Comprehensive Assessment of Left Ventricular Wall Motion Abnormalities in Coronary Artery Disease Using Cardiac Magnetic Resonance

Nineteen patients (65.7 ± 7.4 years old) with triple vessel CAD documented by invasive coronary angiography were included into the study. All patients had proven coronary lesions warranting CABG. The lesions involved at least 70% obstruction in the left anterior descending artery (n=16), circumflex artery (n=12) or right coronary artery (n=15). Apart from this, 14 patients demonstrated stenotic lesions involving their left main segment (at least 50% obstruction). Their pre-and postoperative characteristics, the affected coronary vessels as well as the number of applied coronary grafts have been reported in detail as part of a separate study investigating LV remodeling following CABG [16].

The majority of patients had risk factors for CAD, including hypercholesterolaemia (15 patients), hypertension (12 patients), smoking (9 patients), diabetes mellitus (8 patients), and peripheral vascular disease (6 patients). All patients were receiving medical therapy, including lipid lowering medication (19 patients), ACE inhibitors (17 patients), aspirin (17 patients), beta blockers (15 patients) and diuretics (7 patients). In this group with CAD, Left Ventricular (LV) ejection fraction averaged 40.0 ± 2.5% (mean ± SD), cardiac output 5.6 ± 0.3 l/min, cardiac index 2.7 ± 1.2 l/min/m², end systolic volume 151.5 ± 15.7 ml, end diastolic volume 244.2 ± 17.2 ml, LV mass 162.0 ± 13.9 g, LV mass index 80.1 ± 4.9 g/m². Since CABG is expected to add a variety of confounding factors that affect LV wall motion, only the pre-CABG data were used in the current study.

A group of 9 age-matched healthy controls (66 ± 7 years old) were also studied. All healthy volunteers were non-smokers and underwent a medical assessment prior to acceptance in the study to exclude a history of cardiovascular and metabolic disease. Their baseline data have also been used in a separate study to investigate age-related changes in LV wall motion in healthy individuals [17]. The study was performed according to the principles of the declaration of Helsinki and was approved by a local clinical research ethics committee. Each subject provided written informed consent.

The cardiac MR scans were performed using a 1.5-T Siemens Sonata clinical scanner (Siemens Healthcare, Erlangen, Germany). Pilot images, followed by horizontal and vertical long axis cine images, were acquired using a steady-state Free Precession (TrueFISP) pulse sequence. Cine images for navigator-gated high temporal resolution phase contrast velocity mapping were acquired using a black blood segmented k-space spoiled gradient echo sequence [18,19] with a temporal resolution of 13.8 ms, with prospective ECG-gating and advanced respiratory navigator gating (TR=13.8 ms, flip angle=15º, bandwidth=650 Hz/pixel, FOV=400 x 300 mm, matrix=256 x 96), as previously described [16,20-23]. Three equidistant short axis images were acquired for the LV base, mid-ventricle and apex. The basal slice was positioned parallel to the base of the heart and distal to the LV outflow tract. Basal, mid-ventricular and apical slices were positioned 15-20 mm apart, depending on the heart size. With cardiac and respiratory gating, each short axis acquisition took approximately 3-5 min, with an average of 60-70 phases per cardiac cycle. Velocity encoding was performed by including a phase image with no velocity encoding followed by images with a bipolar gradient in the read, phase or slice direction after each radiofrequency pulse to the otherwise identical sequence (Velocity-Encoded gradient echo imaging (VENC) in-plane=20 cm/s, VENC through-plane=30 cm/s). Post processing was performed in the standard fashion of subtracting the phase from the image with no velocity encoding, followed by conversion of the phase data into velocity maps. Separate measurements were performed for the inner (endocardial) and outer (epicardial) layers. Transmural peak velocities were inferred automatically as averaged peak values of the inner and outer layers. The duration of the cardiac cycle was determined by the R-R interval on the electrocardiogram, with end systole defined as the smallest LV cavity. The cardiac phases were subsequently calculated and normalized for the entire group based on their average duration during a cardiac cycle [16,20,23].

