Future applications of strain imaging

Strain imaging as a biomarker for clinical trials

One of the imminent applications of strain imaging is its use as a surrogate endpoint for clinical trials. Although hard endpoints, such as mortality, are considered important in medical research, observing them requires a large sample size and a long follow-up. Traditionally, LVEF has been considered as the representative surrogate endpoint for reflecting change in cardiac systolic function. However, the reliability of GLS has grown with evidence of reproducibility when observers are properly trained, and clinical trials have begun employing strain as a reproducible and objective endpoint. For example, Ikonomidis et al studied the effect of interleukin-12 inhibition on cardiac function in patients with psoriasis using strain parameters as the primary outcome. They demonstrated that the improvement of GLS was significantly greater in the anti-interleukin drug arm compared with the tumor necrosis factor-alpha and cyclosporine arm. As such, many large cohort studies, such as Copenhagen Heart Study and ARIC Study, use speckle-tracking echocardiography for analysis of LV systolic function. ^, Interestingly, LVEF failed to show changes in systolic function in any arm. The Strain sUrveillance of Chemotherapy for improving Cardiovascular Outcomes (SUCCOUR) study is another interesting and unique study in which strain echocardiography is being compared in efficacy to conventional echocardiographic parameters. In the study, older adult patients who are undergoing chemotherapy using anthracyclines have been randomized to either a GLS-guided or an LVEF-guided arm ( Fig. 10.3 ). In the GLS-guided arm, a relative red uction of GLS by 12% in any follow-up echocardiography (3, 6, 9, and 12 months) is being considered as a threshold for initiating cardioprotective therapy, whereas protective therapy in the LVEF-guided arm is started in response to a conventional LVEF cutoff (a symptomatic drop of >5% of two-dimensional [2D] LVEF, or >10% asymptomatic drop to 2D LVEF <55%). Of 307 patients who were followed up over 1 year, significantly fewer patients developed CTRCD in the GLS-guided arm (9 [5.8%] vs 21 [13.7%], P = .022), although the primary endpoint of change in LVEF was similar (−3.0% vs −2.7%, P = .69).

Thus the ability of strain to identify subtle systolic dysfunction that cannot be appreciated by LVEF is driving its introduction into the mainstream in evaluation of systolic dysfunction especially in early stages of cardiovascular diseases. Collective evidence from randomized clinical trials may further help with formulation of recommendations for the use of strain in clinical practice guidelines. Furthermore, Aguilar and colleagues reported “archeological” strain echocardiography, which indicates assessing strain parameters from old-fashioned analogue archival echocardiographic images by digitalizing them. This technology may open up unique and interesting opportunities to use historical echocardiographic images to study the impact of subclinical LV dysfunction on outcome.

Three-dimensional strain imaging

Although two-dimensional (2D) strain imaging has been shown as a useful tool in clinical studies, there are several inherent limitations. First, the heart is a three-dimensional (3D) structure, and the direction of its movement (contraction and relaxation) occurs not only in the direction of ultrasound beam but also in the direction of the fiber orientation. Consequently, some portion of the myocardium also moves perpendicular to the ultrasound beam (through-plane phenomenon), and 2D speckle tracking may not be tracking the same myocardium throughout the cardiac cycle. Second, 2D strain imaging provides polar map (bull’s-eye map) of the strain values that is extrapolated to represent all segments of the heart. However, it is a reconstruction from only three standard views (apical, short axis, and parasternal views), which may overlook subtle regional abnormalities in areas that are not appreciated in these views. In addition, the orientation and positioning of the probe in acquisition of these three views can vary depending on the individual morphology of the heart. The heartbeat used for each of these three views is also different. These limitations can be overcome by using strain imaging from 3D echocardiography, which theoretically should be a better approach to study mechanical function of the heart. Although 3D echocardiography has had several limitations such as lower image quality and lower frame rates compared with 2D echocardiography, state-of-the-art technologies are making these limitations less significant. The latest 3D applications can scan the entire left ventricle over a single cardiac beat with good image quality with a volume rate of around 30/second, which is analogous to the temporal resolution of DICOM images that are often used for strain in postprocessing. Steps for standardizing the use of 3D strain and further head-to-head comparison with 2D strain in real-world settings will potentially allow verifying its applicability in clinical practice.