Echocardiographic Ischemic Memory Molecular Imaging for Point-of-Care Detection of Myocardial Ischemia

Abstract

Background:

Non-invasive molecular imaging of recent ischemia can potentially be used to diagnose acute coronary syndrome (ACS) with high accuracy.

Objectives:

We hypothesized that bedside myocardial contrast echocardiography (MCE) ischemic memory imaging could be achieved with phosphatidylserine microbubbles (MB_PS) that are retained in the microcirculation via ischemia-associated endothelial activation.

Methods:

A dose-finding study was performed in healthy volunteers (n=17) to establish optimal MB_PS dosing. Stable patients with ACS (n=30) and confirmed antecedent but resolved myocardial ischemia were studied within two hours of coronary angiography and PCI when indicated. MCE molecular imaging was performed 8 minutes after intravenous administration of MB_PS. MCE perfusion imaging was used to assess status of the post-ischemic microcirculation.

Results.

Based on dose-finding studies, 0.10 or 0.15 mL of MB_PS based on body mass was selected. In patients with ACS, all but two underwent primary PCI. MCE molecular imaging signal intensity was greater in the post-ischemic risk area vs. remote territory (median [95% CI]: 56 [33–66] vs 8 [2–17] IU, p<0.001) with a receiver operating characteristic curve c-statistic of 0.94 to differentiate post-ischemic from remote territory. Molecular imaging signal in the risk area was not related to type of ACS (unstable angina=3, NSTEMI=14, STEMI=13), peak troponin, time to PCI, post-PCI myocardial perfusion, GRACE score, or HEART score.

Conclusions:

Molecular imaging with point-of-care echocardiography and MB_PS can detect recent but resolved myocardial ischemia. This bedside technique requires only minutes to perform and appears independent of the degree of ischemia.

CONDENSED ABSTRACT:

We hypothesized that bedside myocardial contrast echocardiography (MCE) ischemic memory imaging is possible with phosphatidylserine microbubbles (MB_PS) that are retained in post-ischemic microcirculation. A dose-finding study was first performed in healthy volunteers. In patients with ACS (n=30) in whom ischemia was confirmed, MCE molecular imaging signal within two hours of angiography and PCI was on average 5-fold higher in the post-ischemic versus remote territory. Risk area MB_PS enhancement was not related to peak troponin, type of ACS, or time to PCI. These results indicate that MCE molecular imaging can be used to detect recent but resolved myocardial ischemia.

Gaps remain in the clinical algorithms used to diagnose or exclude acute coronary syndromes (ACS). These challenges remain despite the introduction of high-sensitivity troponin assays that vary according to age and can be elevated in those with non-ACS conditions (1,2). Molecular imaging techniques capable of detecting recent myocardial ischemia, even hours after its resolution, have been proposed for diagnosing ACS (3). This approach could be impactful in patients with ACS but little myocardial necrosis or in those in whom traditional clinical markers of ischemia are unreliable or non-specific. Additionally, the ability to spatially identify the ischemic region could be useful for risk stratification or procedural planning. Radionuclide molecular imaging of long-chain fatty acid tracers to detect changes in myocardial metabolism that persist after ischemia showed promise in clinical trials (4,5), yet this approach has practical limitations as a point-of-care technology.

Echocardiographic ischemic memory imaging has the ability to be performed at the bedside and to provide results within minutes, assets that have justified the increasing use of conventional echocardiography in the acute care setting. Myocardial contrast echocardiography (MCE) ischemic memory imaging in rodent, canine, and non-human primate models has been achieved using microbubble (MB) contrast agents that are either targeted to endothelial selectins through surface ligand conjugation (6–9), or that contain anionic phospholipids such as phosphatidylserine (MB_PS) that promote agent retention through opsonization or adhesion to the post-ischemic activated endothelium (10,11). Of the two approaches, MB_PS has the advantage of simplicity of its formulation and current availability.

In this proof-of-concept study, we hypothesized that MB_PS would be retained selectively within the post-ischemic myocardial microcirculation in patients with ACS, where they would be detected within minutes by MCE. To test this hypothesis, we first performed a dose-finding study in healthy volunteers which was used to select an appropriate MB_PS dose for MCE ischemic memory molecular imaging. MCE molecular imagine was then performed in patients with ACS and evidence of ischemia within the prior 24 hours. Imaging was performed early after angiography and, when indicated, percutaneous coronary intervention (PCI).