Tissue Phase Mapping (TPM) analysis was performed using customised software (MATLAB^®, v. 6.5; MathWorks^®, Natick, MA). The endocardial and epicardial borders were manually contoured for the base, mid ventricle and apex for each phase of the cardiac cycle. For the TPM analysis, the left ventricle was divided into 16 segments (6 basal, 6 middle and 4 apical) according to the American Heart Association model [24]. In-plane velocities were transformed into an internal polar coordinate system positioned at the centre of the mass of the left ventricle. Global ventricular velocity time courses for radial, circumferential and longitudinal motion were calculated for each group by averaging over the entire segmentation mask. In addition, graphical representations of all ventricular velocities during a cardiac cycle for individual LV segments were obtained. Global ventricular torsion and strain rates were also determined.

MATLAB files containing TPM data were converted and generated into Excel^TM (Microsoft Corporation, Redmond, WA) files for statistical analysis. Subsequently, all data were analysed using SPSS^® v. 16.0 (SPSS Inc. Chicago, IL) and MedCalc statistical software, v. 12.5.0 (MedCalc Software, Ostend, Belgium). The variables were tested for normal distribution and, after the assumption was met, a pooled t-test was used to compare the values between the two groups. Otherwise a non-parametric (Mann-Whitney) test was used. A p-value of less than 0.05 was considered significant.

The peak systolic and diastolic radial velocities at the LV base, mid-ventricle and apex and their corresponding peak times are presented in table 1. Graphical representation of radial velocities at the LV base, mid-ventricle and apex during different phases of the cardiac cycle is shown in figure 1 (a - peak systolic radial velocity, b - peak diastolic radial velocity). Positive values reflect LV segmental motion towards the centre of the ventricle, whilst negative values reflect an outward motion. In patients with CAD, the radial component of ventricular contraction was affected most prominently at the LV apex (peak systolic transmural velocity of 2.4 cm/s in CAD, versus 3.5 cm/s in healthy controls; p-value < 0.01). The contractility of the LV base was also affected, even though to a lower degree (p-value = 0.07). The most affected myocardial layer in patients with CAD was the inner/endocardial layer (Table 1). Whilst LV contraction in CAD was affected more towards the apex (a 11.4% decrease in peak systolic velocity at the LV base versus a 31.4% decrease at the LV apex), the ventricular expansion in diastole was affected to a similar degree at all LV levels (a 30.9% decrease in peak diastolic transmural velocity at the LV base versus a 29.8% decrease at the LV apex). Furthermore, the difference in peak diastolic radial velocities was statistically significant at all ventricular levels (Table 1, all p-values < 0.05). The time to peak systolic and diastolic radial velocities, however, showed no difference between the control and CAD groups.

LV level and myocardial layer		CAD patients	Healthy controls	p- value	CAD patients	Healthy controls	p-value
Peak systolic radial velocity (cm/s)					Time to peak values (% of ES)
LV base	epicardial	2.8 ± 0.2	3.1 ± 0.2	0.11	38.5 ± 2.0	42.7 ± 2.6	0.23
	transmural	3.1 ± 0.2	3.5 ± 0.2	0.07	38.2 ± 2.1	42.6 ± 2.6	0.22
	endocardial	3.4 ± 0.2	4.0 ± 0.2	0.06	38.1 ± 1.9	43.6 ± 2.6	0.11
Mid-ventricle	epicardial	2.7 ± 0.2	3.2 ± 0.3	0.11	36.8 ± 1.8	39.7 ± 1.4	0.33
	transmural	3.0 ± 0.2	3.5 ± 0.3	0.06	39.6 ± 2.3	42.7 ± 2.1	0.42
	endocardial	3.3 ± 0.2	4.0 ± 0.3	0.04	41.1 ± 2.1	44.7 ± 1.5	0.28
LV apex	epicardial	2.4 ± 0.2	3.2 ± 0.2	< 0.01	38.8 ± 1.9	38.6 ± 1.9	0.97
	transmural	2.4 ± 0.2	3.5 ± 0.2	< 0.01	39.7 ± 1.9	39.1 ± 1.7	0.86
	endocardial	2.5 ± 0.2	3.8 ± 0.2	< 0.01	44.3 ± 1.8	40.1 ± 1.4	0.17
Peak diastolic radial velocity (cm/s)					Time to peak values (% of ED)
LV base	epicardial	-3.4 ± 0.3	-4.4 ± 0.4	0.03	18.9 ± 0.9	19.1 ± 1.8	0.90
	transmural	-3.8 ± 0.4	-5.5 ± 0.4	< 0.01	19.0 ± 0.9	18.8 ± 1.9	0.92
	endocardial	-4.3 ± 0.4	-6.6 ± 0.5	< 0.01	19.1 ± 1.0	18.8 ± 1.7	0.85
Mid-ventricle	epicardial	-3.0 ± 0.3	-4.3 ± 0.2	< 0.01	23.8 ± 1.2	23.2 ± 1.2	0.78
	transmural	-3.3 ± 0.3	-5.2 ± 0.2	< 0.01	24.3 ± 1.4	23.5 ± 1.1	0.76
	endocardial	-3.6 ± 0.3	-6.1 ± 0.3	< 0.01	24.1 ± 1.4	23.5 ± 1.1	0.80
LV apex	epicardial	-3.1 ± 0.3	-4.2 ± 0.3	0.01	24.1 ± 1.3	27.4 ± 1.4	0.15
	transmural	-3.3 ± 0.3	-4.7 ± 0.3	< 0.01	24.3 ± 1.2	27.4 ± 1.4	0.16
	endocardial	-3.5 ± 0.3	-5.2 ± 0.4	< 0.01	24.3 ± 1.2	27.4 ± 1.4	0.16

