T2-WEIGHTED CARDIOVASCULAR MRI (CMR)

Cardiovascular MRI uses a high-strength static magnetic field to align protons which “spin” within the body around an axis in line with the main magnetic field. These spins are excited with radiofrequency pulses, after which gradual realignment occurs to a resting state. Energy is dissipated as proton relaxation occurs and this energy is recorded by phased array coils. The energy dissipation is defined by parameters known as T1 and T2 relaxation times, which vary according to differing hydrogen content within tissues. T1 relaxation time measures the time after excitation to recover longitudinal magnetisation in equilibrium. T2 relaxation time measures the transverse magnetisation decay. Different sequences are designed with different preferences (weighting) for these relaxation parameters to enable tissue characterisation and assess soft-tissue water. A linear correlation exists between T2 relaxation times and myocardial water content—the higher the water content, the greater the signal intensity on T2-weighted images.1 This relationship has allowed T2-weighted sequences to become established in the diagnosis of a number of acute inflammatory states—in particular, viral myocarditis.2

MYOCARDIAL OEDEMA

The term myocardial oedema refers to both fluid accumulation in the cardiac interstitium (vasogenic) and to myocyte swelling (cytogenic). The water content of normal myocardium is stable in most mammals: 77% of the 380 ml water per 100 g of dry tissue is intracellular, while 23% is intravascular. Usually, there is only a very small interstitial component, contained within a gel of collagen, elastic fibres and glucosamine glycans, which drains from the subendocardium to subepicardium via the lymphatic system.3 More fluid accumulates in the interstitium when the normal homeostasis between the coronary microcapillaries and myocardial lymphatic glands is disturbed.4 In ischaemia, myocardial microvascular permeability increases owing to endothelial injury, allowing the accumulation of osmotically active substances and the influx of water. At the same time, reduction in myocardial contraction impairs lymphatic drainage. The resulting increase in myocardial water content has been demonstrated in experimental coronary occlusion,5 myocardial infarction (MI),6 reperfusion injury7 and after coronary artery bypass surgery.8 9 Myocardial oedema is not only a consequence but also a cause of decline in cardiac function. A 3.5% increase in myocardial water content, produced experimentally in healthy animals by acute elevation of coronary sinus pressure and reduction of lymph flow, results in a reduction in cardiac output of 40%.10 Acute insults produce a more marked reduction in cardiac output when extravascular fluid content is chronically elevated as, for example, in pulmonary or systemic hypertension.10 Chronic myocardial oedema results in alteration in myocardial structure, with an alteration in collagen subtype and the development of myocardial fibrosis.11

CHANGES IN THE MYOCARDIUM DURING AN ISCHAEMIC INSULT

Myocardial oedema is thus detectable using CMR after infarction; however, the clinical utility of this observation depends upon an understanding of the precise nature and time course of the changes occurring in myocardial tissue during ischaemia, infarction and reperfusion. This understanding is largely derived from experimental cellular work.9 The appearance of oedema depends on the persistence or absence of the microvascular circulation, the perfusion pressure of the epicardial coronary arteries and the extent and transmurality of cell necrosis.12 Short-lived ischaemic episodes, less than 2 h of vessel occlusion, induce an increase in myocardial cellular water content with capillary leakage and accumulation of metabolic products in the absence of cell membrane disruption.9 This results in myofibrillar oedema, and “stunned” myocardium with reduced contractile function.7 Longer, intermediate episodes (90 min to 4 h) of ischaemia induce alterations in the cell membrane cytoskeleton and damage to the microvascular circulation, leading to massive interstitial oedema, an increase in end-diastolic thickness in the oedematous infarct zone and mechanical dysfunction.13 Finally, prolonged ischaemia, more than 7 h from total vessel occlusion in a non-collateralised infarction, results in loss of the microvascular circulation with myocardial dysfunction, myocardial thinning and a lack of intramural oedema.

The influence of reperfusion on myocardial oedema has also been documented experimentally.5 Bragadeesh et al studied the effects of reperfusion on acutely “stunned” myocardium using both MRI tagging and echo.7 They demonstrated that improvements in postischaemic regional myocardial dysfunction over 5 days were related to the reduction in myofibrillar oedema seen at postmortem analysis. The authors hypothesised that increased local oncotic pressures from microvascular capillary leakage induces myocyte swelling and myofibrillar oedema with consequent regional dysfunction.