METHODS

Subjects and Study Design

The study was approved by the Human Investigational Research Board at Oregon Health & Science University (NCT03009266), and by the United States Food and Drug Administration under a physician-sponsored Investigation New Drug application (IND #139347). First, a dose-ranging study protocol was performed to identify a dose of MB_PS sufficient to produce strong myocardial opacification on first-pass, but that at 8 minutes produces minimal opacification from either high concentration of freely circulating agent, or from non-specific charge-mediated retention in the coronary microcirculation (12). Eighteen adult healthy volunteers were recruited. Subjects were excluded for history of cardiovascular disease, diabetes mellitus, systemic inflammatory disorder, left ventricular dysfunction, or right-to-left shunting on screening echocardiography. In the second study, thirty adult patients hospitalized for ACS (unstable angina, NSTEMI or STEMI) and referred for urgent angiography and PCI if indicated were recruited. Patients were studied within two hours of the angiographic procedure. Subjects had to have evidence of ischemia within the preceding 24 hours defined as ST-segment changes on electrocardiogram (ECG), positive troponin-I (non-high-sensitivity assay), or new wall motion abnormality. Subjects not undergoing PCI were eligible only if a culprit artery could be identified, TIMI-grade 3 flow was present on angiography, and there was no clinical evidence of ongoing ischemia. Exclusion criteria included clinical evidence of ischemia at rest at the time of the MCE study, inability to identify a culprit artery, shock, presence of a ventricular assist device, or severe heart failure. For both studies subjects were excluded for pregnancy or allergy to ultrasound contrast agents, eggs, or egg products.

Molecular Imaging Protocols

The MB_PS contrast agent used (Sonazoid, GE Healthcare, Amersham, UK) is approved for use in certain countries for abdominal ultrasound applications. Sonazoid microbubbles are composed of decafluorobutane gas (8 μL/mL), a shell composed almost exclusively of hydrogenated egg phosphatidylserine shell (approximately 0.4 mg/mL), and sucrose excipient; with an average microbubble diameter of 2.1±0.1 μm and a reconstituted concentration of 1.2×10⁹ mL⁻¹ (13,14). For the dose-finding studies, subjects received a single intravenous bolus injection at doses of 0.1, 0.15, 0.3, 0.6, 0.9 or 1.2 mL (n=3 to 4 subjects for each dose). This dose range is equivalent to 0.8 to 9.6 μL decafluorobutane gas volume, or 1.2×10⁸ to 1.44×10⁹ microspheres (13). These studies were used to select dose in the ACS cohort. Contrast-specific amplitude-modulation MCE imaging (iE33, Philips Ultrasound, Andover, MA) was performed in the apical 4-chamber, 2-chamber, and 3-chamber views starting 8 minutes after IV injection of MB_PS. Imaging was performed at a mechanical index 0.25–0.29, which has been shown to be ideal for producing high signal-to-noise without Sonazoid destruction (15). Several end-systolic frames were acquired during ECG-triggered imaging, followed by a 5-frame high mechanical index (>1.0) “flash” sequence to null agent through inertial cavitation (15). Several end-systolic frames >5 seconds after the flash sequence were acquired to measure signal from freely circulating agent. Signal from retained agent alone was calculated off-line (iMCE, Narnar LLC., Portland, OR) by digital subtraction of several averaged post-cavitation frames from several pre-cavitation frames (Figure 1A) (6). All studies were completed within three minutes.

Figure 1. Imaging Algorithm and Dose-finding Studies.

(A) Kinetic model for molecular imaging whereby total myocardial concentration of MB_PS (C_M) represents the sum of retained (R_M) and freely circulating (f) agent (see text). Graphs depict mean (±SEM) myocardial signal from (B) freely circulating; or (C) retained MB_PS. (D) End-systolic apical 4-chamber (top) and 2-chamber (bottom) MCE images (0.6 mL dose) illustrating myocardial enhancement pre-flash and late post-flash. Apical enhancement is absent in the 2-chamber view (arrow) from previous 4-chamber flash. (E) Parasternal short-axis MCE illustrating myocardial signal enhancement from retained agent (1.2 mL dose), and signal voids from previous cavitation of agent in the three apical planes as labeled.