Table 1: Peak radial velocities and time to peak values in patients with CAD and healthy controls*.

* All values are presented as mean ± SEM

 

The radial velocities for individual ventricular segments at the LV base, mid-ventricle and LV apex are displayed in figure 2,3 and 4 respectively. The difference between radial motion of apical segments in the two groups appeared particularly prominent on segmental velocity curves. Segmental graphs also showed that anterior (i.e., segments 1 and 7) and inferior segments (i.e., segments 4 and 10) were affected to a higher degree compared with septal and lateral segments (Figure 2 and 3).

 

Peak clockwise and counter-clockwise circumferential velocities for both study groups are provided in table 2. Positive values reflect clockwise rotation and negative values reflect counter-clockwise rotation as viewed from the apex. The graphical representation of circumferential velocities at the LV base, mid-ventricle and apex is shown in figure 5. An altered pattern of ventricular untwisting was observed in patients with CAD, with much higher peak clockwise velocities at the LV apex (Figure 5, wave h). In healthy volunteers, the circumferential velocity graph at the LV apex showed an undulating pattern of alternating waves of counter-clockwise rotation (Figure 5, dotted line, arrows a, c and e) and waves of sudden deceleration or opposite rotations (arrows b, d and f). In patients with CAD, these waves of alternating rotational motion were less prominent (Figure 5, waves a-h, solid line). For instance, in CAD patients, after the initial counter-clockwise rotation of the entire left ventricle (Figure 5, arrow a), the velocity of the counter-clockwise rotation at the LV apex decreased continually during systole and was subsequently followed by a single wave of clockwise rotation in diastole (arrow h).

 

Figure 5: Circumferential velocity graphs at the LV base (panel A), mid-ventricle (panel B) and LV apex (panel C).

Positive values reflect clockwise rotation and negative values reflect counter-clockwise rotation as viewed from the apex. Solid line - patients with CAD; Dotted line - healthy controls. In healthy volunteers, the circumferential velocity graph at the LV apex showed an undulating pattern of alternating waves of counter-clockwise rotation (panel C, dotted line, waves a, c and e and waves of sudden deceleration or opposite rotations (waves b, d and f). In CAD patients, these waves of alternating rotational motion were less prominent (panel C, solid line, waves a-h). An altered pattern of ventricular untwisting was also observed in patients with CAD, with much higher peak clockwise velocities at the LV apex (wave h). Cardiac cycle phases: 1 - isovolumetric contraction; 2 - rapid ejection; 3 - reduced ejection; 4 - isovolumetric relaxation; 5 - rapid filling; 6 - diastasis; and 7 - atrial systole; ED: End Diastole; ES: End Systole

