Visual left ventricular (LV) wall motion scoring is well established for the assessment of LV function, yet it is subjective, circumstantial, and relative and requires long training. Quantification of myocardial shortening (strain) using two-dimensional speckle-tracking is potentially less subjective. In this study, quantifiable LV contraction (two-dimensional strain) was prospectively cross-related with wall motion score (WMS) and radionuclide myocardial perfusion imaging (MPI) score in 20 patients (mean age, 54 ± 9 years) with acute myocardial infarctions, early and late after percutaneous revascularization.
Echocardiography and rest MPI were performed 3 to 5 days after acute myocardial infarction. Echocardiography was repeated at 4 months. Peak segmental and global endocardial longitudinal strain (LS) and circumferential strain (CS) were measured, and principal strain was calculated. Volumes, WMS, MPI scores, and strain were assessed independently.
Two-dimensional strain, visual WMS, and radionuclide MPI score correlated closely. Strain thresholds for abnormal WMS were 11.7% for early LS, 18.2% for early CS, 13.9% for late LS, and 19.1% for late CS. Late principal strain correlated better with WMS and MPI score than either LS or CS. CS varied minimally over time, while LS improved in most segments. Higher early CS (>15%) was predictive of segmental functional recovery. MPI score correlated better with late rather than early strain, probably because early resting perfusion defects represent permanent damage.
In this pilot study, strain correlated with echocardiographic WMS and the extent of ischemia (MPI score) early and late after revascularization in patients with acute myocardial infarction. Longitudinal and circumferential strain uncoupling was observed. LS appeared to be more sensitive to acute ischemia, whereas CS correlated better with improvement after revascularization.
Visual left ventricular (LV) wall motion scoring is a well-established method for the assessment of LV function. Segmental wall motion assessment plays an important role in the identification of an occluded culprit coronary artery in acute myocardial infarction (MI) and in the assessment of coronary artery disease at rest and during stress. Visual wall motion assessment is performed on the basis of short-axis and long-axis views. Echocardiographers are traditionally trained to use mostly radial thickening and circumferential shortening to score segmental function, while longitudinal shortening serves only as an aid to assessment. Yet visual wall motion scoring has some major limitations. First, it is subjective and usually requires a long period of training to become reproducible. Second, the score is largely comparative, rather than absolute, and varies according to circumstances (e.g., hyperdynamic state, bradycardia). Within a patient, segments are compared with adjacent and remote segments (best contraction may be designated as “normal” compared with worst, but it may still be abnormal). Depending on experience, function is compared with previously assessed contraction and previous impressions of normal and abnormal contraction. Third, longitudinal function, which is less appreciated on visual assessment, may actually be more sensitive to acute ischemia, because it represents subendocardial longitudinal fiber contraction. And last, wall motion score (WMS) is not a continuous variable.
Quantification of myocardial shortening (strain) using two-dimensional (2D) speckle-tracking is now available, with the potential of providing a less subjective assessment of global and segmental function. It may also be capable of detecting much more subtle degrees of functional impairment in ischemic tissue, such as detected by myocardial perfusion imaging (MPI). Uniplane (longitudinal) global and regional strain has been previously shown to correlate well with visual WMS in a mix of patients with acute MIs, noncoronary chest pain, and dilated cardiomyopathy.
The aim of this study was to cross-relate quantifiable LV contraction (longitudinal strain [LS] and circumferential strain [CS]) with visually assessed LV function (WMS) and the extent of ischemia (MPI score) in the first week and 4 months after MI. More specifically, we aimed to answer the following questions: (1) What strain values correspond to echocardiography-based visual WMS, as well as radionuclide MPI scores, and what is the optimal threshold that differentiates between normal and abnormal visual assessment in patients with MI? (2) What is the interaction between the different strain components in the scenario of acute MI and their relationship to functional recovery?
Twenty consecutive patients with first acute ST-segment elevation MIs undergoing primary angioplasty were prospectively included in the study. The study was approved by the internal review board, and all patients provided written informed consent. The study was registered in the National Institutes of Health Clinical Trials Registry ( NCT00285064 ). Exclusion criteria were renal failure (creatinine > 1.3 mg/dL), known allergy to iodine, arrhythmia, and unstable clinical condition.
Invasive Coronary Angiography
This was performed on admission to identify the infarct-related artery, and successful angioplasty was performed in all patients, resulting in Thrombolysis in Myocardial Infarction grade 3 flow.