DURATION OF OEDEMA ON T2-WEIGHTED CMR

Reports have already confirmed the presence of high T2-weighted signal intensity representing oedema in acute MI and its absence in chronic scar. Uncertainty exists, however, about how long myocardial oedema persists after an ischaemic insult. Useful information has been derived from serial measurements of oedema using T2-weighted CMR performed after therapeutic embolisation of the septal artery in patients with hypertrophic obstructive cardiomyopathy.14 In that study, oedema was present up to 28 days after the alcohol-induced infarction. By 3 and 6 months, no oedema was seen in any patients. Other studies have suggested time points extending up to 1 year.15 16

UTILITY OF T2-WEIGHTED DETECTION OF MYOCARDIAL OEDEMA IN THE CLINICAL SETTING

If CMR can reliably detect myocardial oedema representing areas of early or recent but limited cell death, a number of possible clinical avenues open up. Evidence is accumulating as to the clinical utility of this technique.

Acute coronary syndromes

Arai and colleagues first proposed T2-weighted MRI as a measure of the area of myocardium that is hypoperfused during an ischaemic episode. High signal intensity areas obtained from T2-weighted MRI performed 2 days after 90-min left anterior descending coronary artery occlusion in dogs were comparable with the “area at risk” measured by fluorescent microspheres. This area was typically transmural and larger than that of the infarcted region documented in the same scan by delayed gadolinium hyperenhancement. Importantly, the area of high signal intensity on T2-weighted imaging without delayed hyperenhancement was subject to reversible injury, indicated by demonstrating resolution of the T2-weighted signal in association with recovery of contractility at 2 months.17

This ability to quantify an area of myocardium “at risk” has an application in acute coronary syndromes, when patients at high risk need to be identified early in order to optimise treatment. Raised cardiac troponin is currently the most sensitive biomarker but may not be raised within 12 h of the onset of symptoms; however, the earlier acute revascularisation occurs, the more likely that a good outcome will result. The likelihood of success from acute revascularisation drops significantly as early as 4 h after arterial occlusion.18 The addition of T2-weighted sequences in the emergency setting has begun to be explored by Cury and colleagues, who performed CMR as part of the triage of patients with acute chest pain but initially negative cardiac enzymes and no acute ischaemic change on ECG while waiting admission to hospital for a “rule-out MI protocol”.19 CMR incorporating T2-weighted sequences performed within the first 3 h of presentation correctly identified 11 out of those 13 who went on to have an acute coronary syndrome confirmed by eventual clinical criteria. While the focus for Cury and colleagues was that T2-weighted imaging improved the ability of CMR to “rule out” an acute coronary syndrome, the detection of hyperintense signal in those with unstable angina but no subsequent infarction hinted at the possibility of early characterisation of those with ischaemia (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Cardiovascular MRI (CMR) short-axis images from a patient with unstable angina and normal troponin. Left, T2-weighted CMR showing extensive anterolateral oedema; right, no evidence of late gadolinium enhancement.

The culprit lesion

Current guidelines for the management of both ST- and non-ST-segment elevation MI state that “only the culprit lesion should be dilated in the acute setting”.20 21 In an acute coronary syndrome, the electrocardiogram, echocardiogram and coronary angiogram alone may be inadequate to select the acute “culprit lesion”, particularly when multivessel disease is found. The presumed ability of the interventional cardiologist to recognise an “unstable” plaque or to reliably differentiate an acute from a chronic occlusion at angiography has never been validated, while T2-weighted MRI has a high specificity for acute as opposed to chronic MI.22 23 In a study of 15 patients in the acute (1 day) and chronic (2 months) phase of a first MI, both acute and chronic studies displayed delayed enhancement (DE) while high SI in T2-weighted imaging was a feature of acute but not chronic MI. The reliability of this method was confirmed by differentiation of a further 54 events into acute and chronic infarction using the presence or absence of high signal intensity T2-weighted images.22 If the ability of T2-weighted CMR to accurately highlight the area of myocardium subject to a recent insult is confirmed, then more appropriate revascularisation ought to ensue. Recognition or exclusion of an acute event could also prove to be an important aid in deciding on the appropriate treatment (anti-ischaemia or antiarrhythmia device) for patients presenting after life-threatening arrhythmia, in which analysis of cardiac enzymes does not always confidently rule-in or exclude infarction.

