Cardiovascular disease is one of the leading causes of death in the Western World. In the Netherlands 20.000 people die each year of an acute myocardial infarction or stroke. Two-dimensional ultrasound has been the most widely available imaging modality to evaluate cardiovascular structure and function. Improvements in two-dimensional ultrasound image resolution have led to more detailed imaging of the myocardium, heart valves, vessel walls, and atherosclerotic plaques. Doppler techniques are readily available to estimate the blood flow and pressure gradients across heart valves and vascular stenoses. However, two-dimensional B-mode images have a limited reproducibility in three-dimensional structures and do not (yet) allow for detailed tissue characterization.
Specifically, in cardiac function analysis, the standard volumetric biplane measurement of ‘ejection fraction’ to determine systolic function has an inter-rater and longitudinal variability of 5-10%, due to apical foreshortening, suboptimal endocardial contouring and changes in pre- and afterload. Furthermore, it is a rather late marker of systolic dysfunction, knowing that the complex orientation of cardiac muscle fibers allows for compensation of ejection fraction when certain fibers become dysfunctional.
In cardiac ultrasound, two-dimensional myocardial deformation (strain) imaging, by means of tissue doppler imaging and now speckle tracking, has found its way into the clinic. Global longitudinal strain (i.e. shortening of the total muscle length in longitudinal direction) has been shown to aid in the prediction of cardiac events and mortality in various cardiac diseases (heart failure, valve disease, postoperative), especially when ejection fraction is still preserved. Furthermore, it is possible to assess regional wall motion abnormalities (lower strain, pre-stretch, post-systolic strain) in more detail than with the naked eye, provided that there are no image artefacts in these regions. Guidelines in adult cardio-oncology have adopted a reduced global longitudinal strain as an early sign of cardiotoxicity and specific strain patterns point to specific myocardial diseases (e.g. apical sparing in cardiac amyloidosis). It furthermore has high potential to aid in the indication for cardic resynchronization therapy by assessment of dyssynchronous patterns. However, the inclusion of myocardial strain analyses in systolic function quantification guidelines has been hampered by the lack of uniform lower limits of normal. These still depend on the software vendor, and furthermore on the myocardial layer in which strain is assessed and the timing of events. Attempts to resolve these problems are undertaken by a dedicated standardization task force.
For cardiac imaging, our current research aims to identify additional clinical applications of deformation imaging and deeper analysis of its current applications. The fields of research include pediatric cardio-oncology and congenital heart disease.
In the past, we have focussed on the development of strain estimation techniques in two-dimensional and three-dimensional echocardiography data. Currently, we provide an infrastructure with different software packages and technical support to clinicians and clinical studies that involve two-dimensional myocardial strain imaging. Cohort and case-control studies are hosted that compare the hearts of childhood cancer survivors to controls and that assess the hearts of patients with Ebstein anomaly in further detail. Furthermore, we undertake continuous efforts to provide lab- and vendor specific normative strain values.
|example of regional (coloured) and average (white) longitudinal strain curves in an apical 4-chamber view in a childhood cancer survivor, obtained in commercially available, vendor-independent strain software.|
We have developed 2D and 3D strain estimation and segmentation strategies to improve improve accuracy with respect to commercially available techniques. These strategies are based on analysis of the raw ultrasound (rf) signals: rf-based strain estimation demonstrated to be more accurate than B-mode based commercial techniques (Lopata et al 2009, Lopata et al 2009a, Lopata et al 2009b) and these techniques were extended from 2D to biplane strain imaging (Lopata et al 2010) and full 3D (Lopata et al 2011). Automated segementation based on echo statistics (Nillesen et al 2008)) and decorrelation (Nillesen et al 2009) were developed and validated in vivo (Nillesen et al 2009a) and finally evaluated in children with congenital heart disease (Nillesen et al 2011).
Furthermore, clinical application of strain and strain rate with tissue Doppler (Kapusta et al., 2000a, 2000, 2001,) and speckle tracking in childhood cancer survivors (Mavinkurve-Groothuis et al., 2010, 2012; Pourier et al., 2020; Merkx et al., 2021a, 2021b) ) has advanced this field of subclincial detection of cardiac dysfunction.
Most recently, we described our 15-year experience with measuring myocardial strain in the surveillance of childhood cancer survivors and its potential value over LVEF predicting 10-year cardiac events (Pourier et al., 2022).
Also visiting clinical PhD candidates: Lianne Geerdink, Milanthy Pourier.