid regurgitation was graded by performing digital planimetry of the area of the color-flow regurgitant jet, using a modification of the scale proposed by Miyatake and colleagues (14). An area of more than 12 cm2 was considered to represent severe regurgitation, and an area of less than 8 cm2 was considered to represent mild to moderate tricuspid regurgitation.
Technical background. Color kinesis classifies each pixel of an echo cardiographic image as either blood or tissue on a frame-by-frame basis. Pixel transitions from one frame to the next, defined as a change in the pixel designation from blood to tissue in systole or from tissue to blood in diastole, are color encoded by placing a distinct color overlay on the corresponding pixels (Figure 2). This frame-by-frame color encoding of pixel transitions is performed throughout either systole or diastole, with each color overlay representing the endocardial motion that occurred during a constant time interval of 33 ms. Paradoxical pixel transitions, defined as a change in the pixel designation from tissue to blood during systole or from blood to tissue during diastole, are labeled with a unique red overlay that may be toggled on and off, irrespective of the time of occurrence of these transitions (Figure 2). All color tags are removed at the end of image acquisition, and the process is repeated during the next cardiac cycle. Thus, a single end-systolic or end-diastolic color kinesis image displays a cumulative record of endocardial motion for the entire preceding ejection or filling period, respectively.
Fig. 2. Schematic representation of the principles of operation of color kinesis in systole and diastole. Each pixel is classified as either blood or myocardial tissue, and pixel transitions from blood to tissue in systole, or from tissue to blood in diastole, are detected on a frame-by-frame basis and color encoded, using a specific hue for each consecutive frame. Two different color scales are used for systole (11 colors) and for diastole (19 colors), wherein each color represents a 33 ms period. Paradoxical pixel transitions (“dyskinesis”) may be color encoded in red, irrespective of the time of their occurrence. In the present study, this mode was not activated.
[More] [Minimize]Color kinesis data acquisition. Systolic and diastolic color kinesis images of the RV were obtained from the apical four-chamber view during passive end-expiration, with the patient in the left lateral decubitus position. The acoustic quantification system was initially activated, and gain controls were adjusted to optimize tracking of the RV endocardial border (15) within a manually traced region of interest. Color kinesis was then activated to successively color encode systolic and diastolic RV endocardial motion, with the dyskinetic-motion mode turned off. Systolic color encoding was triggered by the R-wave of the electrocardiogram, and was terminated at the end-systolic frame (10). In diastole, color encoding started with the end-systolic frame and terminated with the following R-wave. At times, manual adjustments were necessary to ensure that the initiation of diastolic color-encoding coincided with the first frame depicting outward endocardial motion. The duration of the diastolic color encoding was always set to its maximal value (627 ms) to cover the entire RV filling period. Two nonconsecutive end-systolic and end-diastolic color kinesis images were acquired and stored on an optical disk for off-line analysis.
Analysis of color kinesis images. Through the use of custom software, color-encoded images of the RV were automatically divided into six segments (Figure 3). The segmentation originated from the automatically determined end-systolic centroid of the RV cavity. Two wedge-shaped sectors were delineated to divide the RV into the septum and the free wall. This scheme was based on three manually determined anatomic landmarks, including the RV apex and the junction of both tricuspid valve leaflets with the tricuspid annulus (Figure 3). This segmentation scheme purposely excluded tricuspid valve motion from the analysis. Each sector was then divided into areas subtended by three equal angles. In each of the six resulting RV segments, pixels of each color and pixels marked as blood were counted. As previously described for the LV (10), pixel counts were used to quantify the magnitude and timing of regional RV endocardial motion. Regional endocardial motion was expressed in terms of the fractional area change as a percent of segmental end-diastolic area. The regional fractional area change (RFAC) in both systole and diastole was displayed in the form of stacked color histograms, wherein each time frame (33 ms) was represented by a specific color identical to that used in the color kinesis images.
Fig. 3. Schematic representation of the RV segmentation used in the apical four-chamber view. The six segments originate from the automatically calculated end-systolic RV cavity centroid, and are determined from three easily identifiable anatomic landmarks: the tip of the RV endocardium and both junctions between the tricuspid valves and the tricuspid annulus. (b-sp = basal septum; m-sp = midseptum; a-sp = apical septum; a-fw = apical free wall; m-fw = mid-free wall; b-fw = basal free wall.)
