Diagnostic Accuracy of Strain Imaging of the Left Ventricle in Detection of Obstructive Coronary Artery Disease

Background: Speckle tracking echocardiography using average global strain and strain rate (SR) parameters for early detection of high risk patients with coronary artery disease (CAD) has gained a substantial clinical interest. Ob-jective: Assessment of the diagnostic accuracy of strain imaging of the left ventricle in detection of obstructive coronary artery disease. Methods: One hundred patients were enrolled. They were divided into 3 groups: Group I (n = 40 patients) presented with ST segment elevation myocardial infarction (STEMI), Group II (n = 40 patients) presented with non-ST segment elevation myocardial infarction (NSTEMI) or unstable angina (UA) and Group III (n = 20 patients) with normal coronary angiography served as a control group. All patients were subjected to 2D speckle tracking echocardiography (2D STE) to assess LV longitudinal strain and strain rate (SR). Sensitivity, specificity and diagnostic accuracy of 2D STE in prediction of CAD and its severity using Gensini score were assessed. Results: The mean age for Group I, II and III was 52.20 ± 11.83, 51.97 ± 14.53 and 52.75 ± 10.75 respectively. LV average global systolic strain (AGS) was significantly lower in group I and II when compared to group III. AGS and average global systolic SR showed significant direct correlation with Gensini score. The diagnostic accuracy of 2D STE in prediction of significant LAD stenosis was 92.5%, and it was 89.5% in prediction of 3 vessels CAD. Conclusion: Strain imaging using 2D STE can predict the territory and severity of CAD with high diagnostic accuracy and can be used as a simple noninvasive diagnostic tool to identify high risk CAD patients.


Introduction
Noninvasive identification of patients with coronary artery disease (CAD) remains a clinical challenge despite the widespread use, and possible overuse, of imaging and provocative testing; more than 50% of patients currently referred to coronary angiography show normal or non-obstructive CAD [1].
Recently, the 2-dimensional Speckle Tracking Echocardiography (STE) and its derivatives, including strain and strain rate imaging have been gained substantial clinical interest. Left ventricular longitudinal strain, derived using two-dimensional speckle-tracking echocardiography, has emerged as a noninvasive marker of both global and regional LV dysfunction in patients at risk for developing CAD [2] [3].
Current evidence supports the use of GLS in the detection of moderate to severe obstructive CAD in symptomatic patients. GLS may complement existing diagnostic algorithms and act as an early adjunctive marker of cardiac ischemia [4].
We aimed to assess whether the diagnostic accuracy of 2D-STE global and segmental longitudinal strain (LS) measured at rest may compare to the parameters which can be measured today during conventional Echo. to predict angiographically obstructive CAD.

Patient Population
This is a cross-sectional randomized study including a total of 100 patients admitted to Menoufia university hospitals. Patients were enrolled in the study after obtaining their written informed consent, and approval of the local ethics committee of the hospital.
Patients who included in the study were divided according to their presentation into 3 groups; Group I (n = 40 patients): patients presented with ST segment elevation myocardial infarction (STEMI), Group II (n = 40 patients): patients presented with non-ST segment elevation myocardial infarction (NSTEMI) or unstable angina (UA) and Group III (n = 20 patients): patients whose coronary angiography revealed normal vessels, they were referred for catheterization laboratory for suspected CAD with one or more of: typical ischemic chest pain, electrocardiographic (ECG) changes suggesting CAD, positive exercise stress test, positive myocardial perfusion imaging or echocardiographic regional wall motion abnormality suggesting CAD.

