Heart failure (HF) is a growing and increasingly important chronic disease of the Western world, occurring in at least 2% of the adult population and rising to 3% in those aged over 75 years.1 It is characterised by inadequate systemic perfusion due to impairment of the cardiac pump function. Clinical HF is a progressive condition, typically with high morbidity and mortality rates. It therefore places a significant burden on healthcare resources. One of the keys to reducing the mortality, morbidity, and cost of HF is accurate and early diagnosis of left ventricular systolic dysfunction (LVSD).2 This is essential for successfully addressing underlying diseases or causes and in selection of appropriate therapies.
According to the recently released American College of Cardiology (ACC)/American Heart Association (AHA) guidelines for the diagnosis and management of HF, transthoracic echocardiography is 'The single most useful diagnostic test in the evaluation of patients with HF.'3 As therapeutic techniques such as cardiac resynchronisation therapy (CRT) gain clinical acceptance, new echo-cardiographic imaging modalities such as live 3-D imaging and specialised analysis tools for dynamically assessing regional volumes and differences in timing of contraction promise to play an expanding role in the management of HF. This article discusses these and other new imaging and analysis technologies in echocardiography that are directed at improving accuracy and reproducibility in HF assessment and monitoring.
Aetiology and Prognosis in HF
The most common cause of HF is LVSD (approximately 60% of patients). In this category, most cases are a result of end-stage coronary artery disease (CAD), either with a history of myocardial infarction (MI), chronically underperfused yet viable myocardium or a combination of the two.4 Other disease processes that can lead to HF include valvular heart disease, congenital heart disease, diabetes, hypertensive disease and idiopathic and toxic (e.g. alcohol-induced) cardiomyopathies. Less common causes include viral infections of the heart muscle, thyroid disorders, disorders of the heart rhythm, or a combinations of these. HF associated with LVSD is characterised by progressive structural change in the LV known as remodelling. While the disease progresses, myocyte hypertrophy and elongation gives rise to LV dilatation and hypertrophy. In this situation, stroke volume is increased without an actual increase in ejection fraction (EF). This results in increased wall tension impaired subendocardial myocardial perfusion, and may provoke ischaemia. While this dilatation progresses, separation of the valve leaflets can lead to mitral and tricuspid regurgitation. This may further diminish the cardiac output and increase end-systolic volumes and ventricular wall stress, therefore leading to further dilation, pulmonary congestion and myocardial dysfunction.
LV volumes and EF are therefore important prognostic indicators for morbidity and mortality in HF patients.5
Echocardiography in the Diagnosis and Monitoring of HF
Echocardiographic Analysis of EF
Measurement of EF typically uses manual planimetry of 2-D areas according to Simpson's single plane or biplane method of disks. This method can be time-consuming and has been shown to exhibit inaccuracies when compared with the gold standard of magnetic resonance imaging (MRI).6 Chief sources of inter- and intra-observer variability derive from inconsistent acquisition methods or image plane selection and subjective boundary definition by readers. Variable image quality, particularly in technically difficult patients, may further exacerbate many of these inaccuracies. A further important source of error lies in the geometric assumptions used in the calculation of volumes from 2-D data.
Several new imaging technologies have recently emerged that can significantly improve the speed, reproducibility and accuracy of EF measurements. These can be divided into four categories based on the benefits they offer and are summarised in Table 1.
PureWave Transducer Technology, Image Quality and Harmonic Imaging
Calculation of EF in obese patients or those with poor acoustic windows has traditionally been a problem for echocardiography. Recent advances in system bandwidth and beamforming have allowed technologies such as tissue harmonic imaging and left ventricular opacification (LVO) contrast imaging to make a significant contribution by suppressing artefacts and sharpening endocardial border delineation for improved accuracy and reproducibility of EF assessments.7,8 However, for several decades, no significant innovations have occurred at the level of the most basic determinant of ultrasound image quality: the transducer elements or crystals that convert electrical into acoustic energy and vice versa. Using the Philips iE33 system, a new transducer design based on PureWave mono-crystal technology is able to yield 80% gains in electromechanical efficiency. This has paved the way for a new generation of broadband transducers that offer superior 2-D imaging resolution, more sensitive contrast harmonics, improved colour flow sensitivity, and superior low-frequency tissue harmonics. All of these can significantly improve endocardial border delineation, and therefore provide enhanced accuracy in EF measurements, particularly when addressing technically difficult patients (see Figure 1).
