Correlation of Electrocardiogram and Regional Cardiac Magnetic Resonance Imaging Findings in ST‐Elevation Myocardial Infarction: A Literature Review

Abstract

Background

Patients with acute ST‐elevation myocardial infarction (STEMI) benefit substantially from emergent coronary reperfusion. The principal mechanism is to open the occluded coronary artery to minimize myocardial injury. Thus the size of the area at risk is a critical determinant of the patient outcome, although other factors, such as reperfusion injury, have major impact on the final infarct size. Acute coronary occlusion almost immediately induces metabolic changes within the myocardium, which can be assessed with both the electrocardiogram (ECG) and cardiac magnetic resonance (CMR) imaging.

Methods

The 12‐lead ECG is the principal diagnostic method to detect and risk‐stratify acute STEMI. However, to achieve a correct diagnosis, it is paramount to compare different ECG parameters with golden standards in imaging, such as CMR. In this review, we discuss aspects of ECG and CMR in the assessment of acute regional ischemic changes in the myocardium using the 17 segment model of the left ventricle presented by American Heart Association (AHA), and their relation to coronary artery anatomy.

Results

Using the 17 segment model of AHA, the segments 12 and 16 remain controversial. There is an important overlap in myocardial blood supply at the antero‐lateral region between LAD and LCx territories concerning these two segments.

Conclusion

No all‐encompassing correlation can be found between ECG and CMR findings in acute ischemia with respect to coronary anatomy.

Cardiac magnetic resonance (CMR) imaging is considered the gold standard in the assessment of cardiac function and structure.1 CMR imaging has emerged as a powerful noninvasive technique for the evaluation of myocardial function, perfusion, and viability in a single study without the need for radioactive tracer administration, which is a distinct attribute of the method.2, 3 CMR imaging has been shown to have a higher spatial resolution when compared to nuclear perfusion imaging techniques.4 Late gadolinium enhancement (LGE) CMR after administration of Gadolinium‐chelated contrast agents is a well‐established method for assessing focal fibrosis of the myocardium in myocardial infarction (MI).5 The delayed‐enhancement technique also has high sensitivity for the identification of acute MI (AMI)6 with the ability to detect as little as 1 g of irreversibly damaged tissue.7 Moreover, abnormalities of left ventricular (LV) wall dynamics are sensitive markers of myocardial ischemia.8 The exact measurement of infarct size, which is possible by CMR, may provide valuable information on ventricular remodeling, arrhythmic potential, and prognosis and is a commonly used surrogate endpoint for the evaluation of new MI therapies.9

The most widely used clinical method to diagnose acute coronary occlusion is by observing ST deviation on the electrocardiogram (ECG).10 After acute epicardial coronary artery occlusion, and after the appearance in the ECG of tall and peaked T waves, ST elevation, representing transmural ischemia, appears if the artery remains occluded.11 Methods based on the distribution of ST changes have proven to be useful for localizing the myocardial area at risk (AAR) during coronary occlusion.12, 13, 14 Based on anatomical studies, a relationship between the location of infarcted areas and Q waves in the ECG was accepted and, with minor modifications, implemented in scientific statements and textbooks.

The 12‐lead ECG is a cheap and universally available diagnostic method to detect and risk‐stratify AMI. However, as with many other diagnostic methods, the ECG has its limitations potentially leading to false diagnostic and prognostic conclusions. Investigators originally used experimental and autopsy studies as the reference method to determine the location and to estimate the size of MIs based on the ECG. Later on, ECG findings were compared with coronary angiography, echocardiography, and nuclear imaging. CMR imaging has now emerged as the golden standard with respect to size and location of MI.

Our aim is to give an overview of the literature regarding ECG and CMR correlation in acute ST‐elevation MI (STEMI), with the emphasis on the detection of regional changes. Because the coronary artery circulation is critical in determining the character of both ECG and CMR changes in AMI, coronary anatomic aspects will be covered.

Coronary Artery Anatomy

The coronary artery circulation is composed of two principal coronary arteries, the left (LCA) and the right coronary artery (RCA). Variations in the branching patterns are extremely common in the human heart. According to the Bypass Angioplasty Revascularization Investigators (BARI) classification, the RCA is predominant in approximately 85% of individuals, providing the posterior descending (PD; i.e., posterior interventricular) branch and at least one posterolateral (PL) branch.15 In 7–8% of individuals, the coronary circulation is left dominant; the PL, the PD, and the atrioventricular nodal branches are all supplied by the terminal portion of the left circumflex coronary artery (LCx). In another 7–8% of hearts, there is a codominant or balanced system, in which the RCA gives rise to the PD branch, and the LCx gives rise to all the PL branches. The left anterior descending (LAD) and the LCx arteries are direct continuations of the left main trunk. One or more left diagonal (LD) and approximately 10 septal branches arise from the LAD, while left obtuse marginal branches arise from the LCx. In approximately one‐third of individuals, the left main coronary artery trifurcates; the intermediate branch comes off between the LAD and the LCx.