For dose-finding studies, regions-of-interest were placed over each of the three major coronary artery territories for analysis. For ACS studies, the risk area was adjudicated by consensus of three cardiologists who reviewed angiographic images and, when applicable, the presence of new wall motion abnormalities. Analysis was performed for regions-of-interest placed over the risk area, using at least two views when possible, and over a remote territory for each view.

Myocardial Perfusion Imaging

After molecular imaging, MCE perfusion imaging for all three apical views was performed with amplitude modulation at a mechanical index of 0.2. Sonazoid was diluted 1:10 v:v in normal saline and infused at a constant rate of 1.5 ml/min. End-systolic frames were acquired after a high-mechanical index sequence to assess contrast replenishment kinetics. Regions-of-interest were drawn over the same territories used for molecular imaging. The first post-destructive frame was digitally subtracted from all subsequent frames and background-subtracted time-intensity data were fit to the function:

$$\text{y}=\text{A}(1-{\text{e}}^{-\beta \text{t}})$$

where y is signal intensity at time t, A is the plateau intensity which is an index of relative microvascular blood volume (MBV_i), and β is the rate constant reflecting microvascular blood flux rate. Microvascular blood flow was quantified by the product of MBV_i and β (16).

Echocardiography

Echocardiography using 2-D, Doppler, M-mode, and speckle-tracking strain analysis were performed to assess left ventricular dimensions and function according to guidelines published by the American Society of Echocardiography (17). Images for measuring ventricular dimensions, ejection fraction, and subjective wall motion score index (WMSI) were acquired using contrast-enhanced left ventricular opacification during the Sonazoid infusion.

Statistical Analysis

Data were analyzed using Prism (version 9.0, GraphPad Software, San Diego, CA). For data that were normally distributed by the D’Agostino and Pearson omnibus test, differences were assessed by paired or unpaired Student’s t-test. For non-normally distributed data, differences were assessed by either Mann-Whitney U test or Wilcoxon test. For multiple comparisons, a one-way analysis-of-variance (ANOVA) was performed for normally distributed data, whereas a Kruskal-Wallis test followed by Dunn’s multiple comparison test were performed for non-normally distributed data. Correlations were made using linear regression and a Spearman rho test. Receiver operating characteristic curves were generated in a non-paired fashion to determine an absolute cutoff for differentiating ischemic and non-ischemic myocardium. Differences were considered significant at P<0.05.

RESULTS

Dose-Finding for MB_PS

In the dose-finding study in healthy volunteers, the mean age was 32±8 years, four subjects (24%) were female, and body mass was 75±14 kg. The MCE signal intensity attributable to freely-circulating MB_PS remaining in the microvascular blood pool 8 minutes after injection increased incrementally with dose (p<0.001 for linear trend) (Figure 1B). When evaluating signal attributable to non-specific retention of MB_PS (Figure 1C and 1D, Supplemental Videos), signal increased with dose up to 0.6 mL. Doses above 0.6 mL resulted in minor reductions in signal from retained agent, representing an anticipated effect from very high circulating concentration of MB_PS, the signal from which overwhelms signal from retained agent. Non-specific retention of MB_PS at doses of 0.3 mL and higher was further confirmed by MCE imaging in the parasternal short-axis view after performing destruction-replenishment sequences in three apical views (Figure 1E, Supplemental Video). Short-axis imaging revealed signal voids or “null lines” in the myocardial regions corresponding to each apical imaging plane which had been previously exposed to high-power flash sequence to destroy microbubbles. This phenomenon was also evident by an apical defect seen after performing destruction-replenishment sequence in one apical view and rotating to another apical view (arrow in Figure 1D). Based on the data from these dose-finding protocols, subjects with ACS were studied using a dose of 0.1 or 0.15 mL (corresponding 0.8 to 1.2 mL perfluorobutane, or 1.2 to 1.6 ×10⁸ microspheres) depending on body mass index less than or greater than 30 kg/m².