[More] [Minimize]To evaluate regional temporal filling patterns, fractional area change was integrated with respect to time and was normalized to 100% in each individual RV segment. This eliminated intersegmental differences in the magnitude of endocardial motion. To abolish the confounding effects of intersubject variability in HR, linear interpolation was used to obtain 20 values of segmental fractional area change in 5% increments of RV filling time. Thus, each curve reached 100% of endocardial motion at 100% of filling time, thereby allowing intersubject comparisons of the percentage of total endocardial displacement completed at any specific percent of the RV filling period. In each segment, the proportion of total diastolic endocardial motion completed after 50% of RV filling time was computed. The corresponding values of RFAC were then averaged for all segments, and standard deviation (SD) was used as an index of RV diastolic wall-motion asynchrony in each subject. In normal subjects, regional time curves were averaged to obtain reference profiles of diastolic endocardial motion in each RV segment, which were defined as consisting of values falling within 1 SD around the mean. To objectively identify abnormalities in regional diastolic endocardial wall motion, we superimposed individual regional time–curves obtained in patients with RV pressure overload and RV hypertrophy on the normal reference curves. Instances of delayed regional diastolic endocardial motion, defined as a downward shift of regional time curves relative to the corresponding reference at 50% of RV filling time, were counted in each group of patients.
Mean RV ejection and filling times, which respectively represent the average times required for a pixel to change from blood to tissue in systole or from tissue to blood in diastole (10), were also computed for each segment, and were displayed as bar diagrams and compared among the study groups.
Data obtained from two nonconsecutive end-systolic and end-diastolic color kinesis images were averaged for each subject. Individual data were averaged for each group and compared with those for the normal subjects. The reproducibility of the segmental analysis was evaluated in a subgroup of 16 randomly selected subjects by acquiring and analyzing three nonconsecutive end-systolic and end-diastolic color kinesis images.
Results were expressed as mean ± SD. Intergroup comparisons were made with the Mann–Whitney rank sum test. Differences were considered significant for values of p < 0.05. Reproducibility was quantified by averaging the histograms of the three repeated measurements and calculating for each segment the SD divided by the mean of the RFAC.
In normal subjects, the RV free wall consistently exhibited greater systolic displacement than the septal region, as illustrated by a marked difference in the thickness of the color bands in the end-systolic color kinesis images (Figure 4A). These findings were confirmed by the stacked histograms, which depicted a significantly greater RFAC in the free wall than in the septal segments (57 ± 11% versus 23 ± 8%, p < 0.0001) (Figure 4A).
Fig. 4. End-systolic (A) and end-diastolic (B) apical four-chamber view color kinesis images of the RV obtained from a normal subject with segmentation scheme superimposed. The thickness of the color band reflects the magnitude of inward (systole) or outward (diastole) endocardial motion. Tricuspid valve motion is not analyzed (shaded area). In this normal subject, RV free wall motion is greater than ventricular septal motion, which behaves as part of the left ventricle, as objectively confirmed by color kinesis stacked histograms averaged for the normal group (bottom). The colors in the histograms are identical to those used in color kinesis images. Each color band reflects endocardial motion during a 33-ms period. Dashed areas represent one standard deviation. (RFAC = regional fractional area change; REDA = regional end-diastolic area; other abbreviations as in Figure 3.)
[More] [Minimize]In diastole, the pattern of regional RV endocardial motion was similar to that observed in systole in all normal subjects (Figure 4B). Regional diastolic RV endocardial motion was relatively uniform as reflected by the narrow index of diastolic wall motion asynchrony (10%). However, septal endocardial motion usually occurred earlier in diastole than in the RV free wall. This was evidenced by the steeper slopes of the time curves and shorter mean RV filling time in the septal segments (119 ± 25 versus 208 ± 39 ms, p < 0.0001).
