Heart valve disease: investigation by cardiovascular magnetic resonance
© Myerson; licensee BioMed Central Ltd. 2012
Received: 11 January 2012
Accepted: 19 January 2012
Published: 19 January 2012
Skip to main content
© Myerson; licensee BioMed Central Ltd. 2012
Received: 11 January 2012
Accepted: 19 January 2012
Published: 19 January 2012
Cardiovascular magnetic resonance (CMR) has become a valuable investigative tool in many areas of cardiac medicine. Its value in heart valve disease is less well appreciated however, particularly as echocardiography is a powerful and widely available technique in valve disease. This review highlights the added value that CMR can bring in valve disease, complementing echocardiography in many areas, but it has also become the first-line investigation in some, such as pulmonary valve disease and assessing the right ventricle. CMR has many advantages, including the ability to image in any plane, which allows full visualisation of valves and their inflow/outflow tracts, direct measurement of valve area (particularly for stenotic valves), and characterisation of the associated great vessel anatomy (e.g. the aortic root and arch in aortic valve disease). A particular strength is the ability to quantify flow, which allows accurate measurement of regurgitation, cardiac shunt volumes/ratios and differential flow volumes (e.g. left and right pulmonary arteries). Quantification of ventricular volumes and mass is vital for determining the impact of valve disease on the heart, and CMR is the 'Gold standard' for this. Limitations of the technique include partial volume effects due to image slice thickness, and a low ability to identify small, highly mobile objects (such as vegetations) due to the need to acquire images over several cardiac cycles. The review examines the advantages and disadvantages of each imaging aspect in detail, and considers how CMR can be used optimally for each valve lesion.
Cardiovascular magnetic resonance (CMR) has unique capabilities which can greatly benefit the assessment of the patient with cardiac valve disease. While echocardiography (echo) remains the major imaging modality for assessing valve disease, there are many areas where CMR provides 'added value' to existing assessment and can complement the echo assessment. CMR can also provide a comprehensive 'stand-alone' assessment in some situations, delivering optimal assessment of patients using a combination of several techniques. These include quantifying the severity of the valve lesion, determining aetiology, examining the consequences for the relevant ventricle, and assessment of the surrounding anatomy (e.g. aortic root). Additional information on great vessel anatomy and the presence of myocardial scar (infarction) can also be clinically useful. The modality is used best by harnessing the advantages it brings, rather than attempting to replicate echocardiography or x-ray computed tomography (CT). This review will highlight the optimal use of CMR in valve disease, highlighting the strengths of the technique and also the potential pitfalls when assessing patients with valve disease.
Accurate measurement of left and right ventricular volumes, function and mass are vital for assessing the impact of valve lesions on the ventricles. Excessive dilation or reduced ventricular function are strong indicators of a poor prognosis , and reliable measurement is important. CMR is the most accurate and reproducible technique for assessing both left and right ventricular volumes & mass [7–9], and newer steady-state free-precession sequences appear to be even more accurate than older gradient echo cine sequences [10, 11]. RV volumes are particularly useful as these are difficult to achieve by other methods, though accurate measurement is more difficult than for LV volumes. The role of left ventricular (LV) mass in valve disease has not been studied as extensively as volume, possibly due to the inaccuracies of measurement by M-mode or 2-dimensional echo , and LV mass may become a useful measure in the future, particularly for patients with aortic stenosis. Reproducibility is important for serial assessment of ventricular size, as patients with valve lesions are often monitored for many years if asymptomatic before symptoms and/or ventricular deterioration occur. CMR is highly reproducible , and being a 3-dimensional technique, is more sensitive to changes than one or two-dimensional LV diameters . CMR is also less prone than echocardiographic diameters to variations in measurement position, which can occur, despite standard guidelines . The accuracy of CMR is, however, dependent on correct placement of the basal ventricular image slice, and careful contour placement during post-processing is crucial, with correct differentiation of atrial and ventricular chambers, especially for the right ventricle. Significant error can occur if the basal slice is incorrectly included/excluded from ventricular volumes. Post-processing software that includes long axis visualisation of the valves to ensure appropriate slice inclusion significantly aids accuracy. CMR-derived ventricular stroke volumes can also be used to quantify mitral and tricuspid regurgitation, either in conjunction with flow measurement or with volume data alone if the valve regurgitation is isolated  - see below, but the issues about accuracy of contour placement still apply.
