Functional Near-Infrared Fluorescence Imaging for Cardiac Surgery and Targeted Gene Therapy

Abstract

Cardiac revascularization is presently performed without realtime visual assessment of myocardial blood flow or perfusion. Moreover, gene therapy of the heart cannot, at present, be directed to specific territories at risk for myocardial infarction. We have developed a surgical imaging system that exploits the low autofluorescence, deep tissue penetration, low tissue scatter, and invisibility of near-infrared (NIR) fluorescent light. By completely isolating visible and NIR light paths, one is able to visualize, simultaneously, the anatomy and/or function of the heart, or any desired tissue. In rat model systems, we demonstrate that the heptamethine indocyanine-type NIR fluorophores IR-786 and the carboxylic acid form of IRDye78 can be injected intravenously in the living animal to provide real-time visual assessment of myocardial blood flow or perfusion intraoperatively. This imaging system may prove useful for the refinement of revascularization techniques, and for the administration of cardiac gene therapy.

Introduction

Cardiovascular disease remains the number one cause of mortality in the Western world [1]. In 1964, the first coronary artery bypass graft (CABG) [2] was performed and, in certain select patients, this procedure has been shown to improve survival [3–5]. Although many advances in surgical technique have been made, the CABG operation has not changed significantly since its inception. For example, assessment of the coronary arteries is performed via catheterization and X-ray fluoroscopy only before surgical intervention. Intraoperative monitoring of the heart is not performed routinely since fluoro-scopic dyes can be nephrotoxic, and such monitoring would expose patient and caregivers to additional X-rays.

Without such assessment, it is quite difficult to guide the administration of therapeutic modalities intraoperatively. A novel experimental strategy for treating myocardial ischemia is to induce neovascularization within areas of low blood flow in the heart. A number of mediators, such as vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF), have been shown to induce neovascularization within ischemic regions of the heart [6–8]. This strategy has been used in viable, ischemic regions (as detected by nuclear scans) not amenable to bypass grafting at the time of CABG. However, the ability to assess flow and viability measures intraoperatively would have significant advantages since it could guide surgeons directly to those specific areas of low perfusion and good myocardial viability.

To bypass the need for intraoperative X-ray fluoroscopy, and to permit imaging of specific targets, many investigators have developed visible light-based imaging systems [9]. Although these systems provide the surgeon with a unique view of cardiac function, they do not permit simultaneous imaging of anatomy and function, and suffer from the high background signals associated with visible wavelength fluorescence. These obstacles can be overcome by employing near-infrared (NIR) light.

The NIR portion of the spectrum is arbitrarily defined as those wavelengths above the sensitivity curve of the human eye (i.e., 700 nm) and below the infrared (typically 1000 nm). The range from 700 to 900 nm has been termed the “NIR window,” and is an area of the spectrum with minimal absorption by the biomolecules that constitute living tissue [10]. For visualizing the surfaces of living structures, planar NIR fluorescence imaging has several inherent advantages including low tissue autofluorescence, minimized tissue scatter, and relatively deep penetration depths (discussed in Ref. [11]).

Since NIR light is invisible to the human eye, it is possible to irradiate, simultaneously, a surgical field with visible light and NIR light. By using an optical system that maintains the isolation of each set of wavelengths, it is also possible to simultaneously “see” two images, one corresponding to visible light (what the surgeon is used to seeing) and the other that corresponds to NIR light (invisible to the surgeon). By introducing an NIR fluorophore into the surgical field, one can specifically image whatever function is imparted by the fluorophore. For example, by introducing a fluorophore that remains intravascular, one could visualize blood vessels within the surgical field, or by introducing a fluorophore that partitions into well-oxygenated tissue, one could visualize areas of relative ischemia.

In this study, we do both, by employing two different heptamethine indocyanine NIR fluorophores. We also describe an intraoperative NIR fluorescence imaging system that permits the surgeon to visualize, simultaneously, the anatomy and/or function of the beating heart. We demonstrate the use of this system for visual assessment of myocardial perfusion and viability, and discuss how it can be used for guiding revascularization and/or for targeting gene therapy to specific myocardial territories.