Table 2 summarizes the systolic RFACs of the RV septal and free wall regions, as well as the mean ejection times and ventricular diameter ratios in the different study groups.
| n | RFAC Interventricular Septum (% REDA) | RFAC Free Wall (% REDA) | Mean Ejection Time (ms) | RV/LV Width | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| (1) Normal subjects | 20 | 23 ± 8 | 57 ± 11 | 131 ± 21 | 0.66 ± 0.14 | |||||
| (2) Biventricular dysfunction | 15 | 19 ± 9 | 29 ± 7* | 112 ± 18* | 0.65 ± 0.19 | |||||
| (3) RV pressure/volume overload | 14 | 43 ± 11* | 31 ± 9* | 123 ± 18 | 0.95 ± 0.11* | |||||
| (4) RV pressure overload | 12 | 22 ± 8 | 40 ± 14* | 109 ± 15* | 0.88 ± 0.13* | |||||
| (5) RV hypertrophy | 15 | 23 ± 9 | 54 ± 10 | 130 ± 17 | 0.66 ± 0.07 |
As compared with normal subjects, patients with biventricular systolic dysfunction exhibited markedly decreased endocardial motion along the RV free wall (Figure 5A), whereas the magnitude of inward septal motion in these patients was similar to that observed in the normal group (Figure 4A). Mean ejection time was significantly shorter than that measured in normal subjects (Table 2). In patients with RV pressure/volume overload, the endocardial motion profile was reversed from that of normals, with increased systolic septal inward motion and significantly reduced endocardial excursion of the RV free wall (Figure 5B), whereas mean ejection time was similar to that of normal subjects (Table 2). Patients with relatively pure RV pressure overload had decreased motion of the RV free wall, whereas the magnitude of septal excursion in this group was similar to that of normal subjects (Figure 5C). In this group of patients, mean ejection time was significantly shorter than in normals (Table 2). In patients with RV hypertrophy, the pattern of regional systolic endocardial motion and the mean ejection time were similar to those obtained in the normal group (Table 2).
Fig. 5. Examples of end-systolic color kinesis images of the RV in the apical four-chamber view, obtained from three different study subjects (top), and corresponding stacked histograms averaged in each corresponding group of subjects (bottom). Note the flat pattern of endocardial wall motion in patients with biventricular failure, which reflects decreased RV free wall displacement (A), whereas the pattern of endocardial wall motion was similar to that of normals (Figure 4) in patients with RV pressure overload (C ). In contrast, patients with RV pressure/volume overload (B) exhibited a reversed endocardial motion profile owing to significant ventricular septal recruitment toward RV contraction. (RFAC = regional fractional area change; REDA = regional end-diastolic area; b-sp = basal septum; m-sp = midseptum; a-sp = apical septum; a-fw = apical free wall; m-fw = mid-free wall; b-fw = basal free wall.)
[More] [Minimize]Although patients with biventricular dysfunction had markedly enlarged RV cavities as compared with normal subjects (3.4 ± 0.8 versus 2.7 ± 0.5 cm, p = 0.002), the ratio between ventricle diameters in these patients was not increased because of the presence of associated LV dilatation (Table 2). Patients with isolated RV pressure overload and pressure/volume overload had an increased RV cavity size (4.0 ± 1.0 and 4.3 ± 1.0 cm, respectively; both p ⩽ 0.001) and increased RV/ LV diameter ratios (Table 2). Despite having a smaller mean RV cavity size than normals (2.3 ± 0.6 versus 2.7 ± 0.5 cm, p = 0.02), patients with RV hypertrophy had normal ventricular diameter ratios (Table 2), owing to the presence of associated concentric LV hypertrophy. Patients with relatively pure RV pressure overload exhibited similar eccentricity indices at end-systole and at end-diastole (1.12 ± 0.13 versus 1.14 ± 0.11, p = NS). In contrast, patients with pressure/volume overload had a greater end-diastolic than end-systolic eccentricity index (1.43 ± 0.18 versus 1.16 ± 0.09, p < 0.0001).
Table 3 summarizes the diastolic RFAC of the RV septal and free wall regions, as well as the mean filling times and indices of asynchrony, for the different study groups. Within each group, the diastolic pattern of RV endocardial motion closely matched the profile of systolic wall motion. However, differences in the timing of diastolic endocardial motion were noted in patients with RV pressure overload and RV hypertrophy.