ACS Patient Characteristics

The clinical characteristics for subjects enrolled in the ACS detection study are provided in Table 1. The majority of ACS subjects were classified as NSTEMI or STEMI with only a small number with unstable angina. Of those with NSTEMI or STEMI, all but one had a type I MI. For the entire group, all but two subjects underwent PCI and all but one had post-PCI TIMI-3 grade flow. Segmental wall motion was normal (WMSI= 1.0) in only two subjects. LVEF was only modestly reduced, with only three subjects having an LVEF less than 35%. The median time from symptom onset to molecular imaging study was 448 minutes (95% CI: 270 to 1470).

Table 1.

Clinical Characteristics, Angiographic Data, and Echocardiographic Data for ACS Study Group (n=30)

Age (years)	68±12
Male/Female (n)	23/7
BMI (kg/m2)	28.6±6.1
Dyslipidemia, n (%)	24 (80%)
Diabetes mellitus, n (%)	9 (30%)
Hypertension, n (%)	23 (77%)
Smoking history, n (%)	15 (50%)
Prior revascularization, n (%)	5 (17%)
Troponin-I, ng/mL, median [95% CI]	15.3 [5.9–42.2]
Unstable angina, n (%)	3 (10%)
NSTEMI, n (%)	13 (43%)
STEMI, n (%)	14 (47%)
PCI, n (%)	27 (90%)
Culprit Vessel:
LAD, n (%)	14 (47%)
RCA, n (%)	9 (30%)
LCx or ramus, n (%)	7 (23%)
LVEF, %	51±11
WMSI	1.51±0.40
Cardiac output, L/min	4.2±1.6
Global longitudinal strain, %	−14.8±5.0
GRACE Score	119±28
HEART Score	8.0±1.3

ACS, acute coronary syndrome; BMI, body mass index; LAD, left anterior descending; LCx, left circumflex; LVEF, left ventricular ejection fraction; PCI, percutaneous coronary intervention; RCA, right coronary artery; WMSI, wall motion score index.

Regional Myocardial Microvascular Perfusion

MCE performed at the time of molecular imaging (Figure 2) showed wide variation in MBF in the post-ischemic risk area and the remote territory. This variation likely reflected a wide range in systemic hemodynamics, and variability in either the degree of microvascular reflow or post-ischemic hyperemia in the risk area. Overall, MBF in the risk area was not significantly different than in the remote territory. Myocardial MBVi also varied considerably with no significant difference between the risk area and remote territory, suggesting adequate microvascular reflow in the risk area in most subjects.

Figure 2. MCE Perfusion Imaging in ACS Subjects.

(A, B) Example of MCE perfusion images (3-chamber view) and corresponding time-intensity data illustrating complete reflow in the post-ischemic risk area in a patient 1–2 hours after PCI of the LAD. Background-subtracted color-coded (scale at right) end-systolic images are shown immediately after the cavitating flash sequence (BL) and at incremental end-systolic time intervals. Graphs depict MCE-derived (C) myocardial blood flow (MBF), and (D) microvascular blood volume index (MBVi) in the risk area and remote myocardium. Bar-and-whisker plots show median (bar), interquartile range (box), and range (whiskers).

Molecular Imaging of Recent Myocardial Ischemia

There were no adverse agent-related events associated with MB_PS injection, nor were there any changes in vital signs after injection (Table 2). Image quality was sufficient for analysis in all but two subjects. Signal from MB_PS, which was the primary endpoint, was 5-fold higher in the post-ischemic risk area than in the remote region (Central Illustration). The signal enhancement in the risk area was greater than the remote territory in all but two of the thirty subjects, and on average signal was greater in the risk area by 41.5±33.2 IU (p<0.001). Because MBF influences the mass of contrast material entering into a microvascular bed, MB_PS signal was normalized to risk area MBF. Differences between the two regions remained after normalizing to perfusion. Receiver operating characteristic curve of non-paired data from this test cohort yielded a c-statistic of 0.94; with a signal ≥21 IU yielding a sensitivity of 93% (95% CI: 77%–99%) and a specificity of 78% (95% CI: 59%–89%) for differentiating risk area from remote territory (Central Illustration). After performing log-linear transformation of molecular imaging data, MB_PS signal enhancement in the ischemic territory remained significantly elevated compared to the remote territory, and there was similar discriminatory performance on receiver operating characteristic curves (Supplemental Figure 1).

TABLE 2.