Materials and Methods

NIR Fluorophores

IR-786 perchlorate and sodium ICG were purchased from Sigma (St. Louis, MO). The carboxylic acid of IRDye78 was HPLC-purified after reaction of the N-hydroxysuccinimide ester of IRDye78 (LI-COR, Lincoln, NE) with water at pH 8.5 [12]. 10 mM stock solutions were stored in DMSO at −80°C in the dark.

Animals

Animals were used in accordance with an approved institutional protocol. Male Sprague–Dawley rats were from Charles River Laboratories (Wilmington, MA) and were anesthetized with 65 mg/kg IP pentobarbital.

Pharmacokinetics and Cardiac Biodistribution

IR-786 was diluted to 50 μM in phosphate-buffered saline (PBS) supplemented with 10% Cremophor EL (Sigma) and 10% absolute ethanol. IRDye78-CA was diluted to 50 μM in PBS. Rats (300 g) were injected intravenously with 50 nmol (1 mL) of each solution. Serum concentrations were analyzed fluorometrically as described previously [11]. To estimate cardiac uptake, hearts were removed and flash frozen in LN₂. Cubes (1 mm³) corresponding to areas of normal, ischemic, and scarred myocardium were minced, extracted in absolute methanol, and measured fluorometrically using fluorophore standards diluted in control heart extract.

In Vivo Cardiac Imaging

Anesthetized 350-g animals were ventilated on an SAR-830AP (CWE, Ardmore PA) ventilator and a midline sternotomy was performed. IR-786 was diluted to 25 μM in PBS supplemented with 5% Cremophor EL (Sigma) and 5% absolute ethanol. IRDye78-CA was diluted to 25 μM in PBS. 0.5 mL of each solution (12.5 nmol total, 0.036 μmol/kg) was injected intravenously via the tail vein. For cryonecrosis experiments, a cylindrical rod of dry ice, 2.5 mm in diameter, was applied to the surface of the heart with moderate pressure for 7.5 min. For subendocardial infarction experiments, the aorta was cross-clamped for 20 sec during left anterior descending (LAD) artery ligation and 0.5 mL of 10 μm red fluorescent microspheres (Molecular Probes, Eugene, OR) were injected into the apex. To visualize infarction, heart sections were incubated for 20 min at 37°C in 5% (w/v) 2,3,5-triphenyltetrazolium chloride (TTC; Sigma) in PBS prior to paraformaldehyde fixation.

The exposed heart was imaged using modifications to a previously described small animal imaging system [11]. NIR excitation was via a custom 771 nm, 250 mW laser diode system (Laser Components, Santa Rosa, CA) at a fluence rate of 50 mW/cm². White light excitation was via a 150-mW halogen lamp (Model PL-900, Dolan-Jenner, Lawrence, MA), depleted of wavelengths greater than 700 nm. The Orca-ER (Hamamatsu, Bridgewater, NJ) NIR camera settings included gain 7, 2 × 2 binning, 640 × 480 pixel field of view, and exposure time of 20 msec. Color video camera (HV-D27, Hitachi, Tarrytown, NY) images were acquired at 30 frames per second at a resolution of 640 × 480 pixels. Data were acquired and quantitated on a Macintosh computer equipped with a Digi-16 Snapper (DataCell, North Billerica, MA) frame grabber (for Orca-ER), CG-7 (Scion, Frederick, MD) frame grabber (for HV-D27), and IPLab software (Scanalytics, Fairfax, VA).

Results

NIR Fluorescence Contrast Agents

Heptamethine indocyanines used in this study were IR-786 and IRDye78-CA (Figure 1). IR-786 is a lipophilic cation of limited solubility that partitions into mitochondria and endoplasmic reticulum (ER) and, hence, can be used as a marker of tissue perfusion (A.N. and J.V.F., manuscript in preparation). IRDye78-CA is a highly charged IR-786 derivative that can be used to visualize blood flow. The properties of each fluorophore relevant to this study are shown in Figure 1A.

In initial in vitro experiments with rat neonatal cardiomyocytes, IR-786 exhibited rapid intracellular accumulation, with plateau levels achieved within 30 min (data not shown). For applied extracellular concentrations below 0.25 μM, IR-786 accumulation was exclusively mitochondrial, and for extracellular concentrations above 0.25 μM, accumulation was seen in both mitochondria and ER (data not shown).

