Prognostic significance of infarct core pathology revealed by quantitative non-contrast in comparison with contrast cardiac magnetic resonance imaging in reperfused ST-elevation myocardial infarction survivors

Abstract Aims To assess the prognostic significance of infarct core tissue characteristics using cardiac magnetic resonance (CMR) imaging in survivors of acute ST-elevation myocardial infarction (STEMI). Methods and results We performed an observational prospective single centre cohort study in 300 reperfused STEMI patients (mean ± SD age 59 ± 12 years, 74% male) who underwent CMR 2 days and 6 months post-myocardial infarction (n = 267). Native T1 was measured in myocardial regions of interest (n = 288). Adverse remodelling was defined as an increase in left ventricular (LV) end-diastolic volume ≥20% at 6 months. All-cause death or first heart failure hospitalization was a pre-specified outcome that was assessed during follow-up (median duration 845 days). One hundred and sixty (56%) patients had a hypo-intense infarct core disclosed by native T1. In multivariable regression, infarct core native T1 was inversely associated with adverse remodelling [odds ratio (95% confidence interval (CI)] per 10 ms reduction in native T1: 0.91 (0.82, 0.00); P = 0.061). Thirty (10.4%) of 288 patients died or experienced a heart failure event and 13 of these events occurred post-discharge. Native T1 values (ms) within the hypo-intense infarct core (n = 160 STEMI patients) were inversely associated with the risk of all-cause death or first hospitalization for heart failure post-discharge (for a 10 ms increase in native T1: hazard ratio 0.730, 95% CI 0.617, 0.863; P < 0.001) including after adjustment for left ventricular ejection fraction, infarct core T2 and myocardial haemorrhage. The prognostic results for microvascular obstruction were similar. Conclusion Infarct core native T1 represents a novel non-contrast CMR biomarker with potential for infarct characterization and prognostication in STEMI survivors. Confirmatory studies are warranted. ClinicalTrials.gov identifier NCT02072850.


Introduction
Myocardial infarct size 1,2 and microvascular obstruction 3 -5 revealed by contrast-enhanced cardiac magnetic resonance (CMR) reflect the efficacy of reperfusion therapy and are prognostically important findings in survivors of ST-elevation myocardial infarction (STEMI).
Human tissue has fundamental magnetic properties, including the longitudinal (spin-lattice) relaxation time (native T1 in milliseconds). Native T1 is influenced by water content, binding with macromolecules (water mobility), and cell content. 6,7 Native T1 CMR does not involve an intravenous contrast agent. Tissue water content increases as a result of ischaemia and longer T1 times may represent a biomarker of localized myocardial injury. 8 -16 The clinical significance of tissue changes within the infarct core in patients with acute reperfused STEMI has not been directly assessed. We hypothesized that baseline native T1 values would be (i) inversely associated with the severity of MI, including microvascular obstruction, (ii) independently associated with left ventricular (LV) remodelling, and (iii) independently associated with predefined health outcomes. Should these hypotheses be confirmed then infarct core native T1 mapping without an intravenous contrast agent might have potential as an alternative biomarker to microvascular obstruction revealed by contrast-enhanced CMR.
To investigate these hypotheses, we measured native T1 in myocardial regions of interest in STEMI patients undergoing serial CMR imaging 2 days and 6 months post-MI. We assessed the clinical associates of native T1 within the hypo-intense infarct core and subsequent LV remodelling and examined its association with all-cause death and first hospitalization for heart failure.

Study population and ST-elevation myocardial infarction management
We performed an observational prospective CMR cohort study in a single regional cardiac centre between 14 July 2011 and 22 November 2012. Three hundred and forty three STEMI patients provided written informed consent to undergo CMR 2 days and 6 months post-MI. Patients were eligible if they had an indication for primary percutaneous coronary intervention (PCI) or thrombolysis for acute STEMI due to a history of symptoms consistent with acute myocardial ischaemia and with supporting changes on the electrocardiogram (ECG) (i.e. ST-segment elevation or new left bundle-branch block). 17 Exclusion criteria represented standard contra-indications to contrast CMR, including a pacemaker and estimated glomerular filtration rate ,30 mL/min/1.73 m 2 . The study was approved by the National Research Ethics Service and all participants provided written informed consent. Acute STEMI management followed contemporary guidelines. 17,18 Aspiration thrombectomy, direct stenting, anti-thrombotic drugs, and other therapies were administered according to clinical judgment (Supplementary material online, Methods). The ClinicalTrials.gov identifier is NCT02072850.