| n | RFAC Interventricular Septum (% REDA) | RFAC Free Wall (% REDA) | Mean Filling Time (ms) | Index of Asynchrony (%) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| (1) Normal subjects | 20 | 31 ± 15 | 61 ± 11 | 164 ± 25 | 10 | |||||
| (2) Biventricular dysfunction | 15 | 23 ± 9 | 29 ± 7* | 164 ± 29 | 12 | |||||
| (3) RV pressure/volume overload | 14 | 48 ± 15* | 40 ± 15* | 158 ± 35 | 9 | |||||
| (4) RV pressure overload | 12 | 31 ± 12 | 52 ± 21 | 174 ± 30 | 15 | |||||
| (5) RV hypertrophy | 15 | 34 ± 9 | 68 ± 10 | 208 ± 41* | 16* |
Figure 6 shows examples of end-diastolic RV color kinesis images obtained from three different study subjects. When compared with the images from a normal subject, those of both the patient with RV pressure overload and the patient with RV hypertrophy had a greater proportion of late diastolic colors (yellow-orange hues). This finding is consistent with an augmented contribution of right atrial contraction toward RV filling. In both the RV pressure overload and RV hypertrophy groups, mean filling time was prolonged, although statistical significance was only reached in patients with RV hypertrophy (Table 3). HRs in patients with RV pressure overload were higher than in normal subjects (81 ± 11 versus 63 ± 9 beats/ min, p < 0.0001), whereas in patients with RV hypertrophy (69 ± 12 bpm), they were not significantly different from those of normals.
Fig. 6. Examples of end-diastolic color kinesis images of the RV in the apical four-chamber view, acquired from three different study subjects. When compared with the normal subject (left), patients with RV pressure overload (center) and RV hypertrophy (right) exhibited predominantly yellow-orange late diastolic colors, which reflected a markedly increased contribution of right atrial contraction toward RV filling.
[More] [Minimize]When compared with those of normal subjects, the initial slopes of averaged regional time curves were reduced in patients with RV pressure overload and RV hypertrophy (Figure 7, top panel). In addition, the index of asynchrony in regional RV filling was augmented in these two groups, although statistical significance was reached only in patients with RV hypertrophy (Table 3). As in normal subjects, regional mean time of filling was shorter in septal segments than in the free wall region in both groups (Figure 7, bottom panel).
Fig. 7. Assessment of regional diastolic RV endocardial wall motion asynchrony using color kinesis. In normal subjects (top left), regional RV filling curves were relatively uniform. However, septal endocardial motion occurred earlier in diastole than in the RV free wall, as evidenced by the steeper initial slopes of the time curves in the septal region, and the shorter mean RV filling time, <t\>, in the septal segments (bottom left). In patients with RV pressure overload (middle) and RV hypertrophy (right panels), endocardial wall motion occurred later in diastole, as reflected by the regional filling curves (top) and prolonged regional mean filling time (bottom). (RFAC = regional fractional area change; REDA = regional end-diastolic area; b-sp = basal septum; m-sp = midseptum; a-sp = apical septum; a-fw = apical free wall; m-fw = mid-free wall; b-fw = basal free wall.)
[More] [Minimize]Figure 8 shows an example of delayed regional diastolic endocardial motion objectively identified in one patient from each of the RV pressure overload and RV hypertrophy groups by the superimposition of individual time curves on the corresponding reference profiles. Using this method, we identified 17 of 72 segments (24%) in patients with RV pressure overload and 31 of 90 segments (34%) in patients with RV hypertrophy as segments with delayed diastolic endocardial wall motion. These regional filling abnormalities were equally observed in the septum and in the RV free wall (26 and 22, respectively, p = NS).
Fig. 8. Examples of regional filling curves obtained from a patient with RV hypertrophy and a patient with RV pressure overload. For each RV segment, a reference profile of diastolic endocardial motion was generated by averaging curves obtained in the normal subjects; the dashed lines represent a range of 1 SD around the mean. In these patients, delayed endocardial wall motion during early diastole was evidenced in several RV segments as individual time curves were shifted downward relative to the corresponding reference profile. (RFAC = regional fractional area change.)
[More] [Minimize]Repeated segmental analyses of nonconsecutive color kinesis RV images were found to provide values of RFAC that were reproducible within a mean of 13% in systole and 10% in diastole.
The present study demonstrated the ability of segmental analysis of color kinesis images to: (1) quantitatively reveal RV free wall systolic dysfunction; (2) objectively determine the predominance of pressure versus volume overloading of the RV; and (3) characterize asynchrony in regional diastolic RV endocardial wall motion. This methodology therefore promises to help the clinician in guiding and evaluating the effects of various therapeutic strategies on regional RV systolic and diastolic dysfunction in patients with various cardiopulmonary diseases.