Transthoracic echocardiography was performed using either Acuson Sequoia (Siemens Medical Systems USA Inc., Mountain View, California) or Vivid 7 (GE Healthcare, Milwaukee, Wisconsin) echocardiographic systems. LV segmental and global function was assessed from parasternal long-axis and short-axis views and apical four-chamber, two-chamber, and three-chamber views. Studies were digitally stored in Digital Imaging and Communications in Medicine format for later analysis. Patients underwent three echocardiographic studies during follow-up. The first, a preliminary routine examination, was performed during the first 24 hours after admission, concentrating on 2D imaging of LV segmental function. This examination was performed mainly as a precaution so as not to miss any significant LV dysfunction that may have improved between the following examinations.
During days 3 to 5, a second transthoracic echocardiographic examination was performed with and without intravenous echocardiographic contrast injection (Definity; Lantheus Medical Imaging, North Billerica, Massachusetts). Contrast was used only to enhance endocardial border definition, and the contrast-enhanced images were used only for visual wall motion assessment and to enable accurate LV volume calculation. The non-contrast-enhanced images were used for strain analysis, because contrast may interfere with strain calculations.
Third, because most remodeling is known to occur in the first 3 to 4 months after acute MI, contrast and noncontrast transthoracic echocardiography was repeated at 4 months after admission to evaluate changes in segmental and global LV function and volumes.
For each examination, each of 16 segments was scored by a single observer, giving a score of 1 (normal), 2 (mildly hypokinetic), 2.5 (severely hypokinetic), or 3 (akinetic or dyskinetic). Functional recovery of a segment was defined as any segment having an abnormal score at baseline and a normal score at 4 months. The WMS, defined as the sum of scores for all segments, was calculated as a measure of overall segmental myocardial dysfunction. Segments were divided into three categories: an infarct zone, defined as segments with abnormal scores (2, 2.5, or 3); a border zone, defined as segments with scores of 1 but adjacent to infarct segments; and remote segments, defined as segments having no common border with an infarct segment. A second observer calculated the end-diastolic and end-systolic volumes and ejection fraction, using the biplane area-length method, for the second and third examinations only. Volumes were expressed as indices by normalizing for body surface area.
Rest MPI was performed 3 to 5 days after MI for the assessment of rest ischemia and viability. The studies were performed using a single-photon emission computed tomographic device with low-dose computed tomography for attenuation correction, (Millennium VG or Infinia & Hawkeye; GE Healthcare) using a rest-only single-isotope protocol. Imaging was performed 15 min and 24 hours after the intravenous injection of 3.5 mCi 201 Tl. Seventeen standardized myocardial segments were scored from 0 (zero readings) to 4 (severe defect), and the apical segment was not used in the segmental analysis, to correspond with myocardial mechanics data. A summed score of all segments (summed rest score) was calculated as representative of infarct size.
Measurements were performed using Velocity Vector Imaging (VVI) software version 2.5.1 (Siemens Medical Systems USA Inc.) from archived studies. Using VVI’s clip editor, QRS frames were marked and two or three consecutive cycles were selected for analysis on each view. All analyses were done on non-contrast-enhanced images. Endocardial contours were then traced and processed. Peak endocardial LS was measured from the apical four-chamber, two-chamber, and three-chamber views, in six segments per view. Peak endocardial CS was measured in three parasternal short-axis planes (base, mid, and apex) for six segments per short-axis section. Principal strain (PS) magnitude was calculated from its 2D measurable longitudinal and circumferential components as <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='PS=LS2+CS22′>PS=LS2+CS2−−−−−−−−−√2PS=LS2+CS22
PS = LS 2 + CS 2 2
, using the Pythagorean theorem, representing shortening along the axis of the myocardial fiber. Average global LS, CS, and PS curves were calculated from their respective segmental curves to define average global strain for each strain component. All strain measurements are reported by the magnitude of percentage shortening (absolute values). Segmental and global short-axis thickening (radial strain) was also measured.
Global strain components were compared with the summed score of the early and late radionuclide studies and with echocardiographic WMS and ejection fraction, using Pearson’s correlation coefficient.