LV level and myocardial layer		CAD patients	Healthy controls	p- value	CAD patients	Healthy controls	p-value
Peak clockwise velocity (cm/s)					Time to peak values (% of ES)
LV base	epicardial	2.4 ± 0.2	2.6 ± 0.3	0.35	51.2 ± 2.0	47.5 ± 1.1	0.25
	transmural	2.3 ± 0.2	2.5 ± 0.3	0.27	51.5 ± 2.0	47.5 ± 1.1	0.21
	endocardial	2.2 ± 0.2	2.5 ± 0.3	0.24	51.2 ± 1.8	46.3 ± 1.1	0.10
Mid-ventricle	epicardial	2.9 ± 0.4	2.0 ± 0.2	0.09	137.5 ± 2.3	101.2 ± 0.6† 147.5 ± 5.1†	< 0.01
	transmural	2.7 ± 0.4	2.0 ± 0.3	0.14	137.0 ± 2.2	99.7 ± 2.0† 147.5 ± 5.1†	< 0.01
	endocardial	2.6 ± 0.4	2.2 ± 0.3	0.25	140.3 ± 3.2	99.7 ± 2.0† 147.5 ± 5.1†	< 0.01
LV apex	epicardial	4.0 ± 0.5	2.3 ± 0.3	0.02	137.1 ± 2.5	109.9 ± 4.6	< 0.01
	transmural	3.8 ± 0.5	2.2 ± 0.2	0.02	137.6 ± 2.5	109.9 ± 4.7	< 0.01
	endocardial	3.6 ± 0.5	2.2 ± 0.2	0.03	135.3 ± 3.1	111.4 ± 5.1	< 0.01
Peak counterclockwise velocity (cm/s)					Time to peak values (% of ED)
LV base	epicardial	-3.7 ± 0.4	-3.3 ± 0.4	0.25	117.7 ± 2.1	117.6 ± 1.5	0.98
	transmural	-3.5 ± 0.4	-3.3 ± 0.4	0.34	119.4 ± 1.9	118.1 ± 1.5	0.70
	endocardial	-3.4 ± 0.4	-3.4 ± 0.3	0.48	119.0 ± 2.0	118.6 ± 1.7	0.98
Mid-ventricle	epicardial	-3.8 ± 0.4	-3.5 ± 0.4	0.34	27.8 ± 1.6	25.3 ± 1.9	0.39
	transmural	-3.6 ± 0.4	-3.5 ± 0.3	0.44	27.2 ± 1.6	25.3 ± 1.9	0.51
	endocardial	-3.4 ± 0.4	-3.5 ± 0.3	0.46	27.6 ± 1.4	25.3 ± 1.9	0.41
LV apex	epicardial	-4.7 ± 0.4	-3.8 ± 0.4	0.07	29.5 ± 1.3	26.2 ± 1.1	0.13
	transmural	-4.5 ± 0.4	-3.7 ± 0.4	0.13	24.3 ± 1.2	26.2 ± 1.1	0.11
	endocardial	-4.2 ± 0.4	-3.7 ± 0.4	0.20	29.6 ± 1.3	26.7 ± 1.4	0.24

 

Whilst in healthy volunteers, the LV apex at the end of systole was rotating in a clockwise direction (Figure 5, dashed line, arrow d), in CAD patients the apex continued its counter-clockwise rotation, with a prominent delay in rotational motion. In healthy controls, the recoil clockwise rotation of the LV apex started after repolarization and before the end of systole (Figure 5, dotted line, wave d becomes positive before the end systole), whilst in CAD patients it started much later, during the second part of isovolumetric relaxation. In diastole, the LV apex in healthy controls showed two separate waves of recoil clockwise rotation (Figure 5, dotted line, arrows d and f), alternating with opposing waves of counter-clockwise motion. In CAD patients, however, the recoil clockwise rotation of the LV apex started much later and continued as a single wave (Figure 5, solid line, arrow h). Thus, during the first half of rapid ventricular filling, the LV apex was rotating in a clockwise direction in CAD patients and in a counter-clockwise direction in healthy controls (Figure 5, dotted line is below zero during the first half of phase 5, while solid line is well above zero during the entire phase 5). Of interest are also the higher peak velocities during the initial counter-clockwise rotation at the commencement of systole, obtained in CAD patients (Table 2). The difference was more pronounced at the LV apex (4.5 cm/s in CAD patients compared with 3.7 cm/s in healthy controls), although it did not reach statistical significance in our study (p-values between 0.07 and 0.2).