Myocardial infarction and viability

T1-weighted inversion recovery imaging for the detection of delayed contrast enhancement (DE-CMR) has become the preferred method for the identification of irreversibly damaged myocardium.24 However, while DE accurately reflects fixed expansion of the extracellular, extravascular component of the myocardium due to scar formation in chronic MI, the contrast agent has an increased volume of distribution due to abnormal cell membrane permeability in acute MI. This means that shortly after MI, DE may overestimate the area of irreversible injury by up to 10–15% compared with the true extent of infarction on histology.25 This percentage is not fixed and is likely to depend upon a number of factors, not least of which may be the reperfusion strategy employed. There is some evidence that T2-weighted imaging may give insight into this potentially salvageable myocardium.26 Friedrich and colleagues studied 92 patients 3–12 days after acute MI, all but one of whom had undergone successful reperfusion by angioplasty with or without preceding thrombolysis. A transmural area of oedema was identified in all postinfarct patients by high signal intensity on T2-weighted sequences which consistently exceeded the size of the infarcted myocardium as determined by DE (fig 2). The difference between these areas was quantified and labelled “salvaged myocardium”, compared with the areas with late enhancement that were considered not clinically recoverable.27 Consistent with this hypothesis, there was an inverse correlation between the size of the salvaged myocardium and the delay to reperfusion.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Cardiovascular MRI (CMR) short-axis images from a patient 3 days after presenting with an anteroseptal ST elevation myocardial infarction (STEMI). Left, T2-weighted CMR showing extensive anteroseptal oedema with a thin subendocardial layer of microvascular obstruction visible (arrow); right, anteroseptal late gadolinium enhancement subendocardial microvascular obstruction still visible (arrow). The spatial extent of myocardial injury in the oedema-sensitive T2 imaging is consistently larger than that of the necrosis-sensitive late enhancement.

Acute coronary syndrome with normal coronary arteries

Between 10% and 15% patients presenting with troponin elevation acute coronary syndromes are found to have normal or near-normal coronary arteries.28 T2-weighted CMR has found a further role alongside DE in identifying the aetiology of these cases.29 The differential diagnosis includes a true ischaemic/infarct episode with coronary embolus, spasm or rupture of a non-stenotic plaque, myocarditis and tako-tsubo or other cardiomyopathy. The pattern of high T2 signal intensity in myocarditis is quite different from that of a coronary or embolic event, which will localise to a single coronary territory (fig 3).2 Tako-tsubo is characterised by apical oedema on T2-weighted sequences, with DE uncommon and limited to micro-foci that appear to improve with time.30 Clarification of the diagnosis in such cases, and particularly exclusion of MI, is of utmost importance to guide future medical treatment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Cardiovascular MRI (CMR) vertical long-axis image from a patient with chest pain and symptoms of acute heart failure due to myocarditis. T2-weighted CMR showing patchy areas of oedema not specific to coronary artery territories.

THE USE OF T2-WEIGHTED CMR TO DETECT MYOCARDIAL OEDEMA AS A RESEARCH TOOL

In translational research, the concept of identification of the “area at risk” is an appealing one. For example, to compare reperfusion strategies in STEMI, large randomised studies with long-term follow-up are currently required to detect differences in mortality or major adverse cardiac events. Surrogate end points such as left ventricular function on echocardiogram or scintigraphy may be used in preliminary studies but have limited practical utility for two reasons. First, the degree of intraobserver variability in reporting may be larger than the subtle differences that might be expected to occur in the end point. Second, large sample sizes are needed to overcome the variability in infarct size between patients, which may differ depending on the precise location of the coronary occlusion and on the extent of collateralisation. T2-weighted sequences in combination with DE offer the opportunity to compare size of completed infarction to the area at risk from an occlusion for a given patient before and after intervention. This offers a measure of infarction normalised for the volume of myocardium at risk as a study end point, which could significantly reduce the sample size in any investigation seeking to compare efficacy of an intervention. This concept has already been employed in the assessment of an intravenous peptide FX06 in reducing reperfusion injury during treatment of acute MI, although this study sought to compare necrotic core size by microvascular obstruction by DE compared with infarct size by DE rather than using T2-weighted imaging.31 The potential advantages of an approach including T2-weighted imaging were recognised in the accompanying editorial, and many other potential areas for use exist beyond reperfusion, including human studies on stem cell therapy in the infarct zone and the influence of ischaemic preconditioning.32 Animal work in both of these areas has benefited from being able to use the “area at risk” as a denominator for any benefit gained and this should now become feasible in human studies.