In a normal heart with right‐dominant coronary circulation, the LCA supplies approximately 5 times more blood to the LV than the RCA; the corresponding number for the LAD and LCx is approximately 3.5 and 1.5 times, respectively.16 In the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease (APPROACH) scoring system, the LV is divided into regions according to pathological studies in humans evaluating the relative proportion of the myocardium perfused by each coronary artery.17 Considering the location (proximal, mid, or distal) of the culprit lesion, investigators created a template that calculates the jeopardized myocardium for a given site occlusion, taken into consideration coronary artery dominance.18

Association between Segmental MI ECG Signs of Necrosis

Monumental work in the 1930s and 1940s established the theoretical and empirical basis for the association between myocardial necrosis and ECG findings.19, 20 Based on pathological correlation, a relationship between the location of infarcted areas and Q waves was accepted and, with minor modifications, implemented in scientific statements and textbooks.21 According to these studies, Q waves correspond to the following myocardial segments: leads V₁‐V₂ to the septal wall, leads V₃‐V₄ to the anterior wall, leads I and aVL to the lateral wall, and leads II, III, and aVF to the inferior wall. The mirror pattern (high R wave) in leads V₁‐V₂ was traditionally considered as corresponding to the basal part of the inferoposterior wall (named true posterior MI).22 The traditional view was that Q waves represent transmural MI, while subendocardial MI is reflected by ECG changes other than Q waves. Later on, autopsy studies showed that about half of subendocardial MIs generated pathological Q waves, while about half of transmural MIs did not.23 The terms Q‐wave and non‐Q‐wave MI replaced the former designations. CMR studies have shown that the Q‐wave/non‐Q‐wave distinction is determined rather by the total size rather than by the transmural extent of the underlying MI.24

Already in the 1930s, studies showed different premortal ECG patterns depending on which epicardial coronary artery had thrombotic occlusion on autopsy.25 With the introduction of coronary angiography, it became possible to correlate ECG findings of myocardial necrosis with this new technique.26 Reperfusion therapy, initially by intracoronary thrombolysis in the 1980s, resulted in a clear need to improve ECG diagnostics to be able to predict the culprit artery and the estimated size of the AAR by observing changes from the acute occlusive phase—the ST segment and the T wave. Since then, many investigators have presented ECG algorithms to predict the culprit artery and the level of occlusion within the culprit artery based on the distribution of ST deviations.

The 17‐Segment Model

Nuclear cardiology, echocardiography, cardiac computed tomography (CT), positron emission CT, and CMR are important imaging modalities of the myocardium and the adjacent cavity. The American Heart Association (AHA) writing group on Myocardial Segmentation and Registration for Cardiac Imaging aimed at harmonizing the imaging modalities with respect to the display of the ventricular segments.27 Based on autopsy studies, the group presented a 17‐segment model of the left ventricle (LV) for the visual interpretation of regional abnormalities. According to the statement paper, the model creates a distribution of 35%, 35%, and 30% for the basal, midcavity, and apical thirds of the heart. The authors also assigned individual LV segments to specific coronary artery territories (Fig. 1); segments 3, 4, 9, 10, and 15 were assigned to the dominant RCA, segments 5, 6, 11, 12, and 16 to the LCx, and segments 1, 2, 7, 8, 13, 14, and 17 to the LAD.

Figure 1.

Coronary artery distribution in right (upper row) and left (lower row) dominance. The right coronary artery is shown to the left, the left anterior ascending in the middle, and the left circumflex artery to the right.