Cardiac magnetic resonance acquisition
Cardiac magnetic resonance was performed on a Siemens MAGNETOM Avanto (Erlangen, Germany) 1.5-Tesla scanner with a 12-element phased array cardiac surface coil. 19 The imaging protocol included cine magnetic resonance imaging with steady-state free precession (SSFP), native T1 mapping, 15,20 T2 mapping, 21,22 T2*-mapping, and delayed-enhancement phase-sensitive inversion-recovery pulse sequences. 23 The scan acquisitions were spatially co-registered and also included different slice orientations to enhance diagnostic confidence. Cardiac magnetic resonance was also performed in 50 healthy volunteers of similar age and gender in order to obtain local reference values for myocardial native T1 (Supplementary material online). Patients and healthy volunteers underwent the same imaging protocol except that healthy volunteers ,45 years did not receive gadolinium. The coefficients of variation for native T1 were also measured (Supplementary material online, Results).
Native T1 maps were acquired in three short-axial slices (basal, mid, and apical), using an optimized modified look-locker inversion recovery (MOLLI) T1-mapping investigational prototype sequence 15,20 before contrast administration (Supplementary material online, Methods; work-in-progress 448, Siemens Healthcare). The MOLLI T1 cardiac-gated acquisition involved three inversion recovery prepared Look-Locker experiments combined within one protocol. 15 The CMR parameters were: bandwidth 1090 Hz/pixel, flip angle 358, echo time (TE) 1.1 ms, T1 of first experiment 100 ms, TI increment 80 ms, matrix 192 × 124 pixels, spatial resolution 2.2 × 1.8 × 8.0 mm, slice thickness 8 mm, and scan time 17 heartbeats. The prototype pulse sequence did not involve motion correction.

Cardiac magnetic resonance analyses
The images were analysed on a Siemens work-station by observers with at least 3 years CMR experience (N.A., D.C., I.M, and S.R.). All of the images were reviewed by experienced CMR cardiologists (C.B. and N.T.). Left ventricular dimensions, volumes, and ejection fraction were quantified using computer-assisted planimetry (syngo MR w , Siemens Healthcare, Erlangen, Germany). The late gadolinium enhancement images were analysed for infarct size and microvascular obstruction by observers (N.A. and I.M.) who were blinded to all of the other data. In healthy volunteers, the absence of late gadolinium enhancement was determined qualitatively by visual assessment.

Native T1 mapping: standardized measurements in myocardial regions of interest
Native T1 mapping is a CMR method providing a parametric colourencoded anatomical map in which the T1 value is encoded in each pixel 24 (Figure 1). The native T1 map analyses were informed by contemporary CMR guidelines. 24 Left ventricular contours were delineated with computer-assisted planimetry on the raw T1 image and copied onto the colour-encoded spatially co-registered map. Apical segments were not included because of partial volume effects. Particular care was taken to delineate regions of interest with adequate margins of separation from tissue interfaces prone to partial volume averaging such as between myocardium and blood. 19,24,25 Each T1 map image was assessed for the presence of artefacts relating to susceptibility effects, or cardio-respiratory motion. Each colour map was evaluated against the original images. When artefacts occurred the affected segments were not included in the analysis.
In STEMI patients, myocardial T1 values were segmented spatially and regions of interest were defined as (i) remote myocardium, (ii) injured myocardium, and (iii) infarct core. The regions of interest were planimetered to include the entire area of interest with distinct margins of separation from tissue interfaces to avoid partial volume averaging. The remote myocardium region of interest was defined as myocardium 1808 from the affected zone with no visible evidence of infarction, oedema, or wall motion abnormalities (assessed by inspecting corresponding contrast-enhanced T1-weighted, T2-weighted, and cine images, respectively). The infarct zone region of interest was defined as myocardium with pixel values (T1 or T2) .2 SD from remote myocardium on T2-weighted CMR. 21,22 The hypo-intense infarct core was defined as an area in the centre of the infarct territory having a mean T1 value of at least 2 standard deviations (SDs) below the T1 value of the periphery of the area at risk. 21,22 The assessment of T1 maps and adjudication (present/absent) of a hypo-intense core was performed independently by D.C.
In healthy volunteers, the mid-ventricular T1 colour-encoded map was segmented into six equal segments, using the anterior right ventricular-LV insertion point as the reference point. 26 T1 was measured in each of these segments, and regions of interest were planimetered distinct and separate from blood-pool and tissue interfaces. These segmental values were also averaged to provide one value per subject. Results are presented as average values for segments and slices.