2-DQ Semi-automatic EF and Volume Analysis
Acoustic quantification (AQ) and colour kinesis (CK) provide a promising approach for realtime tracking of the endocardial border for semi-automated EF and volumetric measurements. CK provides a qualitative display of wall displacement in colour-coded time intervals to reduce subjectivity in regional wall motion assessment (see Figure 1), while AQ generates dynamic volume data leading to improved reproducibility and rapidity of EF and LV volume determination. Advances in image quality as well as improvement of the original AQ algorithms for border tracking promise to increase the sensitivity and specificity of AQ.9,10 In Philips QLAB software, which is available both on-cart and as an off-cart tool, it is possible to select either a complex or smooth border for endocardial tracking. The latter method uses a sophisticated multi-factorial algorithm for discriminating tissue from blood pool and offers tissue tracking tools that interrogate the mitral or tricuspid valve annulus for CK assessment of A-V plane motion and long axis shortening (see Figure 2).
Matrix Array for Live xPlane and Live 3-D Imaging
Fully sampled matrix-array transducers have paved the way for both full 3-D acquisition and simultaneous acquisition and control of separate imaging planes. The availability of up to 3,000 elements for simultaneous transmission means that separate, steerable images offering high detail and contrast resolution may be acquired and manipulated in realtime. These images may be independently angled, rotated or tilted to provide greater control and faster access to correctly aligned planes for measurement. Using this method, the problem of incorrect alignment and foreshortening may be greatly reduced while achieving considerable workflow benefits. 2-DQ analysis software includes support for bi-plane acquisitions allowing optimal workflow for EF calculation according to SimpsonÔÇÖs bi-plane method (see Figure 3). Live xPlane in stress echo has obvious and important advantages for workflow and reproducibility.
Live 3-D Image Analysis
The image acquisition and analysis tools reviewed so far provide the promise of improved inter- and intra-observer variability but do not address the important issue of geometric assumptions needed for volume estimation from 2-D image measurements. Second-generation matrix transducers allow realtime acquisition of high-resolution 3-D volumes for live anatomic interrogation of cardiac anatomy and near realtime acquisition of fully sampled LV datasets for functional analysis of global and regional volume data.
3-D Assisted EF Calculation
3-D datasets consisting of the full LV volume allow accurate retrospective selection of measurement planes for calculation of stroke volumes and EF. Traditional problems inherent in the 2-D method, such as incorrect selection of true four- and two-chamber views or foreshortening of chambers, can now be completely obviated by a single volume acquisition. 3DQ also allows analysis of LV mass without the assumptions inherent in the Teichholz approach (see Figure 4). This has important potential workflow benefits; instead of time spent at the time of the scan striving to obtain correct 2-D image planes, the operator may now acquire a single volume that may be analysed on-cart or off-cart using 3DQ multi-planar reconstruction to provide access to any image planes.
3DQ Advanced True Volume Analysis
The 3DQ Advanced analysis tool provides an assumption-free alternative to estimation of LV volumes from 2-D slices11 (see Figure 5). Following a selection of five reference points in end-diastole and end-systole, semi-automated 3-D border detection allows fast surface rendering of the complete voxel data in a full 3-D volume. Before 3DQ Advanced, 3-D volume analysis was typically performed by reconstructing volumes from 'sparse' views or multiple 2-D slices taken from the 3-D volume-rendered dataset. This approach is less computationally intensive than full surface rendering methods and has been preferred due to speed and cost considerations. However, thanks to advances in parallel processing, 3DQ Advanced allows rapid generation of a full 3-D wire-mesh endocardial volume with minimal operator intervention.
Border detection of 3-D datasets presents unique challenges compared with 2-D methods. The algorithm used in 3DQ Advanced applies physics-based modelling plus 3-D pattern matching, which tracks the mitral annulus and apex over time to provide an 'active object' motion presentation of the dynamic 3-D shape. This allows 3-D borders for the endocardial space in each frame to be combined into a smooth beating volume with highly accurate spatial and temporal motion detail.
One of the main drawbacks of motion detection analysis techniques is the problem of overall cardiac translational motion. In 3DQ Advanced, the use of two floating centroids allows effective compensation for translational motion by subtracting the velocity of each floating centroid from the estimated overall velocity.
Echocardiography in the Diagnosis and Monitoring of CHF
In addition to analysis of LV volumes and EF, echocardiography provides insight into numerous other key parameters that are used for differential diagnosis and monitoring in HF.
Identifying the correct aetiology may depend on any of the following details, all of which are highly dependent on image quality:
- confirmation of regional wall motion abnormalities;
- confirmation of LV aneurysm;
- confirmation of LV hypertrophy or dilated cardiomyopathy (accurate determination of LV wall thickness and LV mass);
- detection of pericardial constriction or tamponade; and
- detection of thrombi.
