Absence of visible infarction on cardiac magnetic resonance imaging despite the established diagnosis of myocardial infarction by 4th Universal Definition of Myocardial Infarction

Abstract

Aims

Myocardial scarring due to acute myocardial infarction (AMI) can be visualized by late gadolinium enhancement (LGE) on cardiac magnetic resonance (CMR) imaging. However, a recent study revealed a group of Type 1 AMI patients with undetectable myocardial injury on LGE. This study aims to describe these cases in detail and explore possible explanations for this new phenomenon.

Methods and results

A total of 137 patients diagnosed with either ST-elevation myocardial infarction (STEMI) or non-ST-elevation myocardial infarction (non-STEMI) diagnosed according to the 4th Universal Definition of Myocardial Infarction underwent LGE-CMR after invasive coronary angiography. Fourteen of them (10.2%) showed no LGE and were included in the final study population. Most patients presented with acute chest pain, 3 patients were diagnosed as STEMI, and 11 as non-STEMI. Peak high-sensitive cardiac troponin T ranged from 45 to 1173 ng/L. A culprit lesion was identified in 12 patients. Severe coronary stenoses were found in five patients, while seven patients had subtotal to total coronary artery occlusion. Percutaneous coronary intervention was performed in 10 patients, while 2 patients required coronary artery bypass grafting and no intervention was required in 2 patients. Cardiac magnetic resonance was performed 30 (4–140) days after the initial presentation. Most patients showed preserved left ventricular ejection fraction on CMR. No alternative reasons for the rise/fall of high-sensitive cardiac troponin T were found.

Conclusion

The absence of LGE on CMR in patients with Type 1 AMI is a new finding. While insufficient spatial resolution of LGE imaging, delayed CMR performance, spontaneous reperfusion, and coronary collaterals may provide some explanations, further investigations are required to fully understand this phenomenon.

Graphical Abstract

Open in new tabDownload slide

Introduction

In acute myocardial infarction (AMI), the subtotal or total occlusion of a coronary artery and/or distal embolization of platelet-rich thrombus can result in myocyte necrosis, starting from the subendocardial layers and progressing to the more subepicardial layers.¹ As a result, myocardial cell death occurs which can be detected by diagnostic tools that differ regarding their sensitivities and specificities including cardiac troponins as the most sensitive and cardio-specific biomarker, cardiac imaging including cardiac magnetic resonance (CMR) imaging, echocardiography, ^99mTc-sestamibi-gated SPECT, as well as electrocardiogram (ECG) as the least sensitive and specific method.^2,3 By using extra-cellular contrast media like gadobutrol, it is possible to identify regions of irreversibly damaged myocardium following AMI through late gadolinium enhancement (LGE) in CMR.⁴ Late gadolinium enhancement-cardiac magnetic resonance has become a crucial imaging tool for risk stratification for cardiovascular events.⁵

Previous studies demonstrated a good correlation between cardiac troponin T levels and the extent of infarct mass measured by LGE-CMR in patients with both ST-elevation myocardial infarction (STEMI) and non–ST-elevation myocardial infarction (non-STEMI).^6,7 Recently, we investigated the correlation between single time point measurements of high-sensitivity cardiac troponin T (hs-cTnT) with myocardial infarct mass quantified by LGE-CMR in 137 patients who presented with AMI.⁷ Notably, we discovered that 14 of these patients with confirmed Type 1 AMI did not show any LGE, particularly no subendocardial LGE on their CMR scans. Consequently, we sought to explore the cases of AMI in which myocardial injury was absent on LGE-CMR and attempted to explain this finding.

Methods

Study population

In a recent study, 137 patients who had a confirmed diagnosis of Type 1 AMI per 4th version of the Universal Definition of Myocardial Infarction (UDMI)² and who underwent CMR examination within 6 months after discharge were retrospectively identified from our local clinical database at the Department of Cardiology, Angiology, and Pneumology of the Heidelberg University Hospital, Germany.⁷ In the previous study, patients qualified if admitted between 1 January 2018 and 31 December 2020, at our department with a diagnosis of STEMI or non-STEMI according to the 4th UDMI definition. Patients with AMI and an unexpected absence of LGE were identified and included in this retrospective analysis.

ST-elevation myocardial infarction diagnosis was made based on electrocardiographic presentations as proposed by the 2017 ESC guideline.⁸ Non-STEMI diagnosis was made using the 0/1-h protocol with hs-cTnT as proposed by 2015 and 2020 ESC guidelines.⁹ Cardiac troponin T was measured in plasma on a Cobas E411 and E602 using the hs-cTnT assay (Roche Diagnostics, Rotkreuz, Switzerland). Details of the assay used are shown in the Supplementary material online.