Temporal Aspects of AMI and Its Relation to CMR Imaging

In acute coronary artery occlusion, biomechanical and functional abnormalities begin immediately after the onset of ischemia. There is an immediate loss of contractility during the first minute, and in about half an hour loss of viability develops (Fig. 2). Initially, myocardial perfusion decreases leading to diminished tissue oxygenation, repolarization abnormalities, diastolic and systolic dysfunction, edema, and finally tissue necrosis. The infarct progresses like a wavefront from the endo‐ to the epicardium.28 Progression to final necrosis takes 2–4 hours or longer, depending on factors like collateral circulation, persistent or intermittent coronary arterial occlusion, sensitivity of myocytes to ischemia, preconditioning, and individual demand for myocardial oxygen and nutrients.29 In this cascade myocardial edema can occur in 30 minutes and progress to completely irreversible injury within 6–12 hours. This creates a window of opportunity, during which prompt reperfusion therapy may salvage myocardial tissue, which is still at risk of infarction. The progression of MI is a dynamic process, and this is why the timing between the acute event and evaluation of scar is paramount. The typical techniques used in CMR imaging are shown in Table 1 together with the underlying pathophysiology.

Figure 2.

The ischemic cascade in the acute phase.

Table 1.

The Basic Sequences Used in Myocardial Infarction CMR Imaging

Sequence	Pathophysiology
T2‐weighted imaging	Myocardial edema, area at risk
Kine imaging, tagging techniques	Regional contractile and global ventricular function
Perfusion imaging, early gadolinium enhancement	Myocardial perfusion, microvascular obstruction
Late gadolinium enhancement	Infarct size, transmurality, persistent microvascular obstruction

Technical and Clinical Aspects of CMR Imaging in AMI

Edema Imaging

Acutely infarcted myocardium contains increased amounts of free water as a natural consequence of AMI, including both irreversibly damaged and potentially salvageable myocardium.30 Edema is visible in T₂‐weighted images as bright areas in otherwise dark myocardium. In AMI patients, bright myocardium on T₂‐weighted imaging reflects the myocardium at risk.31 Edema changes are most evident in the acute or subacute phase at least up to 1 week after the acute event,30, 32 after which they slowly fade away during the course of infarct healing with scar formation and resorption of infarct‐related myocardial edema and inflammation.30, 33 T₂‐weighted edema imaging can thus be used in differentiating between recent and healed MI.31, 34

Late Gadolinium Enhancement

Mc Namara et al. introduced the basic principle of using gadolinium enhancement.35 With the combination of the inversion recovery technique described by Simonetti et al.,36 it became possible to maximize the contrast between viable and infarcted myocardium. After gadolinium contrast administration, the enhancement reflects mainly irreversibly damaged tissue. Technically the best time window for infarct imaging is between 10 and 25 minutes postcontrast administration. If the acquisition is performed too early, LGE can lead to overestimation of the MI size.37, 38 There is significant decrease in the total extent of LGE between day 1 and day 7 post‐MI.39, 40, 41 CMR can be used in assessing MI size up to 1 week after the acute ischemic event. With time, the size of the infarct decreases as a result of resolution of tissue edema and gradual contraction of the resultant scar tissue by as much as 25% over a period of 4–8 weeks.42

Myocardial T1 Mapping and Extracellular Volume (ECV) Fraction

It is obvious that late gadolinium enhancement imaging provides crucial diagnostic and prognostic information in MI.43, 44 Nonetheless, LGE is a qualitative, not a quantitative diagnostic method. Previous canine CMR studies have shown alteration of the T1 time with fibrosis, which correlated with myocardial collagen content.45 The T1 values can be encoded in each pixel, and a parametric map (T1‐map) can be provided.46 T1 measures permit noninvasive detection of biologically and pathologically important processes related to excess water in edema, protein deposition, and other T1 altering substances such as lipid or iron in hemorrhage.46, 47, 48, 49, 50

The ECV technique is sensitive to the distribution of the LV myocardium into its cellular and extracellular interstitial compartments.47 Alterations in these compartments occurring from different physiologic and pathophysiologic biologic processes leading to fibrotic changes can be more easily detected with ECV than with LGE.51 This is true especially with the changes resulting from pathologies other than MI, where the differences between the normal and affected myocardium are less distinct.47, 51, 52, 53 Therefore, the more widespread clinical use of these tissue characterization techniques requires further exploration.

Microvascular Obstruction (MVO) and Intramyocardial Hemorrhage (IMH)

Despite the beneficial effects of reperfusion, the process of cell death may continue during the first hours of reperfusion, a phenomenon called myocardial reperfusion injury. It is the consequence of clogging of the small myocardial arterioles with embolic debris, acute inflammation, platelet aggregation, and vasospasm.54 According to Moens et al., reperfusion injury, with MVO as its main feature, has been detected in about 30% of patients.55 The role of the 12‐lead ECG for the assessment of MVO after PCI is not a topic to be discussed here, but it was recently covered in another review article.56