The time to peak counter-clockwise velocity at the LV apex and mid-ventricle in both control and CAD groups was represented by the initial counter-clockwise rotation of the entire ventricle (Figure 5 (B,C), arrow a), whilst at the LV base by the first wave of recoil counter-clockwise rotation of the LV base during untwisting (Figure 5A, arrow e).

Solid line - patients with CAD; Dotted line - healthy controls; Segments: 1 - basal anterior segment; 2 - basal anteroseptal segment; 3 - basal inferoseptal segment; 4 - basal inferior segment; 5 - basal inferolateral segment; and 6 - basal anterolateral segment; Cardiac cycle phases: 1 - isovolumetric contraction; 2 - rapid ejection; 3 - reduced ejection; 4 - isovolumetric relaxation; 5 - rapid filling; 6 - diastasis; and 7 - atrial systole; ED: End Diastole; ES: End Systole

The time to peak clockwise velocity at the LV base was represented in both groups by the clockwise rotation of the ventricular base during ejection (Figure 5A, arrow b). At the mid-ventricular level, the time to peak clockwise velocity in healthy controls was divided in accordance with two different waves of clockwise rotation located at 99.7% and 147.5% of end-systole (Figure 5B, arrows d and f, dotted line). In different subjects, the time to peak clockwise velocity at the mid-ventricle was represented by any of these two peaks, which is also reflected in table 2). In CAD patients, the time to peak clockwise velocity at the mid-ventricular level was represented by a single prominent wave of clockwise rotation transmitted from the apex (Figure 5B, arrow h, solid line). At the LV apex, the time to peak clockwise velocity in healthy controls was represented by the first recoil wave of clockwise rotation (Figure 5C, arrow d, dotted line), whilst in CAD patients-by the delayed wave of apical untwisting with a prominent single peak of clockwise recoil motion (Figure 5C, arrow h, solid line).

The circumferential velocities for individual ventricular segments at the LV base, mid-ventricle and LV apex are displayed in figure 6,7 and 8 respectively. An altered pattern of rotational motion is observed in patients with CAD, particularly affecting apical and, to a lower degree, mid-ventricular and basal LV segments.

Peak longitudinal velocities with corresponding time to peak values for main LV slices are presented in table 3, while the graphical display is provided in figure 9 (a - peak systolic longitudinal velocity, b - peak diastolic longitudinal velocity). Positive values demonstrate downward motion along the longitudinal LV axis (i.e., towards the apex) and negative values show upward motion (i.e., towards the base). Peak systolic longitudinal velocities in both CAD patients and healthy controls were represented by the downward motion of the left ventricle during rapid ejection (Figure 9, arrow a), when the entire ventricle was pushed in an opposite direction to the ejected blood. The values were much lower in CAD patients, being statistically significant at all levels (Table 3). Peak diastolic velocities in both groups were represented by a recoil wave of upward motion in diastole (Figure 9, arrow b). The peak diastolic longitudinal velocities at the LV base in CAD patients were almost half of those recorded in healthy controls (Table 3), but the difference decreased towards the apex. No difference was noted in the time to peak velocity values (Table 3).

 