Infarct definition and size
The presence of acute infarction was established based on abnormalities in cine wall motion, rest first-pass myocardial perfusion, and delayedenhancement imaging in two imaging planes. In addition, supporting changes on the ECG and coronary angiogram were also required. Acute infarction was considered present only if late gadolinium enhancement was confirmed on both the axial and long-axis acquisitions. The myocardial mass of late gadolinium (grams) was quantified using computerassisted planimetry and the territory of infarction was delineated using a signal-intensity threshold of .5 SDs above a remote reference region and expressed as a percentage of total LV mass. 27 Infarct regions with evidence of microvascular obstruction were included within the infarct area and the extent of microvascular LV ventricular mass was also measured. The measurements of infarct size were performed by I.M. and N.A.

Microvascular obstruction
Microvascular obstruction was defined as a dark zone on EGE imaging 1, 3, 5, and 7 min post-contrast injection that remained present within an area of LGE at 15 min. Identification of microvascular obstruction was performed independently by I.M. and N.A.

Area-at-risk
Area-at-risk was defined as LV myocardium with pixel values (T2) .2 SDs from remote myocardium. 4,21,22,28 -30 In order to assess the area-at-risk, the epi-and endocardial contours on the last corresponding T2-weighted raw image with an echo time of 55 ms were planimetered. 21 Contours were then copied to the computed T2 map and corrected when necessary by consulting the SSFP cine images.

Myocardial salvage
Myocardial salvage was calculated by subtraction of per cent infarct size from per cent area at risk. 4,30,31 The myocardial salvage index was calculated by dividing the myocardial salvage area by the initial area at risk.

Adverse remodelling
Adverse remodelling was pre-defined as an increase in LV end-diastolic volume ≥20% at 6 months from baseline. 3

Myocardial haemorrhage
On the T2* maps, a region of reduced signal intensity within the infarcted area, with a T2* value of ,20 ms 32 -35 was considered to confirm the presence of myocardial haemorrhage.

Electrocardiogram
A 12-lead ECG was obtained before coronary reperfusion and 60 min afterwards. The extent of ST-segment resolution on the ECG assessed 60 min after reperfusion compared with the baseline ECG before reperfusion 17 was expressed as complete (≥70%), incomplete (30% to ,70%), or none (≤30%).

Laboratory analyses
The acquisition of the ECGs and blood samples for biochemical and hematologic analyses are described in Supplementary material online, Methods.
D. Carrick et al.

Pre-specified health outcomes
We pre-specified adverse health outcomes that are pathophysiologically linked with STEMI. The primary composite outcome was (i) all-cause death or first heart failure hospitalization (Supplementary material online, Methods). Other health outcomes included major adverse cardiac events (MACEs) defined as cardiac death, non-fatal MI, or hospitalization for heart failure.
Research staff screened for events from enrolment by checking the medical records and by contacting patients and their primary and secondary care physicians, as appropriate with no loss to follow-up ( Figure 2). Each serious adverse event (SAE) was reviewed by a cardiologist who was independent of the research team and blinded to all of the clinical and CMR data. The SAEs were defined according to standard guidelines 36,37 (Supplementary material online, Methods) and categorized as having occurred either during the index admission or postdischarge. All study participants were followed up for a minimum of 18 months after discharge. The median duration of follow-up was of 845 days [post-discharge censor duration (range) 598 -1098 days].

