Author + information
- Published online December 13, 2017.
- aDepartment of Cardiovascular Sciences, Cardiovascular imaging and dynamics, University Hospitals Leuven, KU Leuven, Leuven, Belgium
- bDZHK Centre for Cardiovascular Imaging, Institute for Experimental and Translational Cardiovascular Imaging, University Hospital Frankfurt, Frankfurt am Main, Germany
- ↵∗Address for correspondence:
Dr. Frank Rademakers, Department of Cardiovascular Sciences, University Hospitals Leuven, KU Leuven, Herestraat 49, Leuven, 300, Belgium.
Strain imaging in echocardiography plays an increasingly strong role for diagnoses and risk assessment in a wide spectrum of patients, ranging from low risk (1) to post-myocardial infarction (2). More recently, strain imaging based on standard cardiac magnetic resonance (CMR) images has become possible for tracking mainly endocardial features of the myocardium (feature tracking) (3).
In this issue of iJACC, a study by Gavara et al. (4) assessed the predictive value of various clinical and CMR parameters including strain based on CMR feature tracking in 323 patients referred for CMR after a first ST-segment elevation myocardial infarction (STEMI) and treated with percutaneous coronary intervention, using standard cine imaging and late gadolinium enhancement. Of the original cohort of 542 patients, 98 patients had to be excluded because feature tracking was not possible in these previously obtained images. The authors measured end-diastolic and end-systolic volumes and ejection fraction, as well as infarct size, presence and extent of microvascular obstruction, myocardial edema, and salvage index. Using offline software for feature tracking, the authors also derived segmental and global circumferential, longitudinal, and radial strain values. In a separate, normal control group, cutoff values for these strains were determined.
The validity of the association between strains and outcomes from this retrospective study was ascertained in a prospective cohort of 180 patients who had the same characteristics.
After a median follow-up of 1,085 days, the authors evaluated a composite endpoint, including cardiac death and hospitalization for heart failure or reinfarction. In univariate analysis, lower strains displayed a higher risk for major adverse cardiac events in the entire cohort and in those with a TIMI (Thrombolysis In Myocardial Infarction) flow grade of 3; this was true for the separate components of outcome as well. In a multivariable analysis, time to reperfusion, TIMI risk score, and global longitudinal strain were correlated with outcome. A negative correlation between global longitudinal strain and ejection fraction was noted. In the external validation cohort, the correlation between outcome and longitudinal strain was maintained, but adding longitudinal strain to the multivariate model did not increase reclassification.
Thus, the main finding of the study is that myocardial strain, as derived from CMR feature tracking, contributes significant prognostic information for stratifying risk soon after STEMI. However, it does not substantially improve the risk reclassification of patients compared with the information provided by the baseline characteristics of patients and conventional CMR indices.
It is fascinating to see how we are increasingly able to address the various underlying pathophysiological characteristics in patients after myocardial infarction, ranging from volumes and function to tissue characterization including edema, hemorrhage, no-reflow, and scar tissue. Recently, visualization of infarcted tissue and its components has become an important parameter for predicting outcome in these patients, with clear superiority to volumes and function (5). Will the ability to assess strain reverse this parameter (6) and bring function back into the game?
First, the beauty of strain imaging lies in the ability to measure it in CMR without the need for contrast agents, using standard imaging sequences. This promises to reduce cost, effort, and scan time for CMR examinations. Potentially, the multivariable model proposed in the current work (strain, time to reperfusion, and TIMI risk score) could even be obtained by echocardiography rather than CMR. Although Gavara et al. (4) have demonstrated a strong correlation between strain and outcome, the results are not superior to “classic” CMR-based models. Future studies need to address the question whether the information obtained by echocardiography, including strain and a CMR examination that includes strain but not contrast injection, or a full CMR examination, are similarly suited to predict outcome.
Second, it is important to remember that our ability to improve patient outcome is not based on prediction of prognosis alone but rather on using test results to influence decision making and thus patient management. It remains to be determined whether that is the case for CMR-derived strains (similar to echocardiographic strains): that is, does longitudinal strain provide additional insights which could influence the decisions made about a patient after a first STEMI and improve the components of the outcome studied (i.e., cardiac death or hospitalization for heart failure and re-infarction)?