Early contrast‐enhanced CMR can reliably demonstrate the presence and extent of MVO after reperfused AMI.54 In addition, on LGE images MVO is seen as a central or subendocardial, hypointense, dark region within the enhanced bright area of infarcted myocardium. MVO progresses in size during the first 48 hours,57 after which microvascular damage most likely stabilizes.58 In one study of acutely revascularized MI patients, late MVO proved the most powerful predictor of global and regional functional recovery, even compared with transmural extent of MI.59

Besides MVO, IMH is considered another important feature of reperfusion injury; it is a consequence of blood extravasation through damaged endothelium.60 IMH has been assessed by using different T₂‐weighted imaging techniques.61, 62 Both MVO and IMH have been associated with adverse prognosis, adverse LV remodeling, and diminished recovery of LV function independently of infarct size.63, 64 CMR enables noninvasive evaluation of both MVO and IMH.

Left Ventricular Wall Motion Analysis by CMR

Following reperfused AMI, the LV undergoes structural alterations both within and outside of the area of infarction, referred to as LV remodeling. Ischemic regional remodeling starts soon after the onset of myocardial necrosis and progresses over time to extensive architectural changes of the whole ventricle.65 The pathophysiology of the remodeling process is complex with multiple ultrastructural, metabolic, and neurally mediated processes occurring in infarcted and remote myocardium.66 LGE detected in the infarct zone within 24 hours after primary PCI is an independent predictor of impaired LV systolic thickening and remodeling.67 On the other hand, contractile function improvement has been reported to occur as a result of successful reperfusion.68, 69

CMR is well suitable for the assessment of wall contraction patterns thanks to its capability for volumetric evaluation, and the ability to detect wall motion abnormalities even in regions inaccessible to echocardiography. In addition, CMR is less operator dependent than biplane echo.70

CMR using cine sequences is already considered the standard of reference for measurement of ventricular volume and various parameters of global ventricular function, such as ejection fraction (EF) and ventricular mass.71 The cine‐CMR technique allows accurate delineation of the endocardium and epicardium and offers a reproducible assessment of regional LV wall thickening.72

Masci et al. developed a CMR method to compare global and regional functional parameters.73 One week and 4 months after MI, overall LV functional improvement matched with improvements in regional function of the infarcted region. In contrast, use of the 17‐segment approach failed to depict significant differences over time in the infarcted myocardium, indicating that myocardial segments might contain a mixture of viable and nonviable myocardium, making accurate interpretation difficult.73

Masci et al. reported that improvement in global LV function matched with an improvement in systolic wall motion (SWM) but not with systolic wall thickening (SWT) in the infarcted myocardium.73 This has been explained by the hypothesis that functional recovery of less‐damaged outer myocardial layers may improve the myocardial wall motion leading to improved regional EF without substantial improvement in SWT.74 According to Nowosielski et al., SWT <30% indicates an infarct size that affects LV function with correlation to the LV ejection fraction.70 They also found a highly significant difference in SWT between segments with no LGE and those with transmural LGE of >50%, but pointed out that any infarcted segment is associated with some wall motion abnormality.70 Therefore MI has a certain influence on the contractile function of the LV, not necessarily resulting in impaired ventricular function or clinical deterioration.

LV wall motion, thickening, and strain can be measured also by using the myocardial tagging technique,75, 76 which is performed with cine MR imaging by applying a special radio‐frequency prepulse immediately following detection of R wave on the ECG tracing.77 Myocardial tagging allows the determination and quantification of LV wall thickening of specific myocardial segments.78, 79

Myocardial strain imaging provides a more accurate assessment of LV function, including circumferential and longitudinal contraction, compared with thickening analysis, which only shows cardiac motion in the radial direction.80, 81 Segmental strain analysis can be performed with CMR using myocardial tissue tagging with harmonic phase imaging, but has not yet become clinical routine.

Area at Risk

The AAR represents the amount of myocardium that is hypoperfused during coronary occlusion. Salvaged myocardium is defined as the AAR subtracted by the final infarct size and has an important role in estimating the therapeutic efficacy. Friedrich et al. concluded that the distinction between the areas with edema and LGE can be used as an estimate of AAR.82 They stated that the areas with T₂ edema were invariably transmural and subsequently larger than the regions of LGE. The difference between them thought to most likely represent myocardial salvage.82 This finding has also been supported with previous canine studies.83