Statistical analyses
The sample size calculation is described in Supplementary material online, Methods. We estimated that at least 30 MACE events would occur based on a conservative estimate of the event rate (10 -12%) at 18 months.
Categorical variables are expressed as number and percentage of patients. Most continuous variables followed a normal distribution and are Figure 1 Three patients with acute ST-elevation myocardial infarction treated by primary PCI and with the same anti-thrombotic therapies, including aspirin, clopidogrel, heparin, and intravenous tirofiban. Each patient had normal thrombolysis in myocardial infarction Grade 3 flow at the end of PCI. Cardiac magnetic resonance imaging was performed for each patient 2 days later. (A) Patient with no T1 hypo-intense infarct core and no microvascular obstruction. Native T1 within the injury zone (middle) measured 1211 ms. Acute infarct size revealed by late gadolinium enhancement (right) was 22.2%. The left ventricular ejection fraction and left ventricular end-diastolic volume were 55.2% and 143.1 mL, respectively. Analysis of the repeat magnetic resonance imaging scan after 6 months follow-up indicated that the final infarct size was 15.6% of left ventricular mass and the left ventricular end-diastolic volume had reduced to 103.0 mL. This patient had an uncomplicated clinical course. (B) Patient with both T1 hypo-intense infarct core and microvascular obstruction. T1 mapping (middle) revealed a hypo-intense region within the infarct core, corresponding to the area of microvascular obstruction on contrast-enhanced magnetic resonance imaging (right). Native T1 within the infarct core measured 1036 ms, which was substantially lower than the T1 value measured at the periphery of the infarct zone (1193 ms). Acute infarct size revealed by late gadolinium enhancement (right) was 33.0%. Microvascular obstruction depicted as the central dark zone within the infarct territory was 3.6% of left ventricular mass. The left ventricular ejection fraction and end-diastolic volume were 45.8% and 199.3 mL, respectively. The final infarct size at 6 months was 22.6% of left ventricular mass and the left ventricular end-diastolic volume had increased to 221.8 mL. This patient was re-hospitalized for new onset heart failure during follow-up.
Infarct core native T1 and prognosis post-STEMI therefore presented as means together with SD. Those variables that did not follow a normal distribution are presented as medians with interquartile range. Differences in continuous variables between groups were assessed by the Student's t-test or analysis of variance (ANOVA) for continuous data with normal distribution, otherwise the nonparametric Wilcoxon rank sum test or Kruskal -Wallis test. Differences in categorical variables between groups were assessed using a x 2 test or Fisher's test, as appropriate. Correlation analyses were Pearson or Spearman tests, as indicated. Random effects models were used to compute inter-and intra-rater reliability measures [intra-class correlation coefficient (ICC)] for the reliability of infarct core native T1 values measured independently by 2 observers in 12 randomly selected patients from the cohort.
Univariable and multivariable linear regression methods to identify associates of T1 values for (i) remote myocardium, (ii) injured myocardium within the area at risk, (iii) infarct core in all patients, and (iv) in  Table 1 Clinical and angiographic characteristics of 288 ST-elevation myocardial infarction patients who had cardiac magnetic resonance with evaluable maps for myocardial native T1 magnetization, including the subset of patients with an infarct core revealed by native T1 (all and categorized by tertiles of native T1) Characteristics a All patients (n 5 288) Patients with a native T1 infarct core (n 5 160) (56%) Patients with a native T1 infarct core grouped by tertile of infarct core zone native T1 (ms) at baseline   Missing data: heart rate, n ¼ 1; time from symptom onset to reperfusion, n ¼ 20; ST-segment resolution, n ¼ 1; CRP, n ¼ 7; leucocyte count, n ¼ 1. The patients are grouped according to tertile of T1 in hypo-intense core at baseline. ACE-I or ARB, angiotensin converting enzyme inhibitor or angiotensin receptor blocker; LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; LM, left main coronary artery; RCA, right coronary artery; TIMI, thrombolysis in myocardial infarction grade; PCI, percutaneous coronary intervention. Killip classification of heart failure after acute myocardial infarction: class I-no heart failure, class II-pulmonary rales or crepitations, a third heart sound, and elevated jugular venous pressure, class III-acute pulmonary oedema, and class IV-cardiogenic shock.

Continued
Infarct core native T1 and prognosis post-STEMI patients without late microvascular obstruction are described in Supplementary material online, Methods. Receiver-operating curve (ROC), Kaplan -Meier, and Cox proportional hazards methods were used to identify potential clinical predictors of all-cause death/heart failure events and MACE, including patient characteristics, CMR findings, and native T1. The net reclassification improvement (NRI) was calculated as described by Pencina et al. 38 All P-values are two sided. P-value of,0.05 should be interpreted exploratively. Statistical analyses were performed using R version 2.15.1 or SAS v 9.3, or higher versions of these programs.

