Dynamic changes in injured myocardium, very early after acute myocardial infarction, quantified using T1 mapping cardiovascular magnetic resonance

Following reperfusion treatment in patients with ST-segment elevation myocardial infarction (STEMI), there is an abnormal accumulation of water within the injured myocardium which, when depicted using cardiovascular magnetic resonance (CMR) imaging, defines ‘the area at risk’ (AAR) [1, 2]. The AAR represents tissue that could potentially be transformed into necrotic and infarcted tissue if the injured myocardium were not adequately reperfused; therefore quantifying AAR is important in calculating myocardial salvage index (MSI) [3, 4]. This index is an established measure for testing efficacy of therapies in acute STEMI, but relies on the assumption that AAR is stable post STEMI [3‐5].

Recently, there has been debate over the stability of oedema /AAR immediately after reperfusion [6, 7]. Data from animal models, and subsequently humans, have suggested that oedema, as quantified on T2 weighted imaging and T2 mapping sequences, has a bimodal pattern [6, 7] The initial wave was attributed to reperfusion injury, which diminished after 24 h and was followed by a deferred wave which was related to tissue healing [6, 7]. The use of mapping techniques has added a new domain in measuring the severity of oedema beyond the anatomical area of oedema when quantifying the AAR [6, 7] In other words, quantification of water content becomes a consequence of the interaction of the area and the severity of oedema within AAR. The severity of oedematous reaction post reperfusion, quantified on T2 mapping, followed similar bimodal changes to the anatomical AAR, quantified as percentage of left ventricle (LV) on T2-weighted images [7]. Likewise, quantification of water content using a desiccation method followed bimodal wave, nonetheless, oedema did not normalise in 24 h, unlike T2 relaxation time [6]. This ‘residual’ water content is plausibly derived from persistent oedematous area, since its severity was reduced as measured using T2 mapping. Moreover, it raises the possibility that these previously applied CMR techniques may have failed to detect a relatively subtle persistent increase in water content after 24 h of reperfusion injury.

Native T1 mapping can characterise the injured myocardium and has been shown to characterise viable myocardium in the acute setting [8]. The increase in T1 value is secondary to an increase in myocardial free water content in response to an acute ischaemic insult [9, 10] More importantly, T1 mapping was demonstrated to be superior to T2 weighted imaging for the assessment of myocardial injury by identifying additional areas of abnormal myocardium in the setting of acute infarction and myocarditis [10, 11].

Here, we studied patients very early after reperfusion therapy and sought to investigate the temporal changes in AAR, quantified using a T1 mapping technique post STEMI. We hypothesised that bimodal changes in the intensity of oedematous reaction post ischemia/ reperfusion injury do not necessarily track with changes in the extent of the anatomical AAR (percentage of LV). This may provide a better reflection of actual changes in water content and would afford a better understanding of the application of CMR in assessments of acute myocardial injury.

A total of 48 patients were included in this study with total of 102 CMR exams. Of these, 31 patients underwent the 3 h hyper-acute CMR scan, followed by 24 h (31 patients) and 6 days (23 patients) CMR scans. These patients formed the temporal changes cohort. The remaining 17 patients underwent a single CMR scan at 24 h (Fig. 1). Two patients (1 from each cohort) had unevaluable T1 mapping and were excluded from the analysis. The clinical and CMR characteristics of the remaining 46 patients are shown in Tables 1 and 2. Importantly, there were no differences in clinical or procedural characteristics between two cohorts and, similarly, their CMR characteristics, spanning the whole LV, were comparable.

This is the first study to investigate the dynamic changes of highly sensitive T1 mapping immediately post STEMI. The main findings of this study can be summarised as follows: (1) after reperfusion therapy in patients presenting with STEMI, the intensity of oedematous response assessed using T1 mapping on 3 T CMR was not stable and followed bimodal pattern; (2) this bimodal reaction was still evident when excluding ‘T1 core’, which increased in area and, possibly, in prevalence, over time; (3) the extent of the anatomical AAR did not follow the same course as the T1 value and was not reduced after 24 h; and (4) LGE remained constant within the first week, unlike both MVO and IMH which peaked within the first 7 days.