In addition to flow quantification, recently developed 3-dimensional in-plane CMR flow sequences can measure velocities in 3 dimensions simultaneously [27, 28], allowing the visualisation of complex flow patterns. Future work may examine the utility of this technique for valve lesions and other areas of clinical utility.
Despite the limitations, CMR measurement of velocity is advantageous in angulated roots where correct echo beam alignment with the stenotic jet is difficult. In addition, many stenotic jets are not parallel to the LV outflow tract, and are also inaccurately assessed with echo, even when outflow tract visualisation is good. In-plane velocity mapping in the outflow tract is useful to identify the location of maximal velocity - this is usually just distal to the valve tips in valvar stenosis. This can be followed with through plane velocity mapping in a plane perpendicular to the direction of flow, positioned at the identified location of maximal velocity (Figure 1). This combination reduces partial volume effects while ensuring the peak velocity is measured, and mean velocity can also be assessed from the through-plane velocity measurements. Ensuring the correct slice position for flow measurement is important for accuracy; and although this image appears similar to that through the valve tips, the position of maximal velocity (the vena contracta) usually lies a few millimetres distal to the valve tips, so the images may not be in identical locations.
Other advantages in aortic stenosis include the ability to differentiate sub-valvar and supra-valvar stenosis, which are easily visualised with CMR cine imaging, and the site of velocity acceleration can be accurately located with in-plane velocity mapping. CMR can also accurately assess the ascending aorta, which may be dilated, particularly with bicuspid aortic valves, and may alter surgical management either at the time of valve replacement or by indicating that root ± valve replacement are required due to excess aortic dilation.
Highly accurate LV mass measurement provides a more precise and sensitive measure of the effect of aortic stenosis on the left ventricle than measuring myocardial wall thickness. LV mass has been a poorly examined parameter in aortic stenosis, likely due to the inaccuracies of echocardiographic M-mode measurement, and CMR-derived LV mass may prove to be a useful tool but needs further examination. Late gadolinium enhancement imaging in patients with aortic stenosis has shown patchy mid-wall enhancement in up to a third of patients with severe stenosis, usually in conjunction with significant LV hypertrophy, and often in the basal lateral wall . This likely reflects focal areas of fibrosis, which have also been shown in autopsy studies . Early reports have shown that this is associated with a worse prognosis . Future CMR techniques may include the assessment of diffuse fibrosis using T1 mapping and other CMR techniques, and this is likely to be an exciting area of development.
The accuracy of aortic regurgitation quantification using CMR through-plane velocity mapping is excellent when compared to in-vitro studies  or in-vivo CMR measurement using the difference between ventricular volumes , and it correlates well with angiographic or echocardiographic grades of severity [18, 20, 40]. As it remains the only technique capable of true in-vivo quantification of aortic regurgitation (without calculation), there is no 'Gold standard' for comparison of accuracy. Reproducibility is also good, both for inter-study and intra/inter observer comparisons [39, 40]. The technique is however subject to the same potential problems as all flow techniques, highlighted earlier (under 'flow and velocity quantification'), and care is required to ensure good accuracy of the measurements. In particular, non-breath-hold flow sequences are recommended for their lower background flow offset errors. Quantifying AR with CMR has recently shown a good ability to predict symptom development and the need for valve replacement surgery in the near future , with a regurgitant fraction > 33% providing the optimal threshold for identifying patients likely to require surgery within a few years. Given the difficulty in timing valve replacement surgery in patients with severe AR , this may become a valuable tool in clinical management, with the potential to identify suitable patients for early surgery. The potential improvement in outcome requires confirmation in a clinical trial however.
AR quantification can also be achieved by a comparison of the differences in LV and RV stroke volume from cine imaging alone , but this is less direct and relies on the lack of any other valve regurgitation or shunt. It also assumes accurate contour placement for volumetric assessment, and inaccuracies in measuring any of the four sets of contours (LV and RV in both diastole and systole) can result in significant errors. Careful contour placement is therefore required for this technique, particularly for the difficult RV contours. It is however a useful technique when flow quantification cannot be performed, or as an internal validation of the flow technique. An approximate assessment of the severity of aortic regurgitation can also be obtained by visualisation of the signal void of the regurgitant jet on cine imaging. A narrow jet width at the origin suggests lower degrees of regurgitation, while a wide jet, particularly with a core of high signal from laminar flow, suggests more severe regurgitation. This method is subject to many potential errors however (indicated earlier), and is not recommended for accurate evaluation.