Vital Sign Data in ACS Patients Undergoing Injection of MB_PS (n=30)

	Baseline	5 min	10 min
Heart Rate (min−1)	75 ± 15	73 ± 16	72 ± 13
Systolic BP (mm Hg)	128 ± 18	127 ± 19	127 ± 19
Diastolic BP (mm Hg)	74 ± 12	77 ± 16	77 ± 13
Oxygen Saturation (%)	97 ± 2	97 ± 2	97 ± 3

ACS, acute coronary syndrome; BP, blood pressure.

Central Illustration. MCE Ischemic Memory Molecular Imaging in ACS Subjects.

(A) Apical 3-chamber MCE perfusion imaging late after flash-replenishment which represents microvascular reflow in the adjudicated LAD risk area (RA). (B) Grayscale MCE molecular imaging showing enhancement in the risk area, confirmed to be from MB_PS retention by lack of reflow after the flash sequence. (C) Color-coded (scale at right) MCE molecular imaging derived by digital subtraction of freely circulating (post-flash) frames from pre-flash frames. Box-whisker plots illustrate (D) background-subtracted MB_PS signal in the risk area and remote territory; and (E) signal normalized to MBF. (F) Receiver operating characteristic (ROC) curve for background-subtracted MB_PS signal intensity to differentiate post-ischemic remote territory.

There were no major differences in risk area MB_PS signal intensity based on either the category of ACS or the culprit artery territory (Figure 3). There was no significant relationship between signal from MB_PS in the risk area and either MBF, peak troponin-I concentration, time from symptom onset to PCI, global longitudinal strain, GRACE score, or HEART score (Figure 4). These findings indicate that the use of MB_PS to detect ischemia does not require severe or prolonged ischemia, and that signal is relatively independent of perfusion. In those with normal microvascular reflow, signal enhancement from MB_PS was seen in diffusely within the risk area (Central Illustration). In those that had regions with partial microvascular “no-reflow”, signal enhancement was seen primarily where there was microvascular reflow and not in regions lacking perfusion (Supplemental Figure 2).

Figure 3. MCE Ischemic Memory Molecular Imaging Signal by ACS Type and Vessel.

Box-whisker plots depicting MB_PS signal in the risk area according to (A) type of ACS, and (B) culprit coronary artery territory. Examples show similar degrees of MB_PS signal enhancement on MCE molecular imaging for two patients after PCI of the RCA for either (C) inferior NSTEMI, or (D) inferior STEMI. Images include grayscale MCE in the apical 2-chamber view obtained pre- and post-high-power flash sequence, and corresponding color-coded (scale at right) background-subtracted image. NSTEMI, non-ST-elevation MI; STEMI, ST-elevation MI; USAP, unstable angina pectoris.

Figure 4. Clinical Determinants of Ischemic Memory Molecular Imaging Signal Intensity.

Correlations are shown between the MCE molecular imaging signal for MB_PS in the risk area and: (A) risk area myocardial blood flow (MBF); (B) Peak troponin-I; (C) time from symptom onset to percutaneous coronary intervention (PCI); (D) global longitudinal strain; and (E) GRACE score. Shaded regions depict 95% confidence interval of the regression line. (F) Box-whisker plot illustrates the MCE molecular imaging signal in the risk area and HEART score.

DISCUSSION

To our knowledge, this study is the first to demonstrate echocardiographic molecular imaging in humans. Previous studies in pre-clinical models of myocardial ischemia-reperfusion injury have established that PS in the shell of MBs facilitates their adhesion to activated microvascular endothelium and to certain leukocyte subsets in post-ischemic tissue (11,18), thereby creating an opportunity for imaging recently-resolved ischemia (10,11,19). In the current study, MCE molecular imaging in patients with ACS was performed with Sonazoid, a MB agent composed almost entirely of PS as its shell constituent (14) that is approved for human abdominal contrast ultrasound in several countries. Patients undergoing angiography were recruited, thereby representing a population that afforded high confidence in both the presence and location of recent myocardial ischemia. MCE molecular imaging was performed using a protocol requiring approximately ten minutes between agent injection and completion of imaging. We observed selective signal enhancement in the risk area which was independent of the amount of flow in the reperfused regions the degree of recent ischemia. The latter finding is important when considering the need to eventually apply the technology in a lower risk population.