The segmentation scheme applied to RV images in the apical four-chamber view (Figure 3) was designed to match that proposed by the American Society of Echocardiography (16). This segmentation was applied to color kinesis images to generate quantitative indices of regional endocardial motion (10). In each RV segment, normalization of colored pixel counts by the corresponding segmental end-diastolic area eliminated regional differences in geometry and intersubject differences in RV size. This normalization enabled intersegment as well as intersubject comparisons of RFAC, a parameter reflecting the magnitude of endocardial wall motion, which was used as an index of regional myocardial function in both systole and diastole. To assess the temporal patterns of diastolic endocardial motion on a segmental basis, we displayed RFAC in the form of time curves. The percentage of total endocardial motion completed at half of the RV filling period was computed for each segment, and was used to objectively assess asynchrony in diastolic wall motion (Figure 7). Delayed diastolic endocardial motion was identified by comparing individual curves obtained from patients with normal profiles (Figure 8). The reproducibility of the quantitative segmental analysis of systolic and diastolic color kinesis images of the RV was similar to that previously reported for the left ventricle (9, 11).
Using color kinesis in normal subjects, we found that the magnitude of systolic RV endocardial motion in the free wall was more than twice that measured in the septum (Figure 4A). This finding is consistent with the physiologic interaction between the two ventricles, since in the normal heart, the thin RV free wall is pulled toward the septum by the contracting left ventricle, owing to the presence of intertwining muscle bundles (17).
In patients with biventricular failure, the systolic endocardial motion of the RV free wall was markedly reduced as compared with that of normal subjects, as objectively reflected by the uniformly low RFAC displayed in the respective color kinesis histograms (Figure 5A). Because RV systolic function is load-dependent, the associated RV pressure overload caused by secondary pulmonary hypertension might have contributed to the overall reduction in the free wall motion noted in these patients (18). In patients with preserved LV systolic function and pulmonary hypertension resulting from bronchopulmonary diseases, systolic RV free wall motion was only minimally reduced (Figure 5C), presumably due to the presence of associated RV hypertrophy (77 ± 2.4 mm).
When compared with all other groups, patients with RV pressure/volume overload exhibited markedly increased systolic motion of the interventricular septum toward the RV cavity (Figure 5B). This finding was previously made in animal and human studies showing that in the setting of RV volume overload, the interventricular septum behaves as part of the RV instead of the LV (19-23). This anteriorly directed systolic endocardial motion of the interventricular septum has been termed “paradoxical” by Popp and colleagues (24). In the present study, 10 of 14 patients had systolic pulmonary hypertension in addition to RV volume overload. Since patients with RV pressure/volume overload and those with relatively pure RV pressure overload had preserved LV systolic function, with similar degrees of systolic pulmonary hypertension (55 ± 12 versus 65 ± 10 mm Hg, p = NS), our data suggest that the volume overload was predominantly responsible for the septal recruitment toward RV contraction. This observation was confirmed in four patients with pure volume overload, in whom the pattern of RV systolic endocardial motion was comparable to that of the entire group (Figure 9).
Fig. 9. Average color kinesis stacked histograms obtained from a subgroup of patients with mixed RV pressure/volume overload (left) and a subgroup of patients with pure volume overload (right), both in systole (top) and in diastole (bottom). The endocardial wall motion pattern in the two subgroups, was similar, and reflected a substantial septal recruitment, which compensated for decreased free wall motion. (RFAC = regional fractional area change; REDA = regional end-diastolic area; b-sp = basal septum; m-sp = midseptum; a-sp = apical septum; a-fw = apical free wall; m-fw = mid-free wall; b-fw = basal free wall.)