Accurate LV volumes with CMR can aid clinical assessment of the impact of AR, and the high reproducibility is particularly useful for serial assessment, which is important for the management of a condition that has a long asymptomatic phase. CMR-derived LV end-diastolic volumes have also shown some ability to predict the onset of symptoms or other indications for valve surgery , and although less strong than quantifying the regurgitation itself, it can provide a useful adjunct in predicting outcome. CMR can also provide a detailed assessment of aortic root anatomy, which can assist in identifying the cause of the regurgitation, and/or whether the root needs replacing at the time of valve replacement surgery.
Taken together, the many CMR techniques useful in AR, including regurgitation quantification, LV volumetric assessment and aortic root anatomy, make CMR the optimal tool for comprehensive assessment.
Similarly to AR, an approximate assessment of the severity of mitral regurgitation can be obtained by visualisation of the signal void of the regurgitant jet on cine imaging, with a narrow jet width suggestinglower degrees of regurgitation and a wide jet (particularly with a core of high signal) suggesting more severe regurgitation. However, the same limitations apply, and this method is only useful as a rough guide.
The pulmonary valve and right ventricular outflow tract can be difficult to assess with echo, due to several factors. The location of the valve and outflow tract immediately behind the sternum makes it difficult to position the echo probe adequately to visualise the area. The qualitative echo assessment of pulmonary regurgitation is also less robust than for aortic regurgitation, and grading regurgitation severity can be difficult. Thirdly, the right ventricle can be difficult to assess, particularly for volumetric assessment, due to its unusual shape. CMR is therefore particularly valuable for assessing the pulmonary valve, with its combination of free choice of imaging planes, velocity & flow assessment, and accurate assessment of RV anatomy and volumes. Despite the very thin nature of the normal pulmonary valve, making it difficult to visualise with CMR, these other advantages are considerable, and CMR should be considered the 'Gold standard' for assessment of the pulmonary valve and RV outflow tract.
Trivial or mild pulmonary regurgitation (PR) is common in normal subjects (~30% of the population ), but rarely of importance. More significant degrees of regurgitation are usually related to congenital heart disease, with the largest group being patients with repaired tetralogy of Fallot, who commonly have significant residual PR [58–60]. CMR has revolutionised the investigation and follow up of such patients, as it accurately assesses two important aspects - the quantity of regurgitation and RV volumes/function , and is the method of choice for examining PR in this patient group.
Percutaneous pulmonary valve replacement using a stent-valve is increasing in popularity, and accurate sizing and anatomy of the pulmonary outflow tract is important for determining suitable patients [66, 67]. CMR provides the required detail, in addition to accurate assessment of the valve lesion itself, and is invaluable in the assessment of patients for this procedure . Despite the metal content of the stents, follow-up flow imaging can still occur above & below the stent, and some newer nitinol stents can allow flow assessment within the stent , though the accuracy is more uncertain.
Quantification can be achieved using pulmonary flow measurement (as acquired for pulmonary regurgitation), combined with RV stroke volume to calculate the regurgitant volume (= RV stroke volume - pulmonary forward flow) and the regurgitant fraction (TR/RV stroke volumes × 100%) , in much the same way as for mitral regurgitation. The same limitations apply, as combination MR techniques are used to calculate regurgitation quantity, and care is required with both the flow sequence and RV contouring, which can be especially difficult. Accuracy of the flow sequence in particular is also reduced in very irregular rhythms - not uncommon with TR. Quantifying the TR can also be assessed using the difference in ventricular stroke volumes if only a single valve leak is present , with the same issues about the need for accurate contour placement as in MR.
Patients with abnormal placement of the tricuspid valve (Ebstein's anomaly) often present a challenge for assessing true RV volumes & function, as well as the extent of tricuspid regurgitation, due to the difficulty of identifying the true ventricle on short axis images. Good post-processing software which allows identification of the valve position in the long axis views can help considerably with this, and CMR can produce accurate assessments which aid management [71, 72].