Figure 1.

NIR fluorescence contrast agents and imaging system. (A) Chemical structures of the perchlorate salt of IR-786 (left) and the sodium salt of IRDye78-CA (right). Also shown are physicochemical and photoproperties of fluorophores in 10 mM HEPES, pH 7.4 (A.N. and J.V.F., manuscript in preparation). (B) Simultaneous color video/NIR fluorescence intraoperative imaging system. The NIR light path (dotted lines) is optically isolated from the visible (NIR-depleted) light path (solid lines) by use of separate excitation sources, a dichroic mirror, and barrier filters. All system components are under computer control (not shown). M.W. = molecular weight.

Simultaneous Color/NIR Intraoperative Imaging System

Since NIR light is essentially invisible to the human eye, it should be possible to irradiate a surgical field simultaneously with non-NIR (i.e., visible) and NIR light, and acquire reflected (or emitted) light separately. Such an imaging system is shown in Figure 1B. Visible light is depleted of all IR and NIR components, and NIR light is selected to match the excitation peaks of IRDye78-CA and IR-786. Light emanating from the surgical field is collected by a zoom lens and is split back into visible and NIR components using a dichroic mirror. A second level of filtration ensures that only visible light, or NIR fluorescence emission light, reaches their respective cameras. Since the system has no moving parts, and is under complete computer control, anatomic (color video) data can be acquired essentially simultaneously with functional (NIR fluorescence) data, thus permitting separate or merged display of the information.

Figure 2.

NIR fluorophore biodistribution and pharmacokinetics. (A) Serum concentration of IRDye78-CA (left) or IR-786 (right) over time was measured as detailed in Materials and Methods. Data are from three animals (mean ± SEM). (B) NIR fluorescence emission from the heart was quantitated on the small animal imaging system as a function of time after intravenous injection of 12.5 nmol IRDye78-CA (left) or IR-786 (right). For IRDye78-CA, the time during which the vascular signal is adequate for imaging is indicated by a black bar. Data are representative of four independent experiments. (C) Heart tissue concentration at 5 sec postinjection for IRDye78-CA (left) and at 1 min for IR-786 (right) in normally perfused, ischemic (sutured LAD), and scarred myocardium. Data are from three animals each. (D) Arteriovenous sampling, 20 sec postinjection of 25 nmol IRDye78-CA (left) or 50 nmol IR-786 (right). Data are from three animals each. Statistical significance was measured using the Student's t-test.

Figure 3.

IRDye78-CA as a blood flow indicator in vivo. Color images (left), NIR fluorescence images (middle), and pseudo-colored merged images (right) of the beating rat heart: autofluorescence (top panels), 10 sec after injection of IRDye78-CA (second panels), after ligation of a branch of the LAD artery with a suture (S) and reinjection of IRDye78-CA (third panels), reperfusion after 5 min of occlusion and reinjection of IRDye78-CA (bottom panels). Occluded vessels and a blood flow defect are visible in the NIR fluorescence image and appear as the absence of green pseudo-color in the merged image (white brackets). Data are representative of six independent experiments.

Pharmacokinetics and Cardiac Uptake in the Rat

Based on previously published pharmacokinetics in mice [11], IRDye78-CA was predicted to be intravascular, and exhibits rapid distribution and clearance. Indeed, peak serum levels of IRDye78-CA were achieved within 1 min postinjection in the rat (Figure 2A). Peak serum concentration (2.2 ± 0.1 μM) is consistent with complete distribution in the blood volume of the animal. IRDye78-CA exhibited a two-phase elimination with early and late phase half-lives of 5.0 and 19.3 min, respectively. NIR fluorescence emission from normal heart vasculature peaks at 5 sec, with signal-to-noise adequate for imaging between 2 and 15 sec with a 12.5-nmol intravenous injection (Figure 2B). Tissue concentration of IRDye78-CA in normal, ischemic, and scarred myocardium is shown in Figure 2C. Its arteriovenous gradient is shown in Figure 2D. In a modified Langendorff preparation, only 12% of fluorophore was retained after a 10-mL perfusion, confirming that IRDye78-CA is intravascular (data not shown).