LV level and myocardial layer		CAD patients	Healthy controls	p- value	CAD patients	Healthy controls	p-value
Peak systolic longitudinal velocity (cm/s)					Time to peak values (% of ES)
LV base	epicardial	5.8 ± 0.5	7.8 ± 1.0	0.02	29.5 ± 1.6	26.0 ± 1.2	0.14
	transmural	5.5 ± 0.5	7.9 ± 0.9	0.01	29.8 ± 1.6	25.5 ± 1.1	0.07
	endocardial	5.3 ± 0.5	8.0 ± 0.9	< 0.01	29.5 ± 1.8	25.0 ± 0.9	0.08
Mid-ventricle	epicardial	4.5 ± 0.4	6.0 ± 0.6	0.03	28.1 ± 1.4	25.1 ± 1.2	0.20
	transmural	4.0 ± 0.4	5.9 ± 0.5	< 0.01	28.1 ± 1.4	25.7 ± 1.3	0.31
	endocardial	3.6 ± 0.5	5.8 ± 0.5	< 0.01	28.4 ± 1.3	25.7 ± 1.3	0.24
LV apex	epicardial	2.2 ± 0.4	3.9 ± 0.6	0.01	26.6 ± 1.8	25.8 ± 1.5	0.77
	transmural	2.0 ± 0.4	3.7 ± 0.6	0.02	26.6 ± 1.8	25.8 ± 1.5	0.77
	endocardial	1.9 ± 0.5	3.5 ± 0.7	0.04	26.9 ± 1.7	25.8 ± 1.5	0.68
Peak diastolic longitudinal velocity (cm/s)					Time to peak values (% of ED)
LV base	epicardial	-7.8 ± 0.5	-13.6 ± 1.1	< 0.01	22.6 ± 1.1	20.9 ± 0.8	0.39
	transmural	-7.9 ± 0.6	-14.1 ± 1.1	< 0.01	22.9 ± 1.2	20.6 ± 0.8	0.26
	endocardial	-8.0 ± 0.6	-14.7 ± 1.2	< 0.01	23.5 ± 1.3	20.1 ± 0.7	0.14
Mid-ventricle	epicardial	-7.0 ± 0.5	-10.6 ± 0.7	< 0.01	23.5 ± 1.2	24.4 ± 0.8	0.69
	transmural	-6.6 ± 0.5	-10.7 ± 0.7	< 0.01	23.7 ± 1.4	24.4 ± 0.8	0.78
	endocardial	-6.6 ± 0.6	-10.7 ± 0.7	< 0.01	24.0 ± 1.6	24.1 ± 0.8	0.99
LV apex	epicardial	-6.1 ± 0.5	-7.0 ± 0.6	0.20	21.1 ± 1.4	23.0 ± 2.0	0.52
	transmural	-6.0 ± 0.6	-6.9 ± 0.8	0.18	20.2 ± 1.4	23.0 ± 2.0	0.33
	endocardial	-6.1 ± 0.6	-6.8 ± 0.7	0.23	20.3 ± 1.4	20.7 ± 2.2	0.91

 

The longitudinal velocities for individual ventricular segments at the LV base, mid-ventricle and LV apex are displayed in figure 10,11 and 12 respectively. Since the apical segments (i.e., segments 13-16) were relatively stable with little motion along LV longitudinal axis even in healthy controls, the biggest difference between the two groups was recorded in longitudinal motion of basal LV segments (i.e., segments 1-6).

 

The peak systolic torsion rate was significantly lower in patients with CAD compared with healthy controls (15.1 ± 1.0 versus 20.6 ± 2.0 degrees/s/cm, p-value 0.01, Table 4 and Figure 13). A lower peak diastolic torsion rate was also apparent in CAD group (-19.6 ± 2.1 versus -25.2 ± 1.8 degrees/s/cm), although the difference did not reach statistical significance in our study (p-value=0.12). The greatest difference between CAD and control subjects, however, was recorded for peak diastolic longitudinal strain rate values, which were approximately 2.5 times lower in patients with CAD (1.1 ± 0.2 versus 2.7 ± 0.1 s^-1, p-value < 0.01) (Table 4 and Figure 14). The difference between the two groups was also significant for peak systolic longitudinal strain rate (p-value < 0.01).

 

 

Table 4: Peak ventricular torsion rate and peak longitudinal strain rate in patients with CAD and healthy controls*.

* All values are presented as mean ± SEM

The lower systolic radial velocities in the CAD group (Table 1) underscored decreased inward LV contractility in these patients, with contraction of LV apical segments being affected to the greatest extent. The findings are consistent with previous reports showing a greater dysfunction in distal compared to proximal myocardial segments following induced ischemia in both animal and human studies [25,26]. Radial expansion in diastole was affected even more than inward motion in systole, involving also the LV base in patients with CAD (Table 1). Meanwhile, there was little change in the time to peak radial velocities, indicating these times are not informative for the assessment of CAD. The greatest difference between peak radial velocities was recorded for the inner/endocardial layer of the myocardium, which is consistent with reports that subendocardial layer receives more blood flow and has a higher energy requirement [11,27], being therefore more sensitive to ischemic changes [11,28,29]. In support of this hypothesis, graded ischemia applied in experimental studies also affected initially the subendocardium before extending to the midwall and epicardial layers [1,30]. In patients with severe CAD, the reduced velocities at the subendocardial level can be also related to subendocardial fibrosis and perhaps subendocardial necrosis caused by severe, chronic subendocardial ischemia [31]. It is well known that such lesions are commonly localized in the subendocardium, while non-ischemic cardiomyopathies often display an intramyocardial or subepicardial pattern [32].