Results
Of 343 STEMI patients referred for emergency reperfusion therapy, 300 underwent serial CMR at 1.5 T 2.2 + 1.9 days and 6 months after hospital admission (Figure 2). Two hundred and ninety-two STEMI patients had a T1-map acquisition and 288 (99%) had evaluable T1 data ( Figure 2). Cardiac magnetic resonance follow-up at 6 months was achieved in 267 (93%) of the patients and the reasons for non-attendance are summarized in Figure 2. Information on vital status and SAEs were available in all (100%) of the 288 participants. Table 1 shows the characteristics of the patients, including the patients with a hypo-intense infarct core revealed by native T1 mapping [n ¼ 160 (56%), grouped by thirds of native T1]. The C-reactive protein and leucocyte results are described in Supplementary material online, Table S1. The characteristics of those patients with missing CMR data at 6 months are described in Supplementary material online, Table S2.

Left ventricular function and pathology
Initial cardiac magnetic resonance findings following hospital admission The CMR findings are summarized in Table 2 and case examples are shown in Figure 2. At baseline, the mean (SD) myocardial infarct size was 18 (14) % of LV mass. The average infarct core native T1 (997 (57)) was higher than native T1 in the remote myocardium 961 (25) ms; P , 0.01] but lower than native T1 in the area at risk (1097 (52) ms; P , 0.01). The ICC for T1 core is described in Supplementary material online, Results.
Baseline associates of infarct core native T1 (Hypothesis 1) Native T1 in the infarct core was inversely associated with thrombolysis in myocardial infarction (TIMI) coronary flow grades at the end of emergency PCI, Killip class and neutrophil count at initial presentation (all P , 0.04), independent of left ventricular ejection fraction (LVEF), LV end-diastolic volume, or infarct size ( Table 3).
Relationships for native T1 infarct core vs. infarct pathology, including infarct core T2, myocardial haemorrhage, and microvascular obstruction and 137 (86%) STEMI patients with a hypo-intense native T1 infarct core also had microvascular obstruction. In contrast, only 6.3% of those without hypo-intense infarct core had late microvascular obstruction.   (Figure 2). In all, 42 (4.8%) T1 maps were unsuitable for analysis because of SSFP off-resonance artefacts and 19 (2.2%) T1 maps were affected by motion artefacts. T1 values were higher in infarct tissue surrounding the infarct core than within the infarct core (P , 0.001) and remote myocardium (P , 0.001).
native T1 infarct core for T2 infarct core, myocardial haemorrhage, EGE, and microvascular obstruction are summarized in Supplementary material online, Table S2.
Infarct core tissue characteristics as a marker of subsequent left ventricular remodelling (Hypothesis 2) At 6 months, LV end-diastolic volume increased on average (SD) by 5 (25) ml in 262 patients with evaluable data ( Table 2). Adverse remodelling occurred in 30 (12%) patients and 23 (77%) of these patients had a hypo-intense native T1 core at baseline. Infarct core native T1 (ms) was not associated with change in LV end-diastolic volume at follow-up (P ¼ 0.531). In multivariable regression, native T1 (ms, continuous) within the hypo-intense core was inversely associated with adverse remodelling ( Table 4).
In a sensitivity analysis, the occurrence of a hypo-intense core within the infarct zone on T1 mapping was associated with the odds ratio for being in the top quarter of an increase in LV end-diastolic volume at 6 months (native T1 core to predict Q4 (n ¼ 66) vs. Q1 -Q3 (n ¼ 196) (n ¼ 26 missing); odds ratio 0.994 (0.987, 0.999); P ¼ 0.048).
Native T1 infarct core, microvascular obstruction, and left ventricular outcomes at 6 months The relationships for infarct core native T1 (binary and continuous), T2 core (binary and continuous), microvascular obstruction (binary, % LV mass), and myocardial haemorrhage for LV outcomes, including LV end-diastolic volume and LV ejection fraction, are shown in Table 5. The presence of a hypo-intense infarct core disclosed by native T1 and native T2, the presence and amount of microvascular obstruction, and the occurrence of myocardial haemorrhage, were consistently associated with LV outcomes. Native T1 (ms) was not associated with LV volumes at follow-up, and there was no evidence of non-linearity between infarct core T1 (ms) and LV outcomes.
Infarct core native T1 early post-MI was associated with the concentration of NT-proBNP, a biochemical measure of LV remodelling, at 6 months independent of LV end-diastolic volume at baseline (Supplementary material online, Results).