All patients underwent an evaluation including symptom assessment, physical examination, medical history, 12-lead ECG, standard blood tests, and invasive coronary angiography (ICA). The Rentrop classification was used to grade coronary collateral flow to the culprit lesion based on the angiogram findings.¹⁰ Echocardiographic findings were shown if they were available.

Ethical approval was waived due to the retrospective character of the study by the institutional ethics committee, and the study was conducted in agreement with the Declaration of Helsinki. All data were processed in an anonymized way.

Cardiac magnetic resonance imaging acquisition protocol

Standard CMR was performed supine in a 1.5-T Achieva, 1.5-T Ingenia CX, or 3-T Ingenia whole-body scanner (Philips Healthcare, Best, The Netherlands), with a commercial receiver coil as previously described.¹¹ Cine long-axis two-, three-, and four-chamber views and a short-axis cine stack were obtained.

Native myocardial T1 maps were obtained using a Modified Look-Locker Inversion recovery [5 s(3 s)3 s scheme] sequence acquired prior to contrast agent application. Seven to ten minutes after intravenous administration of gadobutrol (Gadovist™, Bayer HealthCare, Leverkusen, Germany), 0.14 mmol/kg body weight (1.5 T) or 0.1 mmol/kg body weight (3 T) contrast-enhanced Look-Locker imaging was acquired. Regions with LGE were verified in at least one other orthogonal plane being obtained as previously described.¹² Details of acquisition of images and post-processing are available in the supplementary appendix.

Late gadolinium enhancement images were evaluated visually independently by two blinded readers with high expertise in CMR (>2000 CMR). The average voxel size for our local LGE sequence was 0.85 × 0.85 × 10 mm (7.23 mm³).

Subgroup analysis

In a subgroup analysis, we assessed differences between patients with AMI without LGE and patients after AMI with LGE, who were previously investigated.⁷

Additionally, we assessed differences between patients who underwent LGE-CMR within 30 days after AMI and those who underwent LGE-CMR after the 30-day period.

Statistical analysis

Variables were tested for normal distribution using the D’Agostino–Pearson test. Parametric variables are given as mean ± standard deviation, and non-parametric variables as median with 25th and 75th percentiles. Variables with normal distribution were tested for statistical significance using the t-test, and non-parametric variables were analysed using the Mann–Whitney U test. Categorical variables were assessed using the χ² test. Statistical analyses were performed using the statistical software MedCalc Version 20.014 (MedCalc Software, Mariakerke, Belgium). A P-value <0.05 was considered statistically significant.

Results

Population characteristics

Of the previously investigated 137 patients, 14 patients (10.2%) fulfilled the diagnostic criteria for STEMI or non-STEMI and lacked LGE on CMR. The baseline characteristics of all patients are presented in Table 1. Eleven patients experienced typical chest pain, while one presented with a near syncope, one had a syncope, and one had palpitations. Most patients suffered from cardiovascular risk factors, with hypertension and hypercholesterolaemia being the most prevalent. Three patients revealed significant age- and sex-adjusted ST-segment elevations indicating STEMI (Cases 3, 5, and 9). The remaining 11 patients met the non-STEMI criteria with elevation of high hs-cTnT on admission and any later time point but before ICA and a relevant rise or fall of hs-cTnT (median of 23 ng/L absolute change and median of 39% relative increase of hs-cTnT) (Table 1).

In addition to increased hs-cTnT, at least one other feature was present in all cases, including symptoms suggestive of myocardial ischaemia, wall motion abnormalities on echocardiography, electrocardiographic deviations, or thrombus on ICA (Table 2).

Results of invasive coronary angiography

Of the 14 patients, 12 had an infarct-related coronary artery lesion, while 2 had no significant coronary artery disease (CAD) (Cases 1 and 4) (Table 2). Seven patients had a three-vessel disease, two had a two-vessel disease, and three had a single-vessel disease. Severe stenoses were diagnosed in five patients, while seven had subtotal to total coronary artery occlusion. Nine patients had low Rentrop collateral grades (0–1), while one subject (Patient 13) had Rentrop collateral Grade 2 (Table 2). A culprit lesion was identified in 12 patients. Percutaneous coronary intervention was performed in 11 patients and coronary artery bypass grafting (CABG) in 1 patient (Case 6). Figures 1 and  2 show representative cases of patients with STEMI (Figure 1; Supplementary material online, Figure S1) and non-STEMI (Figure 2; Supplementary material online, Figure S2) without detectable myocardial LGE.