The circumferential velocity traces demonstrated not only differences in peak velocities, but also an altered pattern of rotational motion. The higher initial counterclockwise velocities obtained in CAD are consistent with experiments of Kroeker et al., [33], who showed that ischemia induced by a short period of coronary artery occlusion in animal models resulted in increased counter-clockwise rotation. As a potential explanation the authors suggested that subendocardial ischemia led to a loss of counteraction of contraction of subendocardial cardiomyocytes, which have a clockwise orientation [33,34].

The delay in LV apical untwisting in CAD subjects, made the clockwise rotation of the LV apex (Figure 5C, wave h, solid line) coincide at least partially with the clockwise rotation of the LV base (Figure 5A, wave f, solid line). This concordance in motion direction of the LV apex with the massive LV base may explain the much higher peak clockwise velocities at the LV apex (and even mid-ventricle) recorded in patients with CAD (Table 2 and Figure 5C, wave h, solid line). The altered rotational pattern of the LV apex might also explain the higher peak counter-clockwise velocities at the LV base (despite lower peak clockwise values) recorded in CAD patients. During the first half of isovolumetric relaxation, the LV apex in healthy subjects was rotating in a clockwise direction (Figure 5C, wave d, dotted line), which was opposite to the rotation of the LV base (Figure 5A, wave e, dotted line). In CAD patients, however, the LV apex was still rotating counter-clockwise during the first half of isovolumetric relaxation, the velocity trace showing negative values (Figure 5C, solid line). Thus, in healthy controls, the LV base and apex were rotating in opposite directions, whilst in CAD patients the counter-clockwise rotational recoil of the LV base encountered little or no resistance from the LV apex, resulting in higher peak values. These findings demonstrate how an altered rotational pattern of the ventricular apex can directly affect the rotational motion of the LV base. Significant alterations of LV rotational pattern were also noted during the phase of rapid ventricular filling, when the LV apex was rotating clockwise in CAD patients (Figure 5C, wave h) and counter-clockwise in healthy controls (Figure 5C, wave e). Most of these alterations of rotational pattern are expected to directly affect the ventricular physiology and LV filling. Correlative studies involving ventricular hemodynamics may shed new light into these mechanisms.

To provide a better understanding of the altered pattern of rotational motion of the LV apex and base against each other, we compared the circumferential velocity graphs with those obtained in a younger age group (healthy controls 23 ± 3 years old from an unrelated study [17]) (Figure 15). The graphs showed successive changes, with initial alterations of the rotational motions of the left ventricle in older age group and subsequent progressive loss of the undulating pattern of ventricular twisting in patients with CAD. Of particular note is that the peak counter-clockwise velocity at the LV base in younger subjects was represented by the initial wave of counter-clockwise rotation of the entire ventricle at the commencement of systole (Figure 15A, solid line, wave b), whilst in the older age group, as well as in patients with CAD it was represented by a recoil wave of ventricular untwisting in diastole (Figure 15B and C, solid line, wave b).

 