Infarct core tissue characteristics and health outcomes (Hypothesis 3)
All 288 patients had long-term follow-up data completed. Thirty (10.4%) patients died or experienced a heart failure event. These events included 5 cardiovascular deaths, 3 non-cardiovascular deaths, and 22 episodes of heart failure (Killip Class 3 or 4 heart failure (n ¼ 20) or defibrillator implantation n ¼ 2). Thirteen (4.5%) patients died or experienced a first heart failure hospitalization postdischarge, and 8 (61.5%) of these patients had a hypo-intense infarct core at baseline.
of all-cause death or first hospitalization for heart failure postdischarge (for a 10 ms increase in native T1: hazard ratio 0.730, 95% confidence interval (CI) 0.617, 0.863; P , 0.001) including after adjustment for LVEF at baseline, infarct core T2 (10 ms difference), and myocardial haemorrhage ( Figure 3; Table 6). Infarct core T1 retained its prognostic significance over and above infarct core T2 and myocardial haemorrhage ( Table 6, models C -F). The net reclassification index for the inclusion of infarct core native T1 (10 ms) in a multivariable prognostic model for all-cause death or heart failure post-discharge was 1.129 (95% CI 0.516, 1.742); P , 0.001) ( Table 6). Using ROC analysis, the C-index for infarct core native T1 for all-cause death or heart failure was 0.806. The C-indexes for the prognostic model without and with infarct core native T1 (ms) were 0.715 and 0.931, respectively.
Prognostic importance of infarct core native T1: comparisons with microvascular obstruction and longer term health outcomes In univariate Cox models for infarct core native T1 (ms), native T1 core (binary), T2 core (ms), T2 core (binary), myocardial haemorrhage and the presence (binary) and amount of microvascular obstruction (% LV mass), only infarct core native T1 (ms) (P , 0.001) and the amount of microvascular obstruction (% LV mass) (P , 0.001) were associated with all-cause death or first heart failure hospitalisation after discharge. In a post-hoc analysis stimulated by peer review, the odds ratio for infarct core T1 (10 ms) at baseline for the occurrence of all-cause death, heart failure hospitalisation or adverse LV remodelling was 0.92 (95% CI 0.85, 0.99), P ¼ 0.0312. The inverse relationships between infarct core native T1 (ms) and LV surrogate and adverse health outcomes were reasonably linear, and there was no cut-off value for infarct core T1 (ms) for these outcomes.