Results of cardiac magnetic resonance imaging

The median time between the initial presentation and CMR was 30 (4–140; range 1− 211) days. Left ventricular ejection fraction (LVEF) was preserved in 13 patients with a mean of 64.0 ± 5.1%, while 1 patient (Case 6) had a reduced LVEF (41%). Left ventricular end-diastolic volume (LVEDV) was within normal range for most patients, except for Patients 1 and 6. Left ventricular (LV) mass was within normal range in all patients. Native T1 relaxation times were increased in three patients (Cases 6, 8, and 10). Inducible myocardial ischaemia was observed in three patients (Cases 8, 13, and 14). Pathological findings on CMR are shown in Table 3. Assuming a myocardial density of 1.055 g/mL (18), 7.6 mg myocardium is represented by one voxel.

Comparison between patients with and without late gadolinium enhancement after myocardial infarction

We compared the 14 patients without LGE with the 123 patients with LGE on CMR after AMI. The results are presented in Table 4. Patients without LGE had significantly lower hs-cTnT levels upon admission in comparison with patients with LGE [113 (38–189) ng/L vs. 222 (72–1059) ng/L, P = 0.042]. Also, peak hs-cTnT levels were significantly lower in patients without LGE compared with patients with LGE [226 (120–653) ng/L vs. 1567 (456–3245) ng/L, P < 0.001]. Interestingly, patients without LGE showed a significantly longer symptom-to-admission time [>6 h: without LGE: 6 patients (43%), with LGE: 17 patients (14%), P = 0.006].

The time between presentation and ICA was similar in both groups [without LGE: 2.0 (0.5–15.0) h, with LGE: 2.0 (0.5–6.0) h, P = 0.275]. Although not statistically significant, there were more patients with STEMI in the LGE group (41% vs. 21%). Percutaneous coronary intervention was similarly often performed in both groups (without LGE: 79%, with LGE: 88%, P = 0.334). Patients without LGE showed significantly better LVEF compared with patients with LGE [62.5 (56–68)% vs. 53.0 (45–62)%, P = 0.013]. LVEDV and LV mass were reduced in patients without LGE compared with patients with LGE (both P < 0.0001). Further details of the comparison between the two groups are shown in the supplementary appendix.

Comparison between early and late cardiac magnetic resonance imaging after myocardial infarction

Patients who underwent LGE-CMR 30 days after AMI showed a significantly higher peak hs-cTnT level compared with patients who received LGE-CMR within 30 days of AMI (CMR within 30 days of AMI: 164 ± 111 ng/L; CMR after 30 days of AMI: 588 ± 410 ng/L; P < 0.05). The time between the initial presentation and the ICA was significantly shorter for patients who underwent LGE-CMR 30 days after AMI [CMR within 30 days of AMI: 15 (2.3–32); CMR after 30 days of AMI: 0.5 (0.5–2.0); P < 0.05] (see Supplementary material online, Table S1). Additional information of the comparison between the two groups is shown in the supplementary appendix.

Discussion

This study provides detailed descriptions of 14 cases of AMI defined by the 4th UDMI, in which there was no visualization of infarction on CMR using LGE. It has been established that subendocardial LGE after AMI is independently associated with major adverse cardiovascular events.¹³ Our group has previously demonstrated a good correlation between LGE-CMR infarct mass and hs-cTnT levels at admission, as well 24, 48, 72, and 96 h after AMI.⁷ However, in the present study, we observed 14 patients who exhibited significant peak hs-cTnT values following AMI but did not show any LGE including subendocardial LGE on CMR as an indicator of myocardial injury. We attempt to provide explanations for this observation in the following section.

As presented above, a voxel-to-myocardial mass ratio 7.6 mg per voxel is established. However, in our experience, at least 20 voxels of LGE are required for an experienced investigator to visualize the infarcted myocardium, which translates to a minimum of 152 mg of infarcted myocardium detectable by LGE-CMR. Previous studies have suggested that at least 200 mg of infarcted myocardium is necessary to be visualized by LGE.¹⁴ In another study, a strong linear correlation between cTnT and myocardial mass was established.¹⁵ The authors discuss that myocardial necrosis of just 25 mg can exceed the 99th percentile concentration of hs-cTnT. Thus, even small amounts of infarcted myocardium can lead to slight increases in cTnT concentration, which may be less than what is visible on LGE-CMR. Assuming subendocardial LGE requires 152–200 mg of infarcted myocardium, equivalent to 85–112 ng/L hs-cTnT, for infarct visualization, two non-STEMI patients (Cases 4 and 12) in our study had a lower peak hs-cTnT than this threshold. One of the patients had significant CAD excluded, while the other received PCI due to a subtotal occlusion of RCA. However, most patients had a significantly higher peak hs-cTnT (median 226 ng/L) than the threshold (up to 10 times), with some having subtotal occlusion of a coronary artery for a considerable time until revascularization. Therefore, LGE-CMR should have detected the infarcted myocardial mass in these patients. However, other studies report a threshold of at least 1 g of infarcted myocardium, corresponding to 560 ng/L hs-cTnT, for visualization of the infarcted area using LGE-CMR.^15,16 Those high hs-cTnT concentrations were observed in four patients (Cases 3, 5, 9, and 13). Even assuming a low sensitivity in detecting LGE, it should have detected a myocardial scar in these cases.