The greatest difference between myocardial velocities of CAD patients and healthy controls was found in ventricular longitudinal velocities (Figure 9). Hence, determination of longitudinal velocities may be one of the most sensitive tests for evaluation of myocardial contractility. During systole, the entire ventricle was pushed in an opposite direction to the ejected blood, synonymous to the recoil of a canon after being fired. This motion pattern was seen as a wave of downward displacement at all LV levels (arrow a, Figure 9). The amplitude of this longitudinal motion is expected to be proportional to the force of the ejected blood and, consequently, to the summation of all elements of LV contraction, including radial and rotational motions. For example, altered myocardial contractility is likely to result in a lower radial velocity and contractility; altered rotational motion is expected to affect the force of LV twisting; potential myocardial scars might result in akinetic and dyskinetic regions. All these factors potentially affect the overall force of the ejected blood and, consequently, the force with which the ventricle is being pushed in an opposite direction, stretching the great vessels and elastic structures to which it is attached. After repolarization, these stretched elastic structures start pulling the displaced ventricle (i.e., “fired canon”) upward towards its initial location. The velocity of this longitudinal recoil motion (arrow b, Figure 9) was expected to be proportional to the overall force applied during systole to stretch the great vessels and elastic structures attaching the ventricle to mediastinum. Consequently, peak diastolic longitudinal velocity can serve as a measure of global ventricular contractility, the resulted recoil in diastole being proportional to the sum of all applied forces during systole. In addition, peak diastolic longitudinal velocity represents the highest velocity of the myocardial wall. The highest longitudinal velocities recorded in this study exceeded -14 cm/s (Table 3), while the highest radial and circumferential velocities were -6.6 cm/s (Table 1) and -4.7 cm/s (Table 2), respectively. In a failing heart, one might expect to see the greatest difference in the fastest myocardial motions, i.e., in diastolic longitudinal velocities. The findings are consistent with previous reports that the amplitude of long axis motion provides a useful approximation of the ejection fraction [35].

Since the ventricular apex shows little motion along the longitudinal axis, due to concomitant LV shortening/twisting in systole and elongation/untwisting in diastole, the greatest difference in longitudinal motion can be expected at the LV base. In this study, the peak longitudinal velocities at the LV base were approximately two times lower in CAD patients compared with the values registered in healthy controls, representing the most significant difference between the two groups (Figure 9A, wave b). These findings were confirmed by similar differences in longitudinal strain rates (Table 4 and Figure 14). Our results are concordant with the findings reported in patients with acute myocardial ischemia, indicating reduced global longitudinal strain as the single most powerful marker of manifest left ventricular hemodynamic deterioration in the acute phase of myocardial infarction [36]. After the acute phase, longitudinal strain also proved an important predictor for segmental and global LV function improvement [37]. Furthermore, peak systolic longitudinal strain at rest was reported useful in predicting the presence of left main and three vessel coronary artery disease even before the development of regional wall motion abnormalities [38].

We also believe the findings would complement our previous studies [16,17], providing a broader understanding of underlying changes in left ventricular wall motion in elderly individuals, patients with coronary artery disease and subsequent ventricular remodeling following revascularization.

The cardiac phases were calculated and normalized for the entire group based on their average duration during systole and diastole and not on the timing of cardiac valves closure and opening, which was not recorded. The phases are provided for general orientation and even when slightly displaced would not affect data interpretation. Visual assessment of individual segmental radial, circumferential and longitudinal velocity graphs revealed no significant difference to account for a particular motion pattern in any subject (e.g., large scars, regional paradoxical motion, etc.). Since all patients included in the study had triple-vessel coronary artery disease, no significant difference averaged between affected myocardial regions was expected. Given the number of resulting parameters (16 segments x 3 myocardial layers x (3 velocities + 3 time to peak values)=288 parameters), no myocardial velocities for individual segments were calculated in this case, the evaluation being more suitable for patients with regional rather than global CAD. Mechanical tethering of the adjacent segments to infarcted or ischemic regions has also been described [39-41]. In a cohort of patients with triple vessel disease and scattered affected areas throughout the entire LV, evaluating regional motion of the effected LV segments becomes especially challenging. Evaluating segmental velocities and regional myocardial motion in patients with a single-vessel coronary disease and correlating the affected coronary artery with recorded regional wall motion abnormalities may represent an area of interest for further research. Calculating the regional strain and strain rates of the affected LV segments would provide additional insight into regional LV wall motion in such patients [42,43]. Furthermore, correlating the TPM analysis technique with late gadolinium enhancement studies may provide new details about the impact of myocardial fibrosis and infarcted areas on regional and global LV mechanics. This study was also limited to the assessment of ventricular wall motion only. It has been reported that LV wall motion analysis combined with ECG changes and serum markers of myocardial necrosis may allow recognition of stress cardiomyopathies in different clinical settings [44]. In the current study, the endocardial and myocardial borders were delineated manually, being rather time consuming. Programs for automatic delineation that can facilitate the clinical applications of the technique are currently under development. Correlation with other imaging modalities, including with speckle tracking echocardiography in larger populations may further define the suitability and clinical role of TPM among the existing cardiac investigations.