Accurate quantification of myocardial area at risk is essential in calculating MSI, which has been used to assess efficacy of applied therapies in patients presenting with STEMI [5]. Timing of the ‘acute’ CMR scanning following acute myocardial infarction has lacked consistency based on the assumption that AAR assessed by standard T2W techniques is stable over the first one week and its quantification could be measured retrospectively, in part supported by our own earlier observations [21]. Recently, however concerns have been raised about the stability of the AAR and evidence from both animal and human studies have challenged this concept, suggesting that changes in myocardial oedema followed a bimodal pattern [6, 7] Carrick et al. [20, 22] proposed that deoxyhaemoglobin was exclusively responsible for the decrease in T2 relaxation time after 24 h. They suggested that these dynamic changes were a pure CMR phenomenon and did not reflect real changes within AAR [20, 22]. They proposed that changes within the AAR were only apparent in STEMI population with evidence of IMH on CMR. However, data from our study demonstrated that changes in the intensity of oedema post-STEMI followed a bimodal pattern. The extremely early (median 1.6 h) CMR time point adopted in the current study showed that the initial peak took place within the first three hours, a hyper-acute time point that was not previously available [20].

Recently, Fernández-Jiménez et al. [7] reported bimodal changes in humans replicating previously reported findings in an animal model [6]. Using multi-detector computed tomography (MDCT) to quantify AAR at the time of vessel occlusion, they noted that T2W had, in fact, underestimated AAR at 24 h following STEMI in the animal model. Furthermore, the same group reported that water content quantified on desiccation did not normalise, in contrast to the AAR delineated using T2W sequence in addition to T2 relaxation times. Both extent and severity of injured myocardium were reduced after 24 h to match pre-infarct measurements [6]. In other words, it appeared that the earlier CMR techniques were not sufficiently sensitive to detect residual oedema. We demonstrate here, for the first time, that changes in the anatomical area of oedema did not follow the same course as the intensity of the oedematous reaction measured using mapping technique (T1 relaxation time). This implies that the water content of a given volume of myocardium can change, somewhat like a sponge taking water, where changes in water content do not necessitate changes in the overall volume of the sponge. Moreover, we compared the difference between anatomical AAR determined by T1 mapping and T2W imaging and demonstrated that T2W underestimated the injured myocardium 24 h after STEMI. Taking all these observations together, these findings are consistent with the residual water content identified using desiccation method, which can be considered a gold standard. More importantly, it highlights the stability of AAR measured using T1 mapping and would allow the calculation of MSI within the first week.

An area with low T1 relaxation time, termed the ‘T1 core’, within injured myocardium showed some impact on the dynamic changes of AAR. T1 core has been shown to be a significant imaging biomarker to characterise infarcted myocardium and to predict future cardiovascular events [18]. We demonstrated that this area increases over time post STEMI and possibly in prevalence. Therefore, it would be anticipated that it may reduce the T1 relaxation time and possibly ameliorate the bimodal reaction by decreasing the “intensity” of oedematous reaction after 24 h. The significance relevance of expanded T1 core over time may need further assessments in prospective studies.

In this study, patients underwent LGE and T2* mapping sequences at three time points to assess if dynamic changes occur with the first week. Enhanced myocardium showed minimal changes overtime, unlike MVO which appears to peak within the first 24 h. Experimental studies using CMR coupled with blood flow measurement employing microspheres have demonstrated a similar course to our data [23]. Moreover, Robbers et al. [19] demonstrated that areas of MVO lack vascular integrity with higher number of vessels in the gadolinium enhanced myocardium compared to contrast-devoid core [19]. IMH, on the other hand, appears to peak at different timing from MVO. Our study demonstrated a significant increase in IMH after 24 h, which did not parallel MVO changes. Previous reports have illustrated that IMH peaks after 3 days, a time point that is not available in our study [7, 20]. Our data extend these findings by comparing MVO & IMH changes at extremely early time points. While MVO was already established by 3 h, IMH required longer period to reach its peak and lagged behind the no-reflow phenomenon. This adds additional insights to the intricate relationship between the extravasation of erythrocytes with IMH presence and MVO [19, 24] The early presence of IMH in some individuals may also reflect a more severe ischaemic insult or even worse reperfusion injury [19, 24, 25].

Our study in vivo in humans lacks histological / tissue validation to confirm findings using T1 mapping technique. We did not compare T2-weighted and T1 mapping at the hyper acute time point given the time constraint to perform lengthy CMR scan immediately after STEMI in potentially unstable patients. However, we validated our finding using a prospective cohort employing both T2-weighted imaging and T1 mapping sequences to measure AAR in humans. In addition, previous reports have demonstrated that T2-weighted imaging may underestimate AAR when compared to desiccation or MDCT [6, 7]. Finally, our cohort included both left anterior descending (LAD) and non-LAD infarcts and yet this did not influence the dynamic changes within the injured myocardium [7].

The intensity of the oedematous reaction post STEMI follows a bimodal pattern; while the extent of AAR does not track the same course and was stable over the first week. This discrepancy has important implications for use of CMR in this context and may explain the previously reported disagreement between oedema quantified by imaging and by tissue desiccation. We conclude that anatomical area at risk may best be quantified using T1 mapping, undertaken in the first week after reperfusion.