Novel free-breathing high-resolution LGE imaging with an improved spatial resolution of 1.25 × 1.25 × 2.5 mm demonstrated a significant improvement in the diagnostic work-up of patients with MINOCA.¹⁷ However, the clinical relevance and prognostic implications of high-resolution LGE imaging in patients with AMI are currently unknown.

Global as well as regional native T1 relaxation times were increased in three patients. Previous studies have shown a strong correlation between native T1 relaxation time and histological collagen volume fraction, indicating myocardial fibrosis.¹⁸ All three patients also showed hypertrophy of the LV, with elevated native T1 relaxation time correlating with hypertrophic segments, as demonstrated in patients with hypertrophic cardiomyopathy and hypertension.¹⁹ Therefore, we assume that elevated native T1 time is most likely due to diffuse non-ischaemic myocardial fibrosis. However, it is also important to consider that changes in native myocardial T1 values can be attributed to intravascular alterations, such as coronary microvascular dysfunction.²⁰

In recent years, there have been significant advancements in dark-blood LGE techniques which improve the scar-to-blood contrast.²¹ In dark-blood LGE, the inversion time is set to null the LV blood pool, rather than nulling the viable myocardium as inconventional bright-blood LGE. This results in a dark grey appearance of the LV blood pool while keeping the subendocardial LGE bright in appearance. This approach has led to an increase in the scar-to-blood contrast-to-noise ratio.^21,22 When directly compared with conventional bright-blood LGE, dark-blood LGE detected significantly more cases with ischaemic scar tissue and significantly increased scar burden.²² Consequently, dark-blood LGE appears to be a promising non-invasive method for detecting ischaemic myocardial scar tissue, even in small infarct areas. However, during the study period, dark-blood LGE sequences were not integrated in the regular clinical CMR protocols.

When comparing patients post-AMI in the absence of LGE with patients with LGE, symptom onset to hospital admission was significantly longer. Additionally, hs-cTnT levels were significantly lower, indicating fewer myocardial injuries in patients lacking LGE after AMI. However, PCI was still required in a similar proportion of cases (79%), suggesting a significant culprit lesions with total or subtotal occlusion. Nonetheless, the group with LGE had more cases with STEMI characterized by noteworthy periods of no-flow, leading to a potential development of substantial scar tissue. The presence of subendocardial LGE was associated with significant reduced cardiac function and LV dilation compared with patients lacking LGE.

One possible explanation for the absence of LGE could be the late timing of LGE-CMR (median of 30 days, range from 1 to 211 days). The time between presentation and CMR, although statistically not significant, was longer in patients without LGE compared with patients with LGE. However, in the subgroup of patients without LGE and CMR performed 30 days after AMI, hs-cTnT levels on admission were similar to patients with LGE. Therefore, at least a small subendocardial LGE should have been detected in this subgroup. In an experimental study using dogs, subendocardial LGE was observed just 34 min after the onset of ischaemia induced by coronary artery occlusion.²³ As patients with AMI suffer from subtotal occlusion of coronary arteries resulting in ischaemia for sometimes hours, LGE should have detected areas of relevant myocardial injury.

Moreover, in the DANAMI-3-DEFER CMR substudy, infarct size was reduced within the first 3 months after STEMI.²⁴ Hence, a delayed timing of LGE-CMR in some cases might be a reason for the absence of LGE in AMI cases. Nevertheless, undetectable LGE also occurred in patients with LGE-CMR performed 1–2 days after AMI, suggesting that delayed LGE-CMR cannot fully account for this phenomenon. Patients who were investigated 30 days after the initial presentation in our study were more likely to have suffered from STEMI with a higher release of hs-cTnT. They should have had a larger infarct mass even though infarct size may decline within the first weeks after the event.