Our study was predicated on the idea that non-invasive ischemic memory imaging can improve the quality of care and reduce healthcare costs for patients being evaluated acutely for myocardial ischemia. By imaging molecular events that persist for hours after ischemia has resolved, it may be possible to not only expedite care, but to improve the diagnostic accuracy for ACS. The impact will likely be greatest in situations where traditional diagnostic approaches are less reliable, such as the evaluation of patients with pre-existing ECG abnormalities or paced rhythm. The introduction of high-sensitivity troponin assay algorithms has markedly improved sensitivity for ACS. Yet, there are limitations of high-sensitivity troponin assays, particularly with positive predictive value, in certain populations (1,2). They also do not reveal the location and spatial extent of ischemic myocardium which could be impactful even in those with known ACS for either risk stratification or to plan appropriate target vessel revascularization in those with complex coronary anatomy.

The mechanism by which PS mediates retention of lipid MB is only partially understood. Phosphatidylserine imparts a strong negative charge to the MB shell, evidenced by Sonazoid’s known zeta potential of −76 to −82 mV (11,13). Anionic charge together with the lack of surface polymers such as polyethylene glycol increases opsonin-mediated attachment to inflammatory and endothelial cells (12,20). Yet PS also has a distinct biologic role. In most mammalian cell membranes, PS is present but is sequestered to the inner leaflet through an energy dependent process. Outer leaflet PS is a marker of cell apoptosis and triggers immune clearance of cells by a variety of processes that are complement dependent and independent, and that are likely involved in Sonazoid retention (20–22). Experience with liposomal drugs has provided a detailed understanding of how non-self membranes interact with serum complement and other blood proteins, and how both anionic charge and PS can accelerate opsonization (20,23,24). This knowledge has been applied to lipid-stabilized MBs where incorporation of PS has been leveraged to accelerate opsonization involving both the C1q component and C3 deposition on the shell surface (11). Opsonization results in interaction with innate immune cells recruited to the endothelial surface immediately after ischemia-reperfusion (11,18). Yet, this process does not entirely explain observations of MB_PS attachment directly to the endothelium on intravital microscopy of post-ischemic tissue (11), nor the myocardial retention of Sonazoid when given in large doses to healthy subjects such as those in the current study. These mechanisms are still under investigation.

Our study was designed to test MCE molecular imaging in a population in whom ischemia was not in question and the risk area was identifiable. Our first task was to complete a dose-finding study in volunteers intended to minimize non-specific retention and enhancement. The dose that was eventually selected was about half of that used for contrast-enhanced ultrasound of the liver with Sonazoid (25), and far less than typical MB dosing during MCE perfusion imaging. An interesting observation was that Sonazoid at the dose selected produced substantially less non-specific myocardial signal in the remote territory in ACS subjects than in normal controls. This difference can be explained by ACS patients having larger body mass and comorbidities that affect signal intensity (obesity and lung disease), and the use of supplemental oxygen which can influence MB gas exchange and signal intensity over time (26). When considering non-specific retention, it is interesting to speculate that retention of agent in the microcirculation in non-acute settings could provide a simple method to assess myocardial vascular integrity after a single injection without concern for cavity attenuation, similar to what has recently been described with phase-transition ultrasound enhancing agents (27).

In the ACS population, MB_PS signal on MCE molecular imaging was more than five-fold higher in the adjudicated post-ischemic risk area than the remote region. Quantitative signal enhancement was high discriminatory for identifying ischemic from non-ischemic regions. Several features of the study indicate that MCE ischemic memory imaging will be successful in a lower risk population where diagnosis of ACS is less certain. Sonazoid signal enhancement in the risk area was independent of perfusion, the type of ACS, and the severity of ischemia. The latter finding indicates that retention of MP_PS requires only a certain threshold of ischemic injury and does not increase in a linear fashion with injury. This behavior is predictable from multivalent particle-based targeted contrast agents like MBs that mimic the behavior of cell adhesion in comparison to small molecule agents that bind in a 1:1 fashion with a target molecule (3). It is worth noting that the difference in signal intensity in the ischemic and remote territory in these patients with severe ischemia may have been narrowed by inflammatory activation in remote coronary territories in those with acute MI (28). Yet, remote territory signal enhancement was found to be lower than that produced by equivalent MB_PS doses in healthy controls.

