The MRI characteristics of the no-flow region are similar in reperfused and non-reperfused myocardial infarcts: an MRI and histopathology study in swine

The clinical management of myocardial infarction (MI) by early reperfusion therapy aims to reduce the extent of the ischaemic injury and to prevent the long-term consequences of MI [1]. However, Braunwald and Kloner assert that reperfusion may also result in reperfusion injury, i.e., the death of potentially salvageable myocardium [1, 2]. Reperfusion injury has been defined as reperfusion-related expansion of ischaemic cardiac injury resulting in decreased contractility, increased arrhythmogenic effects, irreversible conversion of reversible myocyte injury, and microvascular obstruction (MO) [1, 3].

The presence of MO seems to play a major role in long-term outcome after reperfusion post MI [4‐11]. MO has been described as the strongest predictor of increase over time in left ventricle (LV) end-diastolic and end-systolic volumes post MI [4]. Patients with MO have been shown to have a significantly larger MI than patients without MO [5]. MO detected one week after the onset of MI has been suggested as an independent predictor of the size of MI at one year follow up and was found to be associated with adverse infarct healing, adverse LV remodelling, increased LV volumes, and lower ejection fractions [10]. Patients with MO have been reported as having more cardiovascular events than those without it [12]. Even independently of MI size, the presence of MO may remain a prognostic marker of post-MI complications [11]. Thus, as mentioned previously, the presence and/or the size of the MO might be one of the strong prognostic parameters after reperfusion post MI [4, 5, 8‐11, 13].

MO has been defined as a product of reperfusion injury [1]. It has been shown, however, that the very same magnetic resonance imaging (MRI) phenomenon, i.e. central hypo-enhancement within the infarcted area, also occurs in MI without reperfusion [14, 15]. In non-reperfused MI, with no possibility of reperfusion injury, the underlying pathophysiologic condition of the similar MRI phenomenon might be different. Thus, to prevent any confusion when referring to the unenhanced central area within the MI observed by MRI, we use the term “no-flow region” (NF) instead of “microvascular obstruction”. In this study, we aimed to compare the MRI and histopathologic characteristics of the NF in non-reperfused and reperfused MI.

All animals included in this study survived until day 4 and underwent MRI. The NF was detected in all 14 swine involved in this study. Representative long-axis EGE images (performed 2 min after CA administration) and LGE images (performed 15 min after CA administration) from both experimental groups are shown in Fig. 3.

LVM, MI, and NF measurements obtained by MRI and TTC are shown in Table 1. There was good agreement between MRI and TTC for the measurement of average LVM in each model (all p > 0.077). There were no significant differences in mean MI size measured by MRI versus TTC, either in the reperfused group (p = 0.216), or in the non-reperfused group (p = 0.484). Furthermore, there was no significant difference in the mean MI size in the reperfused and non-reperfused groups (p = 0.090). The mean NF size, and its values normalised as a percentage of MI or LVM measured by EGE MRI, was not significantly different in the two experimental groups (all p > 0.114).

T2 relaxation times measured in the remote, MI, and NF area are reported in Table 2. No significant difference was observed between the groups, or between the remote and the NF area. T2 was significantly increased in the infarct area, i.e., the area showing LGE at 15 min post CA injection, in comparison to both the remote and NF areas (all P < 0.001).

The volume of the NF gradually decreased with time after CA administration in both models (Fig. 4). NF measurements taken at and after 20 min were significantly different from the size of NF obtained in EGE in the reperfused model (P < 0.001), and there was a similar, significant reduction in the NF at 25 min in the non-reperfused model (P < 0.001). The NF was still observed in both models at 45 min after CA administration. The size of the NF, however, was larger in the non-reperfused group (3.05 ± 1.20 versus 1.82 ± 0.91 ml, respectively, p < 0.001), indicating different CA uptake kinetics in the two models. The size of the MI (the enhanced and the NF area combined) did not change significantly during the imaging period in either of the two models. Representative images including both EGE and LGE images are shown in Fig. 5.

Macroscopic and microscopic histopathologic evaluation of representative reperfused and non-reperfused MI is shown in Fig. 6. In the haematoxylin-eosin slides, widespread myocyte necrosis (nuclear shrinkage and nuclear loss), oedema (indicated by wide interstitium), and inflammatory cell infiltration (polymorphonuclear, considered neutrophils) and mononuclear (considered lymphocytes and macrophages) were observed in the central MI regions (corresponding to MRI NF areas) in both models. Using Prussian blue and Von Kossa staining, notable differences were detected between the reperfused and non-reperfused model. In reperfused MI, Prussian blue showed a relevant amount of iron deposits (extensive blue spots) throughout the whole MI area, but at the highest concentration in the border zone of the MI. Iron deposits were not detected at all in the core area of the non-reperfused MI. However, some iron-related blue spots were observed sporadically in the border zone of the non-reperfused MI. In reperfused MI, Von Kossa staining suggested that the periphery of the MI was strongly positive for calcium, while the inner core was negative. Calcium deposits were observed in the border zone of the non-reperfused MI as well, but the amount of the calcium was clearly smaller and less widespread than in the reperfused MI.