The role of edema and AAR imaging with T₂‐weighted technique has been studied a lot. Taking into account the wavefront character of irreversible ischemic injury, beginning from the endocardium and progressing to the epicardium, affects the potential for recovery of the myocardium after an ischemic episode. Different myocardial layers have a different risk potential for ischemic damage. Histopathologic characterization of salvaged myocardium has shown that the subepicardial zone develops cellular necrosis less frequently than the subendocardial core of the infarction.28 The fact that patients with aborted MI have more subendocardial and less transmural infarction by CMR provides imaging evidence for the concept of the wavefront phenomenon.84 Taking these aspects into account, Ortiz‐Perez et al. showed that CMR infarct endocardial surface area (ESA) measured within a week after primary PCI in STEMI patients correlated with the angiographically determined AAR.18

The use of T₂‐weighted images in determining the AAR has its limitations.85 In a review article, Arai et al. discussed the role of T₂‐weighted MR imaging in the assessment of the AAR in patients with AMI.86 More recent study findings, which have evaluated aspects of gadolinium kinetics, indicate that early gadolinium‐enhanced images represent the AAR while LGE images correlate with the size of the MI.86, 87, 88

Also noncontrast T1 mapping technique has given promising results in revealing the AAR in acute coronary syndromes when combined in a clinical scan protocol.59, 89, 90

The Aldrich score is based on the total amount of ST deviation on admission and has been used as an ECG‐based estimate of AAR.91 In 78 reperfused STEMI patients, CMR methods (T₂‐weighted hyperintensity and LGE ESA) were better than ECG methods (Aldrich score) for measuring AAR.92 The authors speculated that the poor performance of the Aldrich score in estimating AAR might be explained by the dynamic nature of ST deviations in acute myocardial ischemia. The Anderson–Wilkins acuteness score was developed to enable assessment of acuteness of myocardial ischemia from the 12‐lead ECG, and studies have indicated that it may be associated with salvageable myocardium.93, 94 A recent study of 38 patients treated with primary PCI, in which SPECT and CMR imaging were used to determine AAR and final infarct size, showed a moderate correlation between the acuteness score and salvageable myocardium in patients with acute RCA occlusion but not in patients with LAD occlusion. At present, there is no clear consensus regarding the preferred ECG method to estimate AAR.

Correspondence between Left Ventricular 17 Myocardial Segments and Coronary Artery Occlusion

Segmental analysis of myocardial perfusion studies is based on certain assumptions concerning coronary anatomy. The AHA working group that recommended the 17‐segment model in LV imaging mentioned the tremendous variability in the coronary artery blood supply to myocardial segments, but the model was believed to be appropriate for assigning individual segments to specific coronary artery territories. Pereztol‐Valdés et al. analyzed the correspondence of the 17 LV segments with each of the three major epicardial coronary arteries by myocardial perfusion single‐photon emission CT (SPECT) studies during elective percutaneous coronary intervention (PCI) in 50 patients.95 The study showed that the most specific segments (anterior, anteroseptal, and all apical segments except the inferoseptal) corresponded to the LAD, and the basal anterolateral segment (segment 6) to the LCx, while no segment could be exclusively attributed to the RCA. Eight segments were not attributed to a single coronary artery. In the working group paper, the inferior segments (4, 10, and 15) were attributed to the RCA, but in the study by Pereztol‐Valdés et al., all patients with LCx disease had involvement of the inferior segments.95 These observations most probably reflect the great variability in blood supply to the inferior and lateral wall.

In 2007, Ortiz‐Perez et al. reported preliminary observations regarding the association between the anatomy of the coronary arteries and LGE localization in CMR.18 In a subsequent study they reported on the correspondence between the 17‐segment model and coronary artery anatomy with contrast‐enhanced CMR imaging in 93 patients flowing AMI treated with primary PCI.96 They found that four segments (2,7,8,and13) were 100% specific for LAD occlusion, while no segments were 100% specific for LCx or RCA occlusion. In a per‐segment analysis, 23% of segments with LGE were found to differ from the coronary distribution empirically assigned to the 17‐segment model. Misassignment of myocardial segments because of a deviation of true coronary anatomy from the standard model could alter the identification of the involved coronary vessel or the location of the culprit lesion.97 The agreement was remarkably low for the midanterolateral (segment 12) and apical lateral (segment 16) empirically attributed to the LCx—for all these segments, the culprit lesion was in the LAD.