There are several limitations of the study that should be highlighted. The design of this proof-of-concept study involved a population in whom ischemia was already recognized, and will be used to justify further investigation in populations such as symptomatic patients in whom diagnosis of ACS is not certain. The full mechanism for MB_PS retention also requires further study. Evaluation of the location and spatial extent of ischemia by a blinded reader was not performed in the current study. This type of analysis was avoided because many subjects continued to have severe wall motion abnormalities and reduced end-systolic wall thickness in the risk area (which can be appreciated in some of the examples shown, appreciated in the Central Illustration), making blinded interpretation impossible. In this study, we did not investigate optimal timing for molecular imaging acquisition which ideally should occur after clearance of most circulating agent, but before loss of gas volume from retained agent. The duration of delayed enhancement was not assessed, although signal is likely to be stable well beyond ten minutes based on the parasternal short-axis images (Figure 1) which were acquired late after apical imaging, and the lack of any signal decay between the first and last imaging plane (separated on average by 128±28 seconds) in the healthy control subjects. During perfusion imaging, we do not believe that retention of MB_PS in the risk area influenced our results because the duration of the post-destruction replenishment is very short and because most curves achieved a flat plateau rather than a continuous slow increased in signal. Finally, we have not definitively excluded the possibility that other inflammatory conditions, such as myocarditis, could produce similar findings.

In summary, we have demonstrated that MCE ischemic memory imaging is achievable in humans using signal enhancement from MB_PS retained in the post-ischemic microcirculation. This approach may allow rapid point-of-care detection of recent but resolved myocardial ischemia at the bedside in a patient population in whom diagnosis of ischemia is uncertain. While this study was performed with MCE technologies currently available, further optimization with respect to the composition of the MB_PS agent or MCE imaging methods may further improve the ability of discriminating retained MB signal in post-ischemic myocardium from non-specific retention in normal tissues.

Supplementary Material

1

Video 1. MB_PS Retention After High Dose Injection in a Healthy Control Subject in the Apical 4-chamber View. End-systolic triggered apical-4-chamber view MCE obtained 8 minutes after I.V. injection of Sonazoid (1.2 mL). The video illustrates signal enhancement pre-flash (first frame) followed by a high-power flash (frames 2–6), followed by several end-systolic frames that fail to show contrast reappearance.

2

Video 2.MB_PS Retention After High Dose Injection in a Healthy Control Subject in the Apical 3-chamber View. Apical 3-chamber view during real-time MCE imaging illustrating myocardial enhancement from MB_PS retention in all regions except the apex which was previously exposed to a high-power flash in the apical 4-chamber view (Video 1).

3

Video 3.Illustration of Absence of MB_PS Signal on Short-axis View Imaging in the “Null Lines” from Previous Apical Imaging. Parasternal short-axis view during real-time MCE imaging after previous flash-replenishment imaging in the apical views illustrates myocardial enhancement from MB_PS retention in all regions except the segments that were previously exposed to high-power flash in the apical 4-chamber, 2-chamber and 3-chamber views (see Figure 1E for labeling).

PERSPECTIVES.

Competency in Patient Care and Procedural Skills:

Non-invasive imaging is a guideline-recommended approach to assessment of the presence, location, and extent of myocardial ischemia in patients with chest pain when the diagnosis is uncertain. Ischemic memory imaging by pairing myocardial contrast echocardiography with phosphatidylserine-containing microbubbles can achieve this rapidly at the point-of-care.

Translational Outlook:

More research is needed to clarify the advantages and limitations echocardiographic ischemic memory molecular imaging to guide clinical management in various clinical settings.

ABBREVIATIONS

CAD

Coronary artery disease

MB

Microbubbles

MB_PS

Phosphatidylserine microbubbles

MBF

Myocardial blood flow

MBV

Microvascular blood volume

MCE

Myocardial contrast echocardiography

NSTEMI

Non-ST-segment elevation myocardial infarction

PCI

Percutaneous coronary intervention

STEMI

ST-segment elevation myocardial infarction

TIMI

Thrombolysis in Myocardial Infarction

WMSI

Wall motion score index