Adenosine Stress and Rest T1 Mapping Can Differentiate Between Ischemic, Infarcted, Remote, and Normal Myocardium Without the Need for Gadolinium Contrast Agents

Objectives The aim of this study was to evaluate the potential of T1 mapping at rest and during adenosine stress as a novel method for ischemia detection without the use of gadolinium contrast. Background In chronic coronary artery disease (CAD), accurate detection of ischemia is important because targeted revascularization improves clinical outcomes. Myocardial blood volume (MBV) may be a more comprehensive marker of ischemia than myocardial blood flow. T1 mapping using cardiac magnetic resonance (CMR) is highly sensitive to changes in myocardial water content, including MBV. We propose that T1 mapping at rest and during adenosine vasodilatory stress can detect MBV changes in normal and diseased myocardium in CAD. Methods Twenty normal controls (10 at 1.5-T; 10 at 3.0-T) and 10 CAD patients (1.5-T) underwent conventional CMR to assess for left ventricular function (cine), infarction (late gadolinium enhancement [LGE]) and ischemia (myocardial perfusion reserve index [MPRI] on first-pass perfusion imaging during adenosine stress). These were compared to novel pre-contrast stress/rest T1 mapping using the Shortened Modified Look-Locker Inversion recovery technique, which is heart rate independent. T1 values were derived for normal myocardium in controls and for infarcted, ischemic, and remote myocardium in CAD patients. Results Normal myocardium in controls (normal wall motion, MPRI, no LGE) showed normal resting T1 (954 ± 19 ms at 1.5-T; 1,189 ± 34 ms at 3.0-T) and significant positive T1 reactivity during adenosine stress compared to baseline (6.2 ± 0.5% at 1.5-T; 6.3 ± 1.1% at 3.0-T; all p < 0.0001). Infarcted myocardium showed the highest resting T1 of all tissue classes (1,442 ± 84 ms), without significant T1 reactivity (0.2 ± 1.5%). Ischemic myocardium showed elevated resting T1 compared to normal (987 ± 17 ms; p < 0.001) without significant T1 reactivity (0.2 ± 0.8%). Remote myocardium, although having comparable resting T1 to normal (955 ± 17 ms; p = 0.92), showed blunted T1 reactivity (3.9 ± 0.6%; p < 0.001). Conclusions T1 mapping at rest and during adenosine stress can differentiate between normal, infarcted, ischemic, and remote myocardium with distinctive T1 profiles. Stress/rest T1 mapping holds promise for ischemia detection without the need for gadolinium contrast.

However, assessment of MBF alone may not reflect all aspects of ischemia (4)(5)(6)(7). Myocardial blood volume (MBV), on the other hand, may be a more comprehensive global marker of ischemia, as it represents the total volume of capacitance vessels in both the microcirculations and macrocirculations (4)(5)(6)8,9). Significant coronary artery stenosis induces capillary recruitment with an increase in resting MBV (9). Myocardial blood volume measurements derived from first-pass contrast-based CMR closely reflect the level of microvascular autoregulation (4,5,9). As a surrogate for epicardial CAD, recent animal studies showed that disturbances in MBV can effectively detect anatomically significant coronary stenoses (4,10), and distinguish their functional relevance (11). Moderate and severe coronary stenoses may be better differentiated using the index of myocardial blood volume reserve than by myocardial perfusion imaging (4). Furthermore, MBV may relate better to cardiomyocyte metabolism by reflecting changes in myocardial oxygen consumption, which is a more reliable marker of cellular ischemia (4,6,11,12). Therefore, MBV determination during vasodilatation and at rest may constitute a more complete assessment of ischemia than MBF (via perfusion imaging) alone.
Native (pre-contrast) T1 mapping is a novel CMR technique that can potentially improve ischemia detection by detecting MBV and myocardial water content. In MRI, hydrogen-proton spin-lattice relaxation time (T1) is a magnetic property of tissue that is prolonged by increased water content (13,14) and, importantly, depends on blood T1 through its partial volume (14). Each tissue type, such as myocardium, has a specific range of normal T1 values, deviation from which is indicative of disease (13,(15)(16)(17)(18)(19)(20). By measuring and displaying T1 relaxation times pixel by pixel, native T1 mapping provides a quantitative biomarker of intracellular and extracellular environments of the myocardium without the need for intravenous contrast agents (13). T1 mapping is highly reproducible with tight normal ranges (13,14), capable of diagnosing a variety of cardiac diseases (13,15,16,(18)(19)(20)(21)(22). Increased myocardial T1 values act as a surrogate for increased myocardial water (13); hence coronary vasodilatation, which increases MBV (4)(5)(6), is expected to prolong T1 and allow detection of microvascular and myocardial blood volume changes during ischemia (9). We have recently demonstrated the ability of stress/rest T1 mapping to detect increases in MBV from coronary vasodilatation in patients with severe aortic stenosis and nonobstructive coronary arteries, with complete reversal and normalization after aortic valve replacement (23). In summary, stress/rest T1 mapping is a highly promising technique for the detection of ischemia and is particularly attractive for applications in patients with CAD.
In this proof-of-principle study, we demonstrate the ability of T1 mapping, during adenosine vasodilatory stress and rest, to distinguish 4 myocardial tissue classes: normal, infarcted, ischemic, and remote myocardium, as a novel gadolinium-free method for ischemia detection. We performed CMR scans in normal controls and patients with known CAD assessing left ventricular (LV) function (cine), viability (late gadolinium enhancement [LGE]), and ischemia (adenosine stress gadolinium first-pass perfusion), and compared them with novel T1 mapping to establish characteristic stress and rest T1 profiles of these 4 myocardial tissues.