In another study, where CMR images were coupled with noninvasive CT coronary angiography, the authors found the maximal discordance between the 17‐segment model and CMR in segments 12,15, and 16, which were supplied by the LAD in most cases.98 In this latter study, only 4 of 26 patients were found to agree with the model in all segments; on average, in a per‐patient analysis there were differences in approximately three segments. LCx segments were most often discordant (35%), with the majority of the segments attributed to this artery being located within the LAD distribution. In fact, the Bayés de Luna group has demonstrated that myocardial injury caused by occlusion of the first diagonal branch involves part of segments 7 and 13 of the anterior wall, but also 12 and 16 of the lateral wall resulting in Q waves in lead aVL and often also in lead I. Accordingly, this type of MI represents an MI of the anterior/midanterior region, not of the basal lateral (LCx occlusion) or high lateral.99, 100 By level, apical segments were found to differ most often (26%), especially segments 15 and 16. Previous observations indicated the greatest variability with segment 17, but Setser et al. found this segment to agree better in their study.27, 98

The LV apical area is more or less supplied by the LAD, portraying the large amount of myocardium supplied by this artery. Javadi et al. studied the coronary artery territories in perfusion images using PET/CT hybrid analysis.97 They agreed with the observations by Pereztol‐Valdés and Ortìz‐Pérez that the standard assignment of vascular territories proposed by the AHA 17‐segment model is frequently inaccurate.95, 96 However, the division by Cerqueira et al. into four walls and 17 segments is easy to understand and probably corresponds to a true anatomic correlation. They also reached promising results with CTA‐derived imaging. Donato et al. performed an ex vivo study of cadaveric hearts using multidetector CT.101 They showed that the LAD territory was larger than the AHA‐proposed 17‐segment model.101 No specific segment could exclusively be attributed to the RCA or the LCx.101 However, although segment 12 belonged to the LAD territory, the LAD and the LCx contributed almost equally to the blood supply of segment 16. Segment 12 corresponds to the lateral wall and due to many anatomic differences of coronary arteries, it may be more or less perfused by the LAD or the LCx. In case of occlusion of a large first diagonal branch, part of segment 12 will be involved.101

In our preliminary study of six patients with proximal LAD occlusion to be discussed more in detail later, either LGE or pathological SWT was assessed in segment 12 in 5/6 of the patients. The corresponding numbers for segments 15 and 16 were 6/6 and 5/6, respectively. Our preliminary observations seem to support the findings from the larger studies by Ortiz‐Pérez and Setser.96, 98 Ubachs et al. analyzed 12‐lead ECGs combined with SPECT imaging within 3 hours after primary PCI, and CMR of the LV within 24 hours in 11 patients with acute STEMI.8 They found that the angiographic culprit artery was correctly designated in 91%, 82%, and 91% of cases with perfusion SPECT, CMR, and 12‐lead ECG, respectively. CMR wall thickening proved to be an accurate method to localize myocardium at risk. The authors used the 12‐segment polar plot system for localization of regional changes, and therefore exact comparison with the 17‐segment model is not possible.

ECG and CMR Correlation

The correlation between the distribution of ST elevations in the 12‐lead ECG and the myocardial AAR in CMR has not been well established. Ortiz‐Pérez et al. presented the correlation between ECG leads with ST elevation and LV segments with hyperenhancement on CMR imaging.96 As previously mentioned, with regard to the culprit artery, segments 12 and 16 showed the maximal discordance from the 17‐segment model. In the patients with LGE of segment 12, 60% had ST elevation in lead I and 65% in lead aVL. In the precordial leads, the corresponding prevalence of ST elevation was less than 40% (highest prevalence in lead V₂: 37%). Regarding segment 16, the numbers for leads I and aVL were identical, while the proportions of patients with precordial ST elevations were somewhat higher (maximal 46% in leads V₄ and V₅). The authors did not report on the prevalence of combined ST elevations in the lateral extremity and the precordial leads.

The authors of this article performed a preliminary (unpublished) study for the correlation of ECG changes in acute proximal LAD occlusions after successful primary PCI with segmental wall motion abnormalities using CMR as the golden standard in six male patients (age 40–70 years). The angiographic lesion site was proximal to the first LD branch in all patients. In four patients the culprit lesion was proximal and in two distal (#4 and 5) to the first septal branch. CMR scanning was arranged within 3 days of the primary PCI. Established criteria for cut points for Q waves, ST elevations, and T‐wave inversions were used and the ECGs were recorded at admission, on the day of the CMR study, and at hospital discharge.102 In the 16‐segment bull's eye map, segments 1, 2, and 8 were considered as corresponding to ECG leads V₁/V₂, segments 7, 13, and 14 to leads V₃/V₄, segments 6 and 12 to leads I/aVL, and segments 11, 15, and 16 to leads V₅/V₆. The distribution of ECG changes in the 12‐lead ECG is shown in Table 2 and the correlation with SWT changes on CMR in Table 3. In four of the six patients, regional CMR changes were observed both in segments 12 and 16. All of these patients had involvement of the lateral extremity leads I and aVL and in the precordial leads V₂ and V₃. The involvement of the other precordial leads varied. We think that our preliminary findings are in line with the observations by Ortiz‐Pérez et al. both in the sense that segments 12 and 16 are involved in LAD occlusion, and that ST elevation in leads I and aVL are frequent in these cases.96 As pointed out earlier, the AHA statement indicated these segments to be associated with LCx disease. It seems that standard models of LV segmentation can be used only partly in assessing individual myocardial segments in relation to coronary artery territories.