Methods
All patients were subjected to detailed history, including CAD risk factors, physical examination, Electrocardiography (ECG) and laboratory investigations included, complete blood picture, liver and kidney function. Some patients-who did not have history of CAD-performed a stress imaging test to diagnose CAD.
Echocardiographic images were obtained in the parasternal long-axis and short-axis and apical two-chamber and four-chamber views using standard transducer positions. Vivid 9, General Electric Healthcare (GE Vingmed, Norway) equipped with a harmonic M5S variable-frequency (1.7 -4 MHz) phased-array transducer was used. LV dimensions and wall thickness, EF, and left atrial diameter and volume were measured in accordance with the recommendations of the American Society of Echocardiography [5].
Peak early (E) and late (A) transmitral filling velocities were measured from mitral inflow velocities. Early diastolic (e') velocity was obtained by placing a tissue Doppler sample volume at the septal and lateral mitral annulus in the apical four-chamber view, and the mean value was obtained. The E/e' ratio was also calculated.
Blood samples were collected before cardiac catheterization. Patients fasted for >12 hours before cardiac catheterization. Blood was collected either from the antecubital vein or indwelling catheter into two 3.2% trisodium citrate tube after discarding the initial 3 ml of blood. Serum was separated by centrifugation at 2000 g for 15 minutes and stored at −70˚C. CBC, liver and renal function were measured.
2D strain analysis was performed offline using the Echopac software (General Electric version 1.8.1.X-Vingmed). All strain images were obtained at a frame rate of 59 -82 frame/s. Three consecutive cardiac cycles were acquired at end expiration breath holding and digitally stored on a hard disk for off-line analysis.
In order to measure the timing of cardiac events, LV inflow (mitral) and outflow (aortic) velocities were recorded using Pulsed wave Doppler. Longitudinal strain and strain rate were assessed in the 6 LV walls and the software algorithm automatically segmented the LV into an equidistant segment model in a ''bull's eye'' plot and each segment was individually analyzed.
The average value of peak longitudinal strain (Esys), the average value of peak systolic strain rate (SRs), peak early diastolic strain rate (SRe), and peak late diastolic strain rate (SRa) were measured at each segment (basal, mid and apical) then the global LV strain and the global LV systolic and diastolic SR were obtained from averaging the peak values of 18 LV segments and used for comparison between the studied groups.
Right and left coronary angiography was performed using multiple projections and analysis was done by professional interventionists who were blind to the cases. Angiographic CAD is defined as ≥50% luminal diameter stenosis of at least one major epicardial coronary artery.
Assessment of the severity of CAD for each coronary angiogram was done  Figure 1). Also severity of CAD was assessed using coronary vessel score in which severity was assessed as no vessel disease, one vessel disease, two vessels disease and three vessels disease.
For STEMI patients, they underwent either rescue PCI after failed thrombolytic therapy or routine PCI strategy after successful reperfusion. Meanwhile for NSTE-ACS very high risk patients were excluded and coronary angiography was done for those who need intervention during the first 24 or 72 hours.

Statistical Analysis
Statistical analysis was performed using Statistical Package for The Social Sciences Version 22 (IBM Corp., Armonk, NY, USA). Quantitative data are expressed as means and standard deviations.
P-Value ≤ 0.05 was considered to indicate significance. Correlation analysis assesses the strength of association between two variables. Multiple logistic regression analysis was used to detect the associations between lipid profile and both maternal and neonatal complications.

Results
The patients were classified into 3 groups according to presentation into group I, 40 patients with STEMI and group II, 40 patients with NSTE-ACS and 20 patients represents control group III. There was a non-significant difference between the studied groups regarding to demographic data (age, gender) and risk factors of CAD (Table 1).
There was a significant difference between group II and control regarding LVESD with p value 0.008, there was also a significant difference between Group I, II and III regarding EF% with p values 0.019, <0.001 and 0.011 respectively.
Regarding FS%, there was a significant difference between Group I, II and III with p values 0.018, <0.001 and 0.013 respectively, with no significant difference regarding other parameters (Table 2).   Figure 2).
There was also a statistically significant difference for all walls AGS when comparing group I with controls and group II with controls with p values less than 0.05 for all, with non-significant values when comparing group I with group II with p values more than 0.05 except for inferior wall AGS with p value 0.043 (Table 4, Figure 2).
Regarding cumulative peak systolic strain rate (SRs), it was significant when comparing group I with controls and group II with controls with p value less than 0.001 for both. For cum. SRe & SRa, it was significant when comparing group I with controls and group II with controls with p value less than 0.001 (Table 5).
As regarding angiographic finding of the patients groups in detection of diseased     (Figure 3). Regarding the validity of 2D speckle tracking in detection of diseased coronary artery, it was good predictor for LAD stenosis with 90% sensitivity and 60% specificity, for LCX stenosis 86.5% sensitivity and 82.1% specificity and finally for RCA stenosis 88.5% sensitivity and 73.7% specificity (Table 6). Also, Figure  7 shows segmental longitudinal strain and strain rate in patient with subtotal proximal LAD occlusion, there is reduced regional strain and strain rate in the distribution of diseased LAD.
Regarding the number of stenotic vessels detected by coronary angiography for patients' groups, 18.75% patients had one vessel disease, 28.75% patients had two vessels disease and 52.5% had three-vessel disease and for suspected by 2D speckle tracking, 20% patients had one vessel disease, 32.5% patients had two vessels disease and 47.5% had three-vessel disease (Figure 4).
Regarding the validity of 2D speckle tracking in detection of number of stenotic coronary arteries, it was good predictor for three vessels disease with 88% sensitivity and 78.9% specificity, for two vessels disease 86.9% sensitivity and 71.7% specificity and finally for one vessel disease stenosis 80% sensitivity and 91.8% specificity ( Table 7).
The best cut-off value considering AGS% in prediction of severity of coronary artery disease based on Gensini score is −11 with 93.3% sensitivity and 98.4% specificity, for SRs is −0.67 with 86.7% sensitivity and 89.2% specificity, for SRe is 0.89 with 86.7% sensitivity and 72.3% specificity, finally for SRa is 0.85 with 93.3% sensitivity and 69.2% specificity (Table 8, Figure 5).      (Table 9, Figure 6). So, the best Table 9. Multivariate linear regression analysis for independent predictors of severity of coronary artery disease based on Gensini score.