Furthermore, previous findings showed that the timing of LGE image acquisition after contrast administration seems to be crucial when evaluating LGE in patients after AMI. The largest LGE volume was measured between minute 1 and 3 post-injection. Also, LGE volume decreased over time when CMR was performed shortly after AMI but remained constant when CMR was performed months after AMI.²⁵ However, a standardized LGE image acquisition protocol was applied to all patients in our study with capturing images 7–10 min post-injection. Hence, the time point of LGE image acquisition after contrast injection might only be a minor confounder.

Another possible explanation for this phenomenon is spontaneous reperfusion of occluded coronary arteries. The timing of revascularization in patients with aborted STEMI was evaluated in a previous study using LGE-CMR to measure myocardial infarct size.²⁶ Interestingly, in patients with delayed revascularization after 22.7 (interquartile range 18.2–27.3) h, a third of the cases had no infarcted myocardial mass. One explanation might be the routine use of antiplatelet therapy to reduce thrombus load. While spontaneous revascularization may explain our findings to a certain extent, some patients still suffered from subtotal coronary artery occlusion causing no-flow or slow flow, which should have resulted in visible myocardial injury on LGE-CMR.

Furthermore, coronary collateral vessels may have provided sufficient blood supply to the area of myocardial ischaemia caused by the culprit lesion. However, coronary collateral and retrograde flow were only visible in a minority of patients, and only Patient 13 showed partial epicardial filling by collateral vessels (Rentrop 2). Collateral and retrograde collateral flow onto the culprit lesion of AMI is inversely associated with infarct size.²⁷ Very small collaterals not visible on ICA may be a reason for undetectable myocardial injury in AMI.

Additionally, several studies have identified a significant increase in cTn levels without myocardial cell death.²⁸ This could be due to alterations in cell membrane permeability caused by mechanical stress or injury-induced cell wounds. The increase in cell membrane permeability is believed to be a result of membrane repair processes aimed at preventing the loss of cardiomyocytes.²⁹ However, the mechanism needs further experimental and clinical confirmation.

As previously shown, in up to 26% of cases where patients meet the diagnostic criteria for MINOCA, CMR detected no significant LGE.³⁰ Importantly, we adhered strictly to the 4th UDMI criteria, performed ICA including a meticulous classification of collaterals and found no alternative reasons for the acute myocardial injury suggested by an elevation of hs-cTnT with a rise and/or fall of hs-TnT in the 14 patients on CMR besides AMI, indicating that this phenomenon seems new in patients with Type 1 AMI.

Limitations

The number of patients without LGE after AMI is relatively small, comprising only 14 individuals from a single centre out of more than 137 patients with AMI who underwent LGE-CMR. This limited sample size, coupled with the primarily descriptive approach of our study, is a limitation in fully explaining the phenomenon of absence of subendocardial LGE in patients with AMI. Furthermore, our findings suggest that the absence of subendocardial LGE in patients with AMI is a rare phenomenon.

It is important to note that patient selection in our study was heterogenous, ranging from patients without significant CAD to those with severe CAD and CABG. In future studies, only patients receiving PCI should be included to provide a more homogeneous study population. Additionally, quantifying the extracellular volume fraction could be useful in determining myocardial fibrosis. The time between AMI and LGE-CMR exceeded 30 days in 50% of the cases. The extended time and variability of timing between AMI and LGE-CMR are notable limitations. To address these issues, we compared patients who underwent LGE-CMR within 30 days of AMI with those who underwent LGE-CMR after the 30-day period. However, the small number of patients in each group is a clear limitation. Further studies with larger groups are essential to assess the impact of the timing on the visualization of potential subendocardial LGE more comprehensively.

Conclusion

This study reveals a new phenomenon of the absence of LGE on CMR in patients diagnosed with STEMI and non-STEM according to the 4th UDMI definition. Although insufficient spatial resolution of LGE-CMR, delayed image acquisition, anti-thrombotic therapy-induced spontaneous reperfusion, and possible untraceable collaterals on the ICA offer some explanation, they only partially justify our observation. Therefore, the pathophysiology of undetectable myocardial scar tissue in patients with AMI remains unclear and requires further investigation.

Supplementary material

Supplementary material is available at European Heart Journal: Acute Cardiovascular Care online.

Acknowledgements

We thank Mrs Heidi Deigentasch, Mrs Amelie Werner, and Mrs Elisabeth Mertz for their valuable support regarding data management. We thank our technologists Mr Daniel Asmussen, Mrs Katharina Boesenberg, Mrs Miriam Hess, and Mrs Alexandra Jeck for image acquisition.

Funding

J.S. was funded by the German Centre for Cardiovascular Research (DZHK—Deutsches Zentrum für Herz-Kreislauf-Forschung).

Data availability

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Reference

Author notes