Table 2.

Distribution of ST Elevations, Q waves, and Inverted T Waves in the 12‐Lead ECG at Admission, at the Day of Cardiac Magnetic Resonance Imaging, and at Discharge in the Six Patients, a Positive Finding Refers to at Least One Affected Lead

	Admission	MRI	Discharge
	ECG	ECG	ECG
ST Elevation	n	N	n
I/AVL	4	2	2
V1/V2	6	5	4
V3/V4	6	5	4
V5/V6	1	1	2
Q Wave
I/AVL	1	2	2
V1/V2	2	2	2
V3/V4	1	3	4
V5/V6	0	1	1
Inverted T Waves
I/AVL	1	5	3
V1/V2	2	3	3
V3/V4	2	3	4
V5/V6	0	2	2

Table 3.

Statistical Correlation between ST Elevations, Q Waves, and Inverted T Waves at Three Different Time Points, and Cardiac Magnetic Resonance Imaging, a Positive Finding Refers to at Least One Affected Lead

Time of ECG	Sensitivity	Specificity	PPV	NPV
Admission
ST elevation	79	42	58	67
Q waves	24	91	73	55
T‐wave inversion	15	73	36	46
MRI Day
ST elevation	79	67	70	76
Q waves	39	82	68	58
T‐wave inversion	52	58	55	54
Discharge
ST elevation	73	77	75	74
Q waves	55	79	72	63
T‐wave inversion	67	61	63	65

Meléndez‐Ramírez et al. used myocardial edema on CMR when correlating the location of regional myocardial changes with the ECG, which included standard and additional leads.103 They included 91 patients with a first STEMI with reperfusion therapy within 12 hours from symptom onset, who had a CMR study (mean 3 days after the ischemic event). The highest sensitivities for the ECG leads compared to the different segments (1–17) were: V₂ for basal anterior and anteroseptal, III and aVF for basal inferoseptal, III for basal inferior and inferolateral, V₇–V₉ for basal anterolateral, V₂ and V₃ for midanterior and anteroseptal, III and aVF for midinferoseptal, inferior, and inferolateral, V₂ and V₈ for midanterolateral, V₂–V₄ for apical anterior and septal, II, III, and aVF for apical inferior and lateral, and V₂–V₄ for apex. For these associations there is fairly good correlation with previously published data,96 but for the somewhat controversial segments 12 and 16 the results differ. In the study by Meléndez‐Ramirez,103 the most sensitive leads for segment 12 and 16 were V₂ and V₈ and I, III, and aVF, respectively, while I and aVL were the most frequently involved leads in the study by Ortiz‐Pérez. Actually, in the latter study, leads II, III, and aVF were involved in only 5–10% of the cases when segment 16 showed CMR changes.96

The different CMR techniques used may explain some of the differences, but it is obvious that especially segments 12 and 16 remain controversial. We conclude that the main explanation for differences in different studies is that no consensus can be made about the exact borders of each vascular territory. There is an important overlap in myocardial blood supply at the antero‐lateral region between LAD and LCx territories concerning these controversial segments.

New Terminology for the LV Walls and Location of Q‐Wave MI

The correlation between Q waves in various ECG leads and the affected myocardium has been studied by CMR. Based on these findings, a new terminology for LV walls and location of Q‐wave MIs based on CMR was proposed.99, 100 A main division into the anteroseptal and inferolateral zones with subdivisions were recommended (Fig. 3). The term midanterior was recommended for MIs located especially in the midlow segments (7 and 13) of the anterior wall. This MI subtype is typically caused by occlusion of the first LD branch, and in the ECG Q waves are present in leads aVL (I) and sometimes V₂. This coincides with the previously published paper on that by Sclarovsky et al., and is also supported by the paper of Meléndez‐Ramirez et al. showing that leads I and V₂ were among the most sensitive ECG leads to detect involvement of segments 12 and 16, typical for first diagonal branch occlusion in STEMI patients.103, 104

Figure 3.