If systolic dysfunction is present, echocardiographic confirmation of regional wall motion abnormalities or LV aneurysm will suggest an ischaemic basis for HF, whereas global dysfunction suggests a non-ischaemic aetiology. Echocardiography is also helpful in determining other aetiologies such as valvular heart disease, cardiac tamponade and pericardial constriction, and provides useful clues about infiltrative and restrictive cardiomyopathies.
In patients with CAD and chronic LV dysfunction, it is crucial to distinguish between viable and fibrotic tissue to make adequate clinical decisions. Non-contractile but viable myocardium may correspond to different states that are important but difficult to distinguish, i.e. ischaemia, stunning, non-transmural infarction, or hibernation and in individual patients these pictures may co-exist.12 In this regard pharmacologic or exercise stress testing can play a key role in evaluating patients with HF. Stress echocardiography is also valuable for evaluating the behaviour of functional mitral regurgitation.
Several complementary analysis tools exist for the assessment of regional function. These include CK, tissue Doppler imaging (TDI) with strain and strain rate analysis, 3-D regional volumes and 3-D multi-planar reconstruction with iSlice view. iSlice imaging allows interrogation of a 3-D volume in nine serial slices. This is an intuitive qualitative method based on an 'exploded bullseye' view for comparative assessment of regional function (see Figure 6).
To monitor the effect of therapies in HF, it is important to identify echocardiographic indicators of reverse remodelling. Apart from the reduction of LV volume, other useful parameters include the sphericity index, which can be most effectively evaluated using 3-D datasets.
CRT and HF
Numerous clinical trials have shown the benefit of cardiac resynchronisation therapy using multi-site, biventricular pacing in patients with severe symptomatic HF and a wide QRS complex. There is convincing evidence that biventricular pacing decreases LV volume (reverse remodelling), increases the LVEF, decreases mitral regurgitation and improves symptoms caused by HF.5-10, 13-18 However, as many as one-third of patients with left bundle branch block (LBBB) or prolonged QRS fail to respond to CRT. This may be due to the fact that many of these patients do not exhibit systolic intra-ventricular mechanical dyssynchrony. New research has shown that mechanical evidence of intraventricular dyssynchrony provides an effective independent predictor of patient response.5 This mechanical evidence can be readily obtained using 3-D regional volume curves or regional tissue Doppler velocity (TDV) curves to derive an index of dyssynchrony (SD index).
Time to peak velocity from multiple regional segments can be measured with QLab to generate TDI-based SD indices according to the methodology of Yu et al.9 Additionally, true regional volume versus time data may be generated to support SD index analysis based on downstream mechanical/functional criteria.10
TDI Approach
Time to peak systolic velocity (tPSV) is measured from onset of QRS to the origin or peak of S2 wave.9 Intraventricular dyssynchrony may be inferred from:
- the maximum difference between two regions (e.g. basal septum and basal lateral wall); and
- standard deviation (SD) of 12 segments (basal and mid segments in both walls in each of three apical views).
QLAB SQ analysis software provides the possibility of defining a multi-region, curved m-line (using user-definable sub-regions) for fast comparison between tPSV in four or more segments per apical view to provide 12 or more segments in total (see Figure 7).
3-D Approach
The 3-D approach is based on the time to minimum systolic volume (tMSV) measured from onset of QRS to minimum value of regional time-volume curves.10 Intraventricular dyssynchrony may be inferred from:
- the maximum difference between two regions (e.g. basal septum and basal lateral wall); and
- SD of up to 17 regional volumes.
The tMSV approach has several potential advantages:
- ease of use and reproducibility - five clicks for end-diastole and end-systole generates full volume data with regional time-volume curves;
- better correlation with clinical end-points - downstream functional data for all segments rather than sampled vectors representing only one component of mechanical contraction;
- ease of interpretation:
- tMSV points are marked and curves are normalised against end-diastolic values to allow rapid qualitative assessment;
- the report page summarises SD index (standard deviation) both as absolute time and R-R ratio for six, 12 or 16 segments or based on a user-selected choice of segments - maximum difference between any segments is also displayed; and - it is more robust - one of the main drawbacks of motion detection analysis techniques is the problem of overall cardiac translational motion and reproducibility.
Conclusion
Transthoracic echocardiography provides vital information for the diagnosis, prognosis and choice of therapies in HF. Reproducible, accurate echocardiographic data relies on high-quality images as well as advanced analysis tools designed to improve accuracy, reduce operator dependence and remove geometric assumptions.
With the growing acceptance of CRT as a therapeutic option, new echocardiographic imaging modalities such as live 3-D imaging and specialised analysis tools for dynamically assessing regional volumes and differences in timing of contraction promise to play an expanding role in the management of HF.