Discussion
Noninvasive identification of patients with coronary artery disease (CAD) remains a clinical challenge despite the widespread use, and possible overuse, of imaging and provocative testing; more than 50% of patients currently referred to coronary angiography show normal or non-obstructive CAD [1].
The 2-dimensional Speckle Tracking Echocardiography (STE) and its derivatives, including strain and strain rate imaging have been gained substantial clinical interest to surmount those limitations [3].
We enrolled 100 patients in our study with suspected coronary artery disease, STE using AGS was done to those patients followed by coronary angiography.
Our results found no statistically significant difference between the three groups as regarding demographic data and risk factors. This was contrary to Conte L. et al. [6] who measured the degree of longitudinal fibers dysfunction by STE in diabetic and obese patients and found reduction in GLPS in those patients. Also, Rostamzadeh et al. stated that of conventional risk factors, diabetes mellitus (55%) and hypertension (67%) were more frequent in the high-risk group when compared with low-risk and normal groups [1].
We found that significant difference was present between the three groups as the optimal values of GLS and GLSR, for detecting the severe CAD were −17% and −1 s −1 with the sensitivities of 77% and 71% and the specificities of 63% and 67% respectively [1].
They also stated that the GLSR was also significantly lower in the high-risk group when compared with the low-risk and normal groups (−0.9 ± 0.2, −1 ± 0.2 and −1.1 ± 0.2 s −1 , respectively, p = 0.004), in agreement with our data [1]. Tanaka et al. stated that the SRe was significantly correlated with SRs in all studied segments (r = 0.55, p < 0.0001). The average values of SR indices in each patient were defined as SSR index and ESR index, which were derived by dividing the summed value of the SRs or SRe by the number of respective assessable segments. A significant positive relationship was observed between the SRs index and left ventricular ejection fraction (r = 0.85, p < 0.0001) [12].
We also found that reduced global systolic strain associated with reduced strain in group of segments (territory), are significantly correlated to coronary artery lesions discovered by coronary angiography.
This was in agreement with Fahim et al. who predicted the extent and site of lesions in coronary arteries in patients suspected to have stable CAD depending World Journal of Cardiovascular Diseases on values of segmental strain [13].
These results were supported by Vrettos et al., who correlated the global longitudinal peak systolic strain with the site and severity of coronary artery lesions assessed by coronary angiography [14].
This relation between segmental longitudinal strain values reduction by STE and site of coronary artery lesion by coronary angiography is also declared by Moustafa et al. [15].
While, Montgomery et al., declared that abnormal myocardial deformation at rest identifies significant CAD especially for LAD stenosis [16].
Also, Raja, established that systolic strain has the highest specificity for detection of obstructive CAD in LAD followed by RCA and LCX [17].

Limitations
The study lacked a large validation population. Further prospective studies are thus needed to confirm our results. Also, the definition of coronary arteries disease was based on angiographic views of 2D X-ray, we did not use IVUS or FFR which may interfere with the decision of the interpretation of coronary angiography.

Conclusion
Strain imaging using 2D STE can predict the territory and severity of CAD with high diagnostic accuracy and can be used as a simple noninvasive diagnostic tool to identify high risk CAD patients.