The new terminology for LV walls and location of Q‐wave MIs based on CMR study (LV, left ventricle; MI, myocardial infarction; CMR, cardiac magnetic resonance).

The term true posterior MI was introduced by Grant and Murray to refer to the basal part of the LV wall positioned on the diaphragm.105 In 1964, Perloff defined the criteria for this clinical entity as the presence of an R/S ratio of >1 and an R‐wave duration >40 ms in lead V₁.22 Later on, the term inferoposterior wall was introduced to refer to the entire wall that lies on the diaphragm. A working group statement, in agreement with the terminology of Cerqueira et al., recommended that the term posterior be abandoned and that the term inferior should be applied to the entire wall that lies on the diaphragm.99 The recommendation was based on findings from CMR imaging, which showed that infarction of the basal and midsegments (4 and 10) of the inferior wall will generate increased R waves in leads V₃ and V₄ instead of in leads V₁ and V₂.106 It was also shown that an infarction located in the lateral LV wall, involving more than the basal segments (segments 5 and 11), may generate increased R waves in leads V₁ and V₂. The working group recommended using the term lateral infarction in case of high R waves in leads V₁ and V₂.99

The observations by Bayés de Luna et al. were supported by two later studies. Van der Weg et al. evaluated the ECG in the identification and quantification of lateral wall involvement in nonanterior MI.107 The CMR study was done 5 (2–12) days after an AMI. Significant correlations were found between lateral LV wall involvement and R waves in lead V₁ and V₆: tall and broad R waves in lead V₁ reflected lateral wall MI, especially in LCx occlusions.107 In patients with LCx occlusion (n = 11), there was a strong correlation between the height, width, and surface of R waves in lead V₁. The surface of R waves in V₁ had the highest correlations with the apical inferior (15) and apical lateral (16) segments. Loss of height of the R wave in V₆ was found to be a sign of lateral MI, and Q waves in V₆ were related to inferolateral involvement. Also the study of Rovai et al. with 50 patients in the post‐MI phase showed that a prominent R wave in lead V₁ is a specific sign of a large and transmural lateral MI.107

The fact that van der Weg et al. found a clear association between LCx occlusions and involvement of segment 16 may seem contradictory to the observations by Ortiz‐Pérez and others discussed earlier.96, 106 However, it is clear that anatomic variation exists with respect to which coronary artery branches irrigate the apical lateral wall. In the paper of Van der Weg et al., the lateral wall included all segments of the lateral zone. Also routine clinical practice shows that there is large variation between individuals in the distribution of side branches to this area, which can be subtended by a distal LD branch from the LAD, from the intermediate artery or from branches of the LCx.

We conclude that there is no doubt that the posterior wall does not exist and that the presence of a prominent R wave in V₁ is due to lateral and not to inferior wall (especially basal) involvement.

CONCLUSION

CMR imaging has offered new insights into the diagnostics of cardiac diseases, including acute myocardial ischemia and infarction. Certain CMR applications, such as those used for the evaluation of myocardial viability, represent the most accurate clinical test, whereas other applications are helpful adjuncts to standard tests such as the ECG.

When using a multidisciplinary diagnostic approach to AMI, the knowledge about strengths and weaknesses of different modalities is of paramount importance. In addition, familiarity with the dynamic character of the disease and the variations in coronary artery anatomy are of utmost importance.

Findings from studies dealing with the ECG–MRI correlation in the chronic MI phase have already resulted in changes of some basic concepts. There is general agreement that the “posterior” wall does not exist more; it is the inferobasal segment of the inferior wall. Also there is agreement that a prominent R wave in V₁ in post‐MI patients is explained by a lateral wall MI. Therefore, there has not only been a change of terminology (posterior vs lateral wall), but a change of MI location (lateral wall, not the inferobasal segment of the inferior wall), and this may have several clinical implications (infarct size, presence of ventricular arrhythmia, risk for sudden cardiac death, and cardiac rupture).

Some limitations still exist. Comparing ECG and CMR findings of acute myocardial ischemia, the best agreement has been reached for the anatomic territory of the LAD, which clearly involves the segments 2, 7, 8, and 13. Regarding segments 12 and 16, which also have been ascribed to the LAD territory, there are controversial findings in the literature. In addition, no segments are 100% specific for LCx or RCA occlusion. At this moment, no all‐encompassing correlation can be found between ECG and CMR findings in acute ischemia with respect to coronary anatomy. When assessing CMR findings in coronary artery disease, one has to take into consideration the variability in coronary artery anatomy.