[More] [Minimize]It has been previously shown that both the direction and magnitude of systolic septal motion are strongly affected by the position of the septum at end-diastole, which in turn depends on the interventricular pressure gradient (21, 25). The proposed mechanism for this was that the heart contracts symmetrically toward its center of mass (25) and tends to restore its normal end-systolic geometry (13). In agreement with Pearlman and colleagues (25), we found that patients who exhibited septal recruitment toward RV contraction had the largest RV/LV diameter ratio, which approximated unity (0.95 ± 0.11). In addition, patients with RV pressure/volume overload exhibited the largest eccentricity index at end-diastole, whereas in patients with relatively pure pressure overload, this parameter was similar to that obtained at end-systole. Accordingly, the latter group of patients exhibited normal septal motion, as is consistent with minor septal distorsion prior to ventricular contraction (13). These findings suggest that septal recruitment occurs when the RV cavity approximates the diameter of the LV chamber (20), and when the timing of maximal ventricular septal flattening occurs at end-diastole (26). Restoring forces, which return the interventricular septum to a nearly normal curvature by end-systole, are thus exerted counter to the normal systolic inward motion of the septum toward the LV cavity. In keeping with this hypothesis and with previous reports (27), was our finding in the present study that the anterior septal displacement occurred predominantly within the first 100 ms of systole, as reflected by the greatest proportion of septal endocardial displacement occurring during the first three systolic frames on the corresponding stacked histograms (Figure 9).
The pattern of systolic endocardial motion in patients with RV hypertrophy (free wall thickness: 9.3 ± 2.1 mm) was similar to that observed in the control group, indicating that regional systolic RV function was preserved in this group. In these patients, RV hypertrophy was not the result of chronic pulmonary hypertension, but was rather associated with concentric LV hypertrophy as a result of systemic hypertension (28, 29).
Few studies have previously attempted to noninvasively assess regional RV diastolic function (6). In normal subjects, regional filling was consistently completed earlier in the septal region than in the RV free wall segments, as reflected by steeper time curves in the segments of the former region. Similarly, mean RV filling time was shorter in the septal region (Figure 7, left). This heterogeneity in regional diastolic RV endocardial motion observed in normal subjects might be explained by the anatomic and functional division of the RV into the sinus (inflow) and conus (outflow) regions (5, 18).
We uniformly noted, in both the RV pressure overload and RV hypertrophy groups, a marked delay in regional diastolic endocardial motion, particularly in the free wall segments. In addition, the RV hypertrophy group exhibited greater RV filling heterogeneity, as reflected by an increased asynchrony index and regional differences in the mean filling time. Using the methodology herein described, we were able to demonstrate in these groups that RV filling was delayed in 48 to 162 segments (30%). These findings show that RV hypertrophy, irrespective of its etiology, results in asynchrony in regional RV diastolic wall motion. Since regional LV diastolic dysfunction is known to negatively affect global LV filling (11, 30), it is likely that a similar mechanism is operant in the hypertrophied RV.
Color kinesis has several limitations. Because this technique does not allow correction for translation or rotation, color-encoded images should be interpreted with caution when significant translation or rotation is noted (e.g., in cases of large pericardial effusions). The accuracy of color kinesis strongly depends on the presence of adequate image quality. However, the exclusion of 16% of our patients because of inadequate RV endocardial boundary detection failed to cause an exclusion bias in our results.
In the present study we used a single echocardiographic view of the RV, which is not sufficient to permit evaluation of the function of all RV segments. Although we found the apical four-chamber view to provide the best suited imaging plane for simple segmentation and reproducible color kinesis analysis, other echocardiographic views have also been successfully used (31). The heavily trabeculated RV inflow and the moderator band, which is echocardiographically identifiable in most individuals, may be tracked and color-encoded by color kinesis imaging. In certain patients, this may lead to overestimation of the motion of apical segments. However, precise adjustment of regional gains generally allows the operator to optimize endocardial tracking in this region of the RV.
Because of the absence of a widely accepted reference imaging modality for the objective assessment of regional RV function, we did not compare our color kinesis findings with data obtained from an alternative approach. Nonetheless, acoustic quantification has been previously validated for the quantitative evaluation of global RV systolic function in both experimental (32) and clinical settings (33, 34). Moreover, color kinesis has been recently validated against conventional right ventriculography for the measurement of RV endocardial wall motion (31).
In summary, segmental analysis of color kinesis images provides on-line quantitative evaluation of regional RV systolic and diastolic properties. In the present study, the proposed methodology was able to clearly identify distinct types of RV endocardial motion profiles (i.e., increased afterload, decreased contractility, volume overload), and allowed quantification of the septal recruitment toward RV contraction in patients with RV volume overload. Additionally, our methodology provided objective detection of asynchrony in diastolic RV wall motion. With these capabilities, color kinesis promises to enable easy, noninvasive assessment of regional RV functional properties at patients' bedsides in various clinical settings. The clinical value of this new approach for the diagnosis, prognosis, and guidance of therapeutic management of RV dysfunction remains to be determined.