METHODS
Ethical approval was granted for all study procedures and all subjects gave written informed consent. NORMAL CONTROLS. Twenty normal controls without history of cardiovascular disease, not on cardiovascular medications, and with normal electrocardiograms were recruited. Ten volunteers (7 males, 33 AE 10 years of age) underwent CMR scans at 1.5-T (Avanto, Siemens Healthcare, Erlangen, Germany) and 10 volunteers (7 males, 36 AE 11 years of age) were scanned at 3.0-T (TimTrio, Siemens Medical Solutions), all with identical CMR protocols. All subjects avoided potential adenosine antagonizers (e.g., caffeine) for $24 h before CMR scans.
Cine images were obtained as previously described (23). T1 mapping was performed using the Shortened Modified Look-Locker Inversion recovery (ShMOLLI) sequence, which has been shown to be heart rate independent over a wide range of T1 values (13), as previously described (13,14,18,19,21). In brief, T1 maps were acquired at rest and during peak adenosine stress (140 mg/kg/min, intravenously for $3 to 6 min) in 3 short-axis (basal, midventricular, apical) slices  (23). The basal slice was carefully planned to exclude LV outflow tract. First-pass perfusion imaging was performed, on matching short-axis slices to T1 maps, during peak adenosine stress with an intravenous bolus of gadolinium (0.03 mmol/kg; Dotarem, Guerbet, Villepinte, France), followed by a 15 ml saline flush (12). Rest perfusion images were acquired >15 min after adenosine discontinuation (12,23). LGE imaging was performed w8 to 10 min after an additional bolus of gadolinium (0.1 mmol/kg) (12). PATIENTS WITH CAD. Ten patients with angiographically significant stenosis (>50%) in $1 coronary artery, who underwent CMR at 1.5-T using cine, adenosine stress/rest T1 mapping, adenosine stress/ rest perfusion, and LGE imaging, were included to illustrate the ability of stress/rest T1 mapping to distinguish myocardial tissue classes. IMAGE ANALYSIS. LV function and first-pass myocardial perfusion were analyzed as previously described (23). Short-axis T1 maps were manually contoured to outline the endocardium and epicardium using dedicated software and underwent strict and extensive quality control process as previously described (13)(14)(15)18,19,21,23), which resulted in exclusion of 11.7% of segments (for additional details, see the Online Appendix). For normal controls, mean myocardial T1 values were derived from T1 maps at rest and during adenosine stress per subject, per slice, and per segment (American Heart Association 16-segment model) (21). T1 reactivity to adenosine stress was expressed in absolute terms: DT1(ms) ¼ T1 stress -T1 rest and as percentages: dT1(%) ¼ DT1OT1 rest Â 100. T1 rest and T1 stress represent mean T1 values at rest and during adenosine stress, respectively.
In CAD patients, the mean T1 of ischemic myocardium was measured by placing a region of interest (ROI) in an area corresponding to the area of reversible perfusion defect on first-pass stress and rest imaging but without LGE, accompanied by angiographic evidence of significant coronary stenosis as assessed by an expert interventional cardiologist. Infarcted myocardium on T1 maps was defined by placing a ROI corresponding to an area of infarction on LGE images, defined as enhancement involving the subendocardium of >50% transmurality, as assessed by 2 experienced observers. To avoid partial volume contamination from the LV blood pool, all infarct ROIs were placed in the core of the infarcts and away from the endocardial and epicardial borders carefully referenced against corresponding cine images in the same phase of the cardiac cycle. Remote myocardial ROI on T1 maps were placed contralateral to the ischemic myocardium in areas without evidence of first-pass perfusion defects, regional wall motion abnormalities, LGE, or significant upstream angiographic coronary stenosis. A reference ROI was also placed in the LV blood pool.  Table 1).  (14) ( Figure 2B). There was no significant correlation between age and T1 reactivity in the normal controls at 1.    is expected to increase the myocardial water content and prolong T1 relaxation time. T1 mapping offers highly reproducible pixel-wise estimations of myocardial T1, both in normal and pathologic states (15)(16)(17)(18)(19)(20)(21)24    Each patient showed an area of inducible ischemia on stress/rest perfusion images (A, red arrows) and an area of infarction (B, white arrows). On T1 maps (C and D), the corresponding remote, ischemic, and infarcted regions as well as the LV blood pool are as labeled. In all 3 patients, the remote myocardial T1 at rest was within normal ranges (14), which increased significantly with adenosine stress (marked by [). No significant T1 reactivity was observed in the ischemic or infarcted regions, or the left ventricular blood pool. Reference color T1 maps are shown in the Online Figure 1 for comparative/illustrative purposes. CMR ¼ cardiac magnetic resonance.