In this study we qualitatively investigated the MRI CA uptake characteristics of the NF area in non-reperfused and reperfused MI on day 4 after experimental induction of MI. Based on previous experience [15], we expected emergence of an NF area in the majority of the non-reperfused animals. The literature indicates that the NF is detected in human reperfused MI in around 68% of patients [19] and our past experience of experiments in reperfused swine suggests that the NF can be expected in at least two thirds of animals. In this study, all reperfused animals presented with a significant NF on day 4 post MI.

Our study indicates that the NF size strongly depends on the time elapsed after the administration of the CA. Of note, with time the NF was able to take up the CA even in animals with non-reperfused MI. According to the current definition, the NF should be measured during the early CA kinetics [20]. Due to the significant CA uptake into the NF, this protocol should indeed be followed, as evaluating the NF 20–25 min later in the post-contrast period would result in notably larger enhanced MI regions and smaller NF areas. As previous reports highlight the prognostic value of the NF, its accurate quantification is even more relevant [21].

The reason for CA uptake in the reperfused animals may be explained by the observation that ischaemia/reperfusion injury does not really cause total occlusion of the microvascular system [1]. Apparently, blood flow in the area of the early NF is not completely absent but rather very low [1]. The NF first appears centrally in the infarct core extending toward the epicardium over time. Myocardial blood flow progressively decreases within 2 to 3 hours after reperfusion, resulting in a twofold increase in the area of the NF. A further increase in NF size has been demonstrated up to 48 hours after initiation of reperfusion [1]. In contrast, myocardial blood flow in non-reperfused MI is completely absent. Nevertheless, as our results indicate, the NF area in non-reperfused MI also takes up CA, which has also been reported in prior studies.

Wang et al. have shown that the NF in non-reperfused MIs can be detected by EGE and LGE imaging up to 22–32 min after CA administration. It then disappears as the area defined as the NF on EGE images becomes enhanced [14]. Our study partially supports this observation, but it also indicates that the NF in non-reperfused MI can be detected beyond approximately 30 min and even at 45 min after the administration of CA.

Assuming total coronary artery occlusion in the non-reperfused MI model, the only possible mechanism for CA uptake is slow interstitial diffusion from the surrounding tissues that are still well-supplied [14]. The rim of the MI is highlighted even as early as the time of the EGE acquisition, presumably by CA from intact collaterals. This rim area of the MI is usually composed of a mixture of viable and nonviable myocytes [22]. The collateral circulation supplying this peri-infarct territory keeps the myocardial tissue partially alive, and serves as a CA source as well. The peri-infarct area, however, also contains vessels with damaged endothelial lining, allowing the accumulation of the CA in the extracellular space. This extracellular space may serve as a pool to supply the adjacent infarct area, including the NF, with CA. Interestingly, the reperfused and the non-reperfused MIs accumulate CA at a similar rate. The latter observation may confirm the hypothesis that most of the CA uptake is through slow interstitial diffusion in both MI models. If the blood flow is not completely absent in the reperfused case, as suggested by others [20], a vascular component may contribute to the access to CA. This flow, however, may be extremely limited, as the rate of contrast uptake in both models seems similar.

In this project we studied the presence of hemosiderin components by T2 relaxation rate measurements. Although T2* mapping is the most widely used technique for the detection of haemorrhage, T2 mapping has also been established for such assessment [23]. Our results indicate that T2 measured in the NF was not significantly different from the T2 observed in the remote myocardium and also did not differ between the models. Although this observation seems unexpected given that our histopathologic studies showed extensive hemosiderin deposition in the reperfused but not in the non-reperfused NF, it might be explained by the possible overlap of oedema (causing T2 increase) and haemorrhage (causing T2 decrease) resulting in an apparent net zero T2 change compared to normal tissue in the reperfused model. On the other hand, an absence of both oedema and haemorrhage would also result in an unchanged T2 in the non-reperfused model.

Another histopathologic difference between the two models was the presence of calcium deposits in the NF, which may also contribute to T2 shortening. Calcium deposition is a component of reperfusion injury [24], and its presence only in the reperfused animals further supports the successful total occlusion of the coronary arteries in the non-reperfused animals.

NF in non-reperfused and reperfused MIs have similar characteristics on MRI despite the different pathophysiologic and underlying histopathologic conditions, indicating that the presence of the NF alone cannot differentiate between these two types of MI.