Segmental strains were averaged for different WMS; for infarct, border, and remote zones; and for MPI segment scores. Because the strain software produced scores for 18 segments, we used 18 segments for the echocardiographic and MPI scores as well, by dividing the apical septal and inferior segments into two segments, with each resulting segment receiving the original score given to the parent segment. For each group of scores, analysis of variance was performed to test for differences between groups and thereafter t tests with Bonferroni’s correction for multiple comparisons. Paired t tests were applied to test for differences between the early and late echocardiographic studies. Nonpaired t tests were used to compare normal control to normal segments in our patient population. Results of 1,116 normal control segments were obtained from a previous study of normal strain (62 patients aged > 35 years; mean age, 52 ± 10 years; 55% men). These were healthy patients recruited consecutively during 2009. They were referred to our lab for 2D Doppler echocardiographic studies to evaluate minor symptoms or signs. To evaluate the ability of each strain component to differentiate between normal and abnormal WMS, receiver operating characteristic (ROC) curve analysis was performed, and the optimal threshold defined as the point at which sensitivity equaled specificity. ROC analysis was also performed on early strain values to test whether each strain component could differentiate between abnormal segments that recovered and those that failed to recover on the basis of visual assessment.
Statistical tests were considered significant at P < .05. Statistical analyses were performed using SPSS version 17.0 (SPSS, Inc., Chicago, IL). ROC curve analysis and comparisons were done using MedCalc version 12.5.5 (MedCalc Software, Mariakerke, Belgium).
The study population consisted of 20 patients (17 men; mean age, 54 ± 9 years). All patients had ST-segment elevation MIs at admission and underwent successful primary angioplasty within 12 hours of admission (most within 1 hour). The infarct-related artery was the left anterior descending coronary artery in 13, the left circumflex coronary artery in four, and the right coronary artery in three. Sixteen patients had one-vessel disease, three had two-vessel disease, and one had three-vessel disease (vessel stenoses > 50%). Two patients failed to undergo radionuclide imaging. One patient missed the echocardiographic examination on day 3, so the examination on day 1 was used instead. One patient died 1 month after the ST-segment elevation MI and thus did not undergo the third echocardiographic examination.
The examinations on day 1 and the early examination during the first week had very similar WMS (26.6 ± 6.4 vs 26.4 ± 6.7). There were also no significant differences in WMS (26.4 ± 6.7 vs 25.2 ± 7.7) or ejection fraction (52.3 ± 12.5% vs 54.1 ± 12.3%) between the early and late echocardiographic examinations.
Perfusion defects were detected in 15 of 18 patients, with summed segment scores ranging from 2 to 36. No significant differences were found between the immediate and the 24-hour redistribution scans, so the initial scan was used for all further analyses.
Segmental Strain: Relationship to Echocardiography and MPI Scores
For the sake of simplicity, we will henceforth relate to the absolute value or magnitude of all strain values. The magnitudes of segmental LS, CS, and PS decreased as visual segmental WMS increased ( Table 1 , Figure 1 ). Normal WMS had the highest strain values, which, however, were significantly lower than normal (control group). On the other side of the spectrum, visually akinetic segments (WMS = 3) demonstrated substantial shortening (mean PS, 15 ± 6%). Similarly, increasing WMS were accompanied by increasing MPI segment scores ( Table 2 , Figure 2 ). Remote and border zone segments were scored normal by visual inspection, yet strain and MPI scores differed significantly between the two zones. The remote zone had near normal strain and MPI scores, while the border zone had midway values between normal and infarct segments. Segments with normal MPI scores ( Table 3 ) also demonstrated abnormally low mean strain values. Mildly abnormal MPI segment scores (2) demonstrated lower CS, while LS was similar to normal MPI segments. Both LS and CS were significantly lower in segments with MPI scores of 3, without further change for MPI scores of 4. Radial strain showed the same trend of decrease from WMS = 1 to WMS = 3, but its variability was very high (31 ± 22% and 15 ± 15%, respectively, P < .05). Radial strain (thickening) correlated poorly with PS (shortening vector) ( r = 0.43 early, r = 0.28 late).