Although extremely rare, and not routinely assessed with CMR, this valve lesion can be examined when required. Valve area can be measured by placement of an image slice through the valve tips in diastole, as for mitral stenosis, and forward velocity through the valve can be measured, though this parameter is perhaps less useful.
CMR can be used to assess patients with multiple valve lesions, obtaining a detailed assessment of the severity of each component whether these occur in the same valve (i.e. mixed valve disease) or in different valves. As an extreme example, a patient with both mixed aortic and mixed mitral valve disease could have the opening area of each valve measured by direct planimetry with cine imaging to assess stenosis, the aortic regurgitation quantified from the diastolic (regurgitant) flow above the aortic valve and the mitral regurgitation quantified by subtracting the systolic (forward) flow above the aortic valve from the LV stroke volume. LV volumes and function would also be assessed. In this way, a comprehensive assessment can be undertaken.
Prosthetic valve dysfunction can also be assessed. If one leaflet of a bi-leaflet tilting disc valve fails to open properly, one side of the flow pattern will be missing and this may be identified with CMR. Paraprosthetic leaks can be visualised with careful image positioning, and can even be quantified with careful through-plane velocity mapping. While less conventional than other imaging modalities, these techniques can be extremely helpful in selected patients. CMR is also particularly good for assessing the anatomy of the aortic root around the valve, including grafts and valve-graft conduits, and any dissection or false aneurysm that may be present.
Irregular cardiac rhythms degrade image quality, which affects the assessment of ventricular function, though the effect on this is usually small. The accuracy of flow measurement can also be reduced  as flow sequences are acquired over several cardiac cycles (typically 10-12), with data acquired in a complex manner, and reconstructed based on the assumption of a regular rhythm. In subjects with an irregular rhythm, the complexity of the data acquisition means that although acquired over several cardiac cycles, flow is not averaged over these, as might be expected. Where the beat-to-beat variability is small (e.g. atrial fibrillation with a controlled rate), these errors are usually not clinically significant . Very irregular rhythms however (e.g. uncontrolled atrial fibrillation, multiple ventricular ectopics) can present a challenge, particularly to acquiring accurate flow data. Intelligent planning of the timing of data acquisition to the ECG can offset some of the problems, but some patients still present a challenge and caution should be exercised in interpreting the flow data in these. Cine imaging may be more reliable as it is more amenable to intelligent ECG gating and is less affected by arrhythmias, particularly during systole. Ventricular volumes vary with differing heart rates due to differential filling, and these physiological changes do of course still occur, and result in slight blurring of the myocardial borders. The result is an approximate averaging of the volumes over the several cardiac cycles of image acquisition, which is usually acceptable. Flow imaging may be improved with the newer ECG gating techniques for arrhythmias, which typically omit cardiac cycles outside a defined range.
One limitation that remains with CMR is the inability to measure directly the pressure inside a vessel or cardiac chamber - a limitation of all imaging modalities. As for echocardiography, CMR can measure velocity across a stenosis and derive a pressure drop from this, but absolute pressure quantification remains elusive. Cardiac catheterisation remains the most accurate method for assessing this. One paper has identified how CMR may indirectly indicate pressure, by examining the complex flow patterns in the pulmonary artery using 3-dimensional flow imaging . The paper suggested that pulmonary pressure (measured invasively) was strongly linked with the type of flow pattern in the main pulmonary artery. The data requires further validation but indicates the novel approaches to assessment that CMR may bring in the future.
CMR can provide a comprehensive assessment of valvular heart disease, including quantification of valve regurgitation and other flows, and accurate cardiac volumes and mass for assessing the effect on both ventricles. Combined with the ability to image all areas of the heart (including difficult areas such as the right ventricle and pulmonary valve), it is an excellent adjunct to echocardiography for investigating patients with valve disease. Further studies of clinical outcome, using quantitative CMR data to guide management, are needed to enhance it as a strong tool for guiding clinical practice.
SGM has 15 year's experience of all areas of cardiovascular magnetic resonance, in addition to many years of echocardiography and clinical cardiology. He is the clinical lead for cardiac imaging in a major tertiary centre university hospital in Oxford, UK, and is part of Prof. Stefan Neubauer's acclaimed CMR research group. His clinical and research interests focus on valvular and aortic disease.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.