LIMITATIONS

Some questions remain to be answered before T2-weighted sequences can be recommended in all CMR protocols, including those for clinical research into outcomes. In vivo visualisation of myocardial oedema has proved difficult and there are a number of technical considerations.

First, T2 imaging sequences are susceptible to artefact from pulsatile blood flow. Conventional T2 sequences have relied on passage of blood through an imaging plane to remove signal and produce tissue–blood contrast. This washout, however, may be ineffective and more recent sequences have been developed using a double-inversion blood-nulling preparation to further suppress signal.33 Despite these developments, optimal tissue–blood contrast is still produced when flow is perpendicular to the structure being imaged, and non-suppressed blood can still produce difficulties in interpretation, particularly when there are areas of slow-flowing blood near the apex of the ventricle. One method to reduce this in clinical practice is to compare T2-weighted images side by side with cine images to confirm anatomical borders.34

Second, longer TR times required to minimise T1 weighting come at the cost of long breath-holds which are required to minimise respiratory artefact. With the introduction of multi-echo spin-echo sequences (turbo spin echo or fast spin echo), breath-holds may be as low as 10–15 s but the long echo times inherent in T2-weighting reduce the amount of signal. One method to improve signal is to increase slice thickness but this carries the cost of lower spatial resolution.33

Third, a further problem particular to the inherently low signal-to-noise ratio (SNR) of T2-weighted images arises owing to the signal intensity gradient profile of the surface coils used in clinical MR imaging. This means that there is a progressive loss of signal moving away from the surface coil, leading to a drop in the sensitivity of T2-weighted imaging to detect oedema in the posterior and lateral myocardium. Figure 4 is a good example of a surface coil image with significantly less signal in the inferolateral (distant) region, leading to a loss of sensitivity. The use of the body coil however, comes at the cost of lower SNR, which may be partly compensated by imaging thicker slices. A preferable option has been the introduction of phased array surface coils with signal intensity correction algorithms.33

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Cardiovascular MRI (CMR) short-axis images from a patient 12 h after primary percutaneous intervention to an occluded circumflex artery. Left, T2-weighted CMR showing no evidence of oedema; right, lateral late gadolinium enhancement (arrow). Left, is a good example of a surface coil image with significantly less signal in the inferolateral (distant) region, leading to a loss of sensitivity and hindering interpretation.

It also remains unclear just how much of an ischaemic insult is needed to generate the degree of myocardial oedema that is visible on T2-weighted CMR. The high sensitivity of this technique in detection of acute coronary events has been derived largely from cohorts of patients with STEMI. Thresholds need to be established for duration of ischaemia, severity of ischaemia, size of territory involved and the relative impact of timeliness and completion of revascularisation (fig 4).

CONCLUSIONS

T2-weighted CMR images can identify an area of previously viable myocardium that has been subject to a recent insult in the form of ischaemia, infarction, reperfusion or inflammation. In the context of patients presenting with symptomatic coronary disease, this has potential utility in identification and to some degree quantification of risk of an acute event. In addition, information concerning the timing of the insult as well as its localisation to the territory of a coronary artery is possible. When combined with other CMR modalities such as late contrast enhancement and stress perfusion imaging, valuable information sufficient to determine the benefit of any medical or invasive intervention might be derived. Further clinical research is required to determine in which patients the identification of oedema will be most useful. Meanwhile the potential applications in translational research, particularly the direct quantification of myocardial salvage in acute MI, appear promising.