Third, the most intriguing question is why global longitudinal strain outperforms other strains for prognosis. Does longitudinal strain extract a pathophysiological component not available from other strains? Echocardiography has shown the predictive value of feature-derived strains in many clinical situations, including ischemic heart disease (1). As with echocardiography, longitudinal strain in this CMR study is clearly a better predictor than circumferential or radial strain. Why is this? Is there an intrinsic difference between these strains? Neither longitudinal strain nor circumferential strains are direct measurements of fiber strain. The deformation of the myocardial wall, as defined by the intricate interplay between the contracting fibers oriented at different angles, can only be fully evaluated by obtaining all 6 strains, the 3 normal strains (radial, circumferential, and longitudinal) and the 3 shear strains (1 of them a circumferential-longitudinal strain or torsion). One strain can only partially represent the whole picture. Actually, mid-wall circumferential strain is the strain coming closest to fiber strain as fibers run mainly circumferentially at the mid-myocardium. It has been argued that longitudinal strain is better suited to evaluate ischemia or partial infarction because longitudinal fibers and ischemic damage are more prevalent at the endocardium. This reasoning is flawed as all but the circumferential fibers are equally distributed across the wall and have an oblique angle, thus contributing to circumferential as well as longitudinal deformation. By the conservation of mass and the morphology of the ventricle, this leads to wall thickening; no fiber contraction directly causes thickening; if anything thickening is counteracted by the imbrication of contracting fibers through the wall.
Why then is longitudinal strain a better predictor? Longitudinal strain has the advantage of being fairly constant across the myocardial wall whereas circumferential strain varies from approximately 5% to 10% epicardially to more than 40% endocardially. This makes global circumferential strain very dependent on epicardial and endocardial contouring, as well as the method used for tracking. Although CMR-based feature tracking mainly follows the endocardial features despite the excellent depiction of the epicardial wall, echocardiography is better suited for tracking speckles within the myocardium. Similarly, radial strain (going from <10% epicardially to >60% endocardially) is strongly influenced by endocardial and epicardial contours. As longitudinal strain is more homogeneous across the wall, identification of the exact epicardial and endocardial borders is less relevant, resulting in higher accuracy, reproducibility, and consequently, prognostic value than other strains.
Thus, although longitudinal strain holds the lead with respect to predictive value for both echocardiographic and CMR-derived strains, with either feature or speckle tracking, this could be due to technical reasons rather than an intrinsic characteristic of this strain. Improving our capability to accurately measure circumferential and radial strain is thus needed before we can compare their pathophysiological value to longitudinal strain.
Finally, it remains to be shown if strain per se holds more information than volumes and ejection fraction or whether the observed advantages are similarly due to a more reproducible measurement technique. Once this has been ascertained, the next step is a systematic comparison of the relative contribution of function (measured with the best possible technique), characteristics of the infarcted myocardium, as well as characteristics of the remote myocardium including remaining myocardial ischemia and/or hibernating myocardium. If we were able to separate subgroups of patients in whom these relative contributions had a different importance, we would be able to stratify patients by risk and provide them with individualized optimal therapy.
The authors have made a significant contribution toward guiding clinicians to a better use of the technical arsenal at our disposal to treat patients after a first STEMI, but we need more insight into the technical issues related to the black boxes of feature and speckle tracking to really understand and interpret their relative importance. In addition, we need to show that this knowledge is not only of scientific interest but also relevant to patient management and outcome.
↵∗ Editorials published in JACC: Cardiovascular Imaging reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Imaging or the American College of Cardiology.
Dr. Nagel has received support from Medis, TomTec, CVI42, and NeoSoft. Dr. Rademakers has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation
- Biering-Sørensen T.,
- Biering-Sørensen S.R.,
- Olsen F.J.,
- Sengeløv M.,
- Jørgensen P.G.,
- Mogelvang R.,
- et al.
- Haugaa K.H.,
- Grenne B.L.,
- Eek C.H.,
- Ersbøll M.,
- Valeur N.,
- Svendsen J.H.,
- et al.
- Claus P.,
- Omar A.M.S.,
- Pedrizzetti G.,
- Sengupta P.P.,
- Nagel E.
- Gavara J.,
- Rodriguez-Palomares J.F.,
- Valente F.,
- et al.
- Aidi El H.,
- Adams A.,
- Moons K.G.M.,
- Ruijter Den H.M.,
- Mali W.P.T.M.,
- Doevendans P.A.,
- et al.
- Mangion K.,
- McComb C.,
- Auger D.A.,
- Epstein F.H.,
- Berry C.