The authors gratefully acknowledge their cardiac sonographers and nurses for their precious cooperation and continuous help during this study.
| 1. | Yu C. M., Sanderson J. E., Chan S., Yeung L., Hung Y. T., Woo K. S.Right ventricular diastolic dysfunction in heart failure. Circulation93199615091514 |
| 2. | Klein A. L., Hatle L. K., Burstow D. J., Taliercio C. P., Seward J. B., Kyle R. A., Bailey K. R., Gertz M. A., Tajik A. J.Comprehensive Doppler assessment of right ventricular diastolic function in cardiac amyloidosis. J. Am. Coll. Cardiol.15199099108 |
| 3. | Polak J. F., Holman B. L., Wynne J., Colucci W. S.Right ventricular ejection fraction: an indicator of increased mortality in patients with congestive heart failure associated with coronary artery disease. J. Am. Coll. Cardiol.21983217224 |
| 4. | Lee F. A.Hemodynamics of the right ventricle in normal and disease states. Clin. Cardiol.1019925967 |
| 5. | Chin K. W., Daughters G. T., Alderman E. L., Miller D. C.Asynergy of right ventricular wall motion in man. J. Thorac. Cardiovasc. Surg.971989104109 |
| 6. | Marzullo P., L'Abbate A., Marcus M. L.Patterns of global and regional systolic and diastolic function in the normal right ventricle assessed by ultrafast computed tomography. J. Am. Coll. Cardiol.17199113181325 |
| 7. | Ratner S. J., Huang P. J., Friedman M. I., Pierson R. N.Assessment of right ventricular anatomy and function by quantitative radionuclide ventriculography. J. Am. Coll. Cardiol.131989354359 |
| 8. | Perez J. E., Waggoner A. D., Barzilai B., Melton H. E., Miller J. G., Sobel B. E.On-line assessment of ventricular function by automatic boundary dectection and ultrasonic backscatter imaging. J. Am. Coll. Cardiol.191992313320 |
| 9. | Lang R. M., Vignon P., Weinert L., Bednarz J., Korcarz C., Sandelski J., Koch R., Prater D., Mor-Avi V.Echocardiographic quantification of regional left ventricular wall motion with color kinesis. Circulation93199618771885 |
| 10. | Mor-Avi V., Vignon P., Koch R., Weinert L., Garcia M. J., Spencer K. T., Lang R. M.Segmental analysis of color kinesis images: new method for quantification of the magnitude and timing of endocardial motion during left ventricular systole and diastole. Circulation95199720822097 |
| 11. | Vignon P., Mor-Avi V., Weinert L., Koch R., Spencer K., Lang R. M.Quantitative evaluation of global and regional left ventricular diastolic function with color kinesis. Circulation97199810531061 |
| 12. | Abergel E., Chatellier G., Toussaint P., Dib J. C., Menard J., Diebold B.Doppler-derived pulmonary arterial systolic pressure in patients with known systemic arterial pressures. Am. J. Cardiol.771996767769 |
| 13. | Ryan T., Petrovic O., Dillon J. C., Feigenbaum H., Conley M. J., Armstrong W. F.An echocardiographic index for separation of right ventricular volume and pressure overload. J. Am. Coll. Cardiol.51985918924 |
| 14. | Miyatake K., Okamoto M., Kinoshita N., Ohta M., Kozuka T., Sakakibara H., Nimura Y.Evaluation of tricuspid regurgitation by pulsed Doppler and two-dimensional echocardiography. Circulation661982777784 |
| 15. | Bednarz J. E., Marcus R. H., Lang R. M.Technical guidelines for performing automated border detection studies. J. Am. Soc. Echocardiogr.81995293305 |
| 16. | Henry, W. L., A. DeMaria, H. Feigenbaum, R. Kerber, J. Kisslo, A. E. Weyman, N. Nanda, R. L. Popp, D. Sahn, N. B. Schiller, and A. J. Tajik. 1982. Report of the American Society of Echocardiography Committee on Nomenclature and Standards: identification of myocardial wall segments. American Society of Echocardiography, Raleigh, NC. 3–4. |
| 17. | Clyne C. A., Alpert J. S., Benotti J. R.Interdependence of the left and right ventricles in health and disease. Am. Heart J.117198913661373 |