Discussion
The main findings of our study are (i) native T1 mapping revealed without an intravenous contrast agent resulted in evaluable scans in 96% of STEMI survivors 2 days post-MI; (ii) acute culprit coronary artery blood flow and circulating measures of systemic inflammation at the time of the hospital admission were multivariable associates of native T1 within the hypo-intense infarct core revealed by T1 mapping 2 days later; (iii) native T1 values (ms) within the infarct core were clinically meaningful since they were independently associated with adverse remodelling, NT-proBNP concentrations at 6 months,   Infarct core native T1 and prognosis post-STEMI and all-cause death or heart failure hospitalization post-discharge during longer term follow-up; (iv) compared with infarct core T2 or myocardial haemorrhage revealed by T2* mapping, infarct core T1 was more consistently associated with LV surrogate outcomes and all-cause death or heart failure hospitalization ( Table 6), implying T1 core is more closely linked with infarct pathology; (v) compared with microvascular obstruction revealed by contrast-enhanced CMR, a hypo-intense infarct core revealed by T1 mapping had similar prognostic significance for LV outcomes at 6 months and for post-discharge cardiac events including all-cause mortality and heart failure hospitalization in the longer term (Tables 5 and 6). Finally, our paper adds to the emerging literature on the prognostic value of quantitative native T1 CMR 39 and reaffirms the prognostic importance of MVO post-STEMI. 3 The results of this study extend what is known about infarct core pathology, and also provide a potential mechanistic explanation. Infarct size 1,2 and pathology, including microvascular obstruction, 3 haemorrhage, 5 and salvage, 4 predict cardiac morbidity and mortality post-MI. These pathologies are revealed by contrast-enhanced CMR, and until recently, the assessment of infarct tissue without an intravenous contrast agent has been limited to T2-weighted and T2* imaging of myocardial haemorrhage. 5,11,29,40,41 11 or in proof-of-concept clinical studies involving much smaller numbers of MI patients. 9,12 -15 Our study extends these findings in a much larger STEMI cohort and provides new evidence that native T1 core is more reflective of the severity of infarct injury and its prognostic importance than infarct core T2 and potentially also myocardial haemorrhage.
We also compared infarct core pathology delineated by native T1 mapping with microvascular obstruction, which is an established prognostic CMR biomarker post-MI. 3 Native T1 mapping is obtained without the use of an intravenous gadolinium-based contrast agent whereas microvascular obstruction is revealed by CMR  Thirteen (8.1%) patients experienced all-cause death or heart failure hospitalization post-discharge. Given the limited number of adverse events, the models were specified to assess the prognostic relationships of infarct core native T1 vs. circulating markers of systemic inflammation, LV function, LV volume, and infarct characteristics that were measured at approximately the same time 2 days after hospital admission. LVEF, left ventricular ejection fraction. Infarct core T1 (10 ms difference) is highlighted in bold.
imaging of EGE and LGE after intravenous contrast administration. We observed a high degree of concordance between the occurrence of a hypo-intense infarct core depicted by native T1 CMR (56%) and microvascular obstruction (50%) as revealed by contrast-enhanced CMR. Although both a native T1 core and microvascular obstruction are depicted as a hypo-intense core within the hyperintense infarct zone (Figure 2), the physics of the CMR techniques is entirely different. A hypo-intense infarct core depicted by non-contrast native T1 mapping is due to local destruction of the T1 magnetization signal. On the other hand, microvascular obstruction ( Figure 2) is due to a failure of gadolinium contrast to penetrate within the infarct core. Both CMR methods are T1-weighted but contrast kinetics are not relevant for native T1 mapping since intravenous contrast is not administered. Accordingly, T1 mapping avoids the theoretical risks and actual restrictions associated with contrast-enhanced CMR. Furthermore, acquisition of the native T1 map does not prolong the CMR scan, in contrast to late gadolinium enhancement imaging for microvascular obstruction which is typically imaged 10-15 min after dosing. 19 Culprit artery coronary flow at the end of emergency PCI reflects the efficacy of coronary reperfusion, and reduced coronary flow initially independently predicted native T1 relaxation time within infarct core as assessed by CMR 2 days later. Similar associations also exist for microvascular obstruction, 44,45 and in our study, both infarct core native T1 and microvascular obstruction were independently associated with circulating biomarkers of acute systemic inflammation. The occurrence of an infarct core disclosed by native T1 mapping, and the nature of the core (i.e. the native T1 value), was associated with the initial severity of MI (i.e. Killip heart failure class), systemic inflammation (i.e. leucocyte counts), and LV remodelling and health outcomes in the longer term. We think that the prognostic significance of native T1 values within the hypo-intense core are a distinctive attribute compared with microvascular obstruction since signal-intensity values within microvascular obstruction are not clinically meaningful beyond binary categorization (i.e. present/absent).

Limitations
We performed a single centre natural-history study involving nearconsecutive STEMI admissions. The STEMI patients in our naturalhistory study were recruited 24/7 therefore flow cytometry and routine NT-proBNP testing in all participants was not pragmatically possible.
T1 assessment is sensitive to motion artefacts and imperfect breath holding, which and may reduce image quality. A shortened version of this sequence (ShMOLLI) involving only nine heart beats has been developed. This method shortens breath hold time and may help to account for these limitations. 42 Despite this, the MOLLI method has high precision reproducibility. Our T1 measurements are in good agreement with in vivo data published in the literature, including previous measurements using the ShMOLLI sequence. 42,43 The limited number of adverse events constrained the number of variables that could be included in the multivariable models (e.g. Tables 4 and 6); however, the associations between infarct core native T1 and a range of surrogate and clinical outcomes including adverse remodelling revealed by CMR, NT-proBNP, and the primary health outcome (all-cause death/heart failure), supports the adverse prognostic importance of infarct core native T1. Our study does not permit inference on causality, and other interpretations of our data are possible and further studies are warranted.

Conclusions
We found that infarct core pathology revealed by native T1 maps had similar prognostic value compared with microvascular obstruction revealed by late gadolinium enhancement CMR. Native T1 mapping is potentially widely applicable in clinical practice, not limited by renal disease, and so potentially could represent an alternative non-contrast CMR option for the assessment of infarct pathology.

Supplementary material
Supplementary material is available at European Heart Journal online.