|( n = 172)||( n = 51)||( n = 14)||( n = 75)||( n = 1,116)|
|LS (%)||15.7 ± 5.8||11.1 ± 5.0 ∗||8.2 ± 4.1 ∗||7.7 ± 3.4 ∗ †||19.5 ± 4.3 ‡|
|CS (%)||23.5 ± 9.3||18.3 ± 9.9 ∗||13.8 ± 8.0 ∗ †||11.1 ± 6.3 ∗ †||27.8 ± 3.7 ‡|
|PS (%)||29 ± 8.9||22.2 ± 9.3 ∗||17.0 ± 7.3 ∗||14.1 ± 6.0 ∗ †||33.9 ± 5.7 ‡|
|MPI||0.3 ± 0.8||0.8 ± 1.1 ∗||2.1 ± 2.0 ∗ †||1.9 ± 1.9 ∗ †|
|LS (%)||17.0 ± 6.2 §||13.7 ± 6.0 ∗||13.2 ± 5.6 ∗ §||10.6 ± 6.4 † §||19.5 ± 4.3 ‡|
|CS (%)||23.0 ± 8.4||19.6 ± 7.8 ∗||13.8 ± 6.4 ∗ †||13.4 ± 7.6 ∗ † §||27.8 ± 3.7 ‡|
|PS (%)||29.2 ± 9.1||24.2 ± 8.1 ∗||20.3 ± 7.0 ∗ † §||17.8 ± 8.2 ∗ † §||33.9 ± 5.7 ‡|
|MPI||0.2 ± 0.8||0.4 ± 1.0 ∗||2.3 ± 1.9 ∗ †||1.8 ± 1.9 ∗ †|
|infarct||border||Remote||All normal||Normal controls|
|( n = 140)||( n = 92)||( n = 80)||( n = 172)||( n = 1,116)|
|LS (%)||9.0 ± 4.4||14.4 ± 5.5 ∗||17.1 ± 5.7 ∗ †||15.7 ± 5.8||19.5 ± 4.3 ‡ †|
|CS (%)||14.0 ± 8.6||21.7 ± 8.9 ∗||25.5 ± 9.3 ∗||23.5 ± 9.3||27.8 ± 3.7 ‡ †|
|PS (%)||17.3 ± 8.4||26.6 ± 8.7 ∗||31.7 ± 8.5 ∗ †||29.0 ± 8.9||33.9 ± 5.7 ‡ †|
|MPI||1.4 ± 1.8||0.3 ± 1.0 ∗||0.1 ± 0.6 ∗||0.3 ± 0.8|
|LS (%)||12.1 ± 6.3 §||14.9 ± 5.8 ∗||19.2 ± 5.9 ∗ † §||17.0 ± 6.2 §||19.5 ± 4.3 ‡ †|
|CS (%)||15.7 ± 8.1 §||21.8 ± 8.0 ∗||24.8 ± 8.7 ∗ †||23.0 ± 8.4||27.8 ± 3.7 ‡ †|
|PS (%)||20.4 ± 8.5 §||26.6 ± 8.2 ∗||32.1 ± 9.3 ∗ †||29.2 ± 9.1||33.9 ± 5.7 ‡ †|
|MPI||1.4 ± 1.8||0.3 ± 1.0||0.1 ± 0.6||0.2 ± 0.8|
|( n = 257)||( n = 7)||( n = 21)||( n = 39)||( n = 1,116)|
|LS (%)||13.8 ± 6.6||13.5 ± 4.6||10.4 ± 3.4 ∗||8.0 ± 4.2 ∗||19.5 ± 4.3 ‡|
|CS (%)||21.4 ± 10.0||15.1 ± 10.3 ∗||14.5 ± 8.0 ∗||11.6 ± 8.2 ∗||27.8 ± 3.7 ‡|
|PS (%)||26.1 ± 10.2||21.5 ± 9.4 ∗||18.3 ± 7.6 ∗||14.4 ± 7.5 ∗||33.9 ± 5.7 ‡|
|WMS||1.5 ± 0.7||1.0 ± 0.0 ∗||2.6 ± 0.7 ∗||2.6 ± 0.7 ∗||1.0 ± 0.0|
|LS (%)||15.4 ± 6.8 §||17.3 ± 5.7 §||13.5 ± 5.0 ∗ §||8.9 ± 5.4 ∗||19.5 ± 4.3 ‡|
|CS (%)||20.3 ± 8.9||22.4 ± 8.6||16.5 ± 5.1 ∗||13.1 ± 8.5 ∗ §||27.8 ± 3.7 ‡|
|PS (%)||26.1 ± 9.7||29.0 ± 7.8||21.5 ± 5.4 ∗ §||16.4 ± 9.0 ∗ §||33.9 ± 5.7 ‡|
|WMS||1.4 ± 0.7||1.0 ± 0.0 ∗||2.4 ± 0.7 ∗ †||2.6 ± 0.7 ∗ †||1.0 ± 0.0|