| 18. | Oakley C.Importance of right ventricular function in congestive heart failure. Am. J. Cardiol.62199814A19A |
| 19. | Diamond M. A., Dillon J. C., Haine C. L., Chang S., Feigenbaum H.Echocardiographic features of atrial septal defect. Circulation431971129135 |
| 20. | Laurenceau J. L., Dumesnil J. G.Right and left ventricular dimensions as determinants of ventricular septal motion. Chest691976388393 |
| 21. | Kingma I., Tyberg J. V., Smith E. R.Effects of diastolic transseptal pressure gradient on ventricular septal position and motion. Circulation68198313041314 |
| 22. | Meyer R. A., Schwartz D. C., Benzing G., Kaplan S.Ventricular septum in right ventricular volume overload: an echocardiography study. Am. J. Cardiol.301972349353 |
| 23. | Kerber R. E., Dippel W. F., Abboud F. M.Abnormal motion of the interventricular septum in right ventricular volume overload: experimental and clinical echocardiographic studies. Circulation4819738696 |
| 24. | Popp R. L., Wolfe S. B., Hirata T., Feigenbaum H.Estimation of right and left ventricular size by ultrasound: a study of the echoes from the interventricular septum. Am. J. Cardiol.241969523530 |
| 25. | Pearlman A. S., Clark C. E., Henry W. L., Morganroth J., Itscoitz S. B., Epstein S. E.Determinants of ventricular septal motion. Influence of relative right and left ventricular size. Circulation5419768391 |
| 26. | Louie E. K., Lin S. S., Reynertson S. I., Brundage B. H., Levitsky S., Rich S.Pressure and volume loading of the right ventricle have opposite effects on left ventricular ejection fraction. Circulation921995819824 |
| 27. | Reynertson, S. I., M. G. Mullen, M. R. Costanzo, T. L. McKiernan, and E. K. Louie. 1996. Right ventricular volume overload causes early systolic septal paradox which results in a reduction in left ventricular ejection fraction: evidence for systolic ventricular interaction due to diastolic right ventricular loading (abstract). Circulation 94:I–445. |
| 28. | Nunez B. D., Messerli F. H., Amodeo C., Garavaglia G. E., Schmieder R. E., Frohlich E. D.Biventricular cardiac hypertrophy in essential hypertension. Am. Heart J.1141987813818 |
| 29. | Cuspidi C., Sampieri L., Angioni L., Boselli L., Bragato R., Leonetti G., Zanchetti A.Right ventricular wall thickness and function in hypertensive patients with and without left ventricular hypertrophy: echo-Doppler study. J. Hypertens.71989S108S109 |
| 30. | Nakashima Y., Nii T., Ikeda M., Arakawa K.Role of left ventricular regional nonuniformity in hypertensive diastolic dysfunction. J. Am. Coll. Cardiol.221993790795 |
| 31. | Hayabuchi Y., Matsuoka S., Kubo M., Kuroda Y.Usefulness of Color Kinesis imaging for evaluation of regional right ventricular wall motion in patients with surgically repaired tetralogy of Fallot. Am. J. Cardiol.82199812241229 |
| 32. | Oe M., Gorcsan J., Mandarino W. A., Kawai A., Griffith B. P., Kormos R. L.Automated echocardiographic measures of right ventricular areas as an index of volume and end-systolic pressure-area relations to assess right ventricular function. Circulation92199510261033 |
| 33. | Forni G., Pozzoli M., Cannizzaro G., Traversi E., Calsamiglia G., Rossi D., Cobelli F., Tavazzi L.Assessment of right ventricular function in patients with congestive heart failure by echocardiographic automated boundary detection. Am. J. Cardiol.78199613171321 |
| 34. | Jennesseaux C., Metz D., Maes D., Nazeyrollas P., Tassan-Mangina S., Heuls S., Jamet B., Liehn J. C., Elaerts J.Estimation par quantification acoustique des dimensions et de la fonction systolique du ventricule droit. Arch. Mal. Cœur911998231238 |
Dr. Vignon was supported by the French Ministère des Affaires Etrangères (Fondation Lavoisier), the Société Française de Cardiologie, and the Société de Réanimation de Langue Française.