IR-786, however, had vastly different behavior in vivo. Peak serum concentration was reached within 1 min, however, the absolute serum concentration was only 6% that of IRDye78-CA (Figure 2A). Moreover, a low, but detectable, serum concentration was measurable 1 hr postinjection. These data are most consistent with first-pass extraction of IR-786, with slow redistribution from tissue into the bloodstream. NIR fluorescence emission from the heart, as quantitated on the imaging system (see below), revealed peak fluorescence intensity 1 min postinjection (Figure 2B). Intensity then decreased in an exponential fashion, most likely due to metabolic conversion [13], photo-catalyzed conversion [14], and/or excretion, returning to baseline approximately 1 hr postinjection. Tissue concentration of IR-786 at 1 min postinjection in normal, ischemic, and scarred myocardium is shown in Figure 2C, with uptake in normal myocardium representing 1.7 ± 0.2% injected dose. This uptake compares favorably to the approximately 0.8% injected dose uptake in mice reported for ^99mTc-Sesta-mibi [13] (Cardiolite, DuPont Pharma, Wilmington, DE), a radioscintigraphic agent used routinely for assessment of myocardial perfusion. The arteriovenous gradient of IR-786 is shown in Figure 2D. In a modified Langendorff preparation, 82% of fluorophore was still retained after a 10-mL perfusion, confirming that IR-786 is extracted into the myocardium (data not shown).

In Vivo Imaging of Acute Vascular Occlusion Using IRDye78-CA

Being poly-sulphonated and highly charged, we hypothesized that IRDye78-CA would remain intravascular and permit high-sensitivity, high-resolution imaging of blood flow in the beating heart (Figure 3). NIR autofluorescence from the heart is negligible. When IRDye78-CA is injected intravenously, the entire vasculature is well delineated, with a signal-to-background ratio at 5 sec of 22:1. When a vascular occlusion was introduced by suturing a branch of the LAD artery, a large flow defect was visualized after reinjection of IRDye78-CA. After 5 min of occlusion, the suture was removed and IRDye78-CA was reinjected; vascular flow returned to baseline.

In Vivo Imaging of Acute Vascular Occlusion Using IR-786

The rapid uptake of IR-786 by mitochondria and ER, and its first-pass extraction from the bloodstream, suggested that it might function as an optical indicator of myocardial perfusion. In a normally perfused heart, IR-786 NIR fluorescence was homogeneous with an average signal-to-background ratio of 5:1 (Figure 4). When acute vascular occlusion was introduced by LAD artery occlusion and IR-786 was injected intravenously, myocardium distal to the occlusion had a 4-fold lower signal than well-perfused myocardium, and appeared black on the NIR image and “transparent” on the pseudo-colored merged image (Figure 4). After 5 min of vascular occlusion, the suture was removed. Within 1 min after restoration of blood flow, IR-786 fluorescence was uniform throughout the heart, suggesting that myocardium distal to the occlusion was adequately reperfused. Indeed, the heart continued to beat without any evidence of damage or toxicity for over 1 hr after suture removal. These results are in stark contrast to those seen with prolonged ischemia (see below). It should be noted that IR-786 did not have to be reinjected to acquire these revascularization images since the blood level of IR-786 (Figure 2A) was still adequate for imaging.

In Vivo Assessment of Myocardial Vascularity and Perfusion after Prolonged Vascular Occlusion

With short occlusion times, there was no apparent damage to heart vasculature (Figure 3) or myocardium (Figure 4) as judged by IRDye78-CA and IR-786 NIR fluorescence, respectively. We next tested how these patterns of distribution would change after prolonged ischemia. The vasculature of a normally perfused, resting heart was mapped using intravenous injection of IRDye78-CA (Figure 5). Prolonged ischemia was then induced by suturing a branch of the LAD artery. Reinjection of IRDye78-CA identified occluded vessels, venous engorgement, and poorly vascularized territory. After 30 min of ischemia, the suture was removed, and IRDye78-CA was reinjected. Hyperemia of the previously occluded vessels is seen clearly, although much of the distal myocardium is not well perfused. IR-786 was then injected approximately 2 min after the final IRDye78-CA injection. Unlike the homogenous uptake seen with transient ischemia (Figure 4), there was now a large defect suggesting that the myocardium was either at risk or destined for necrosis (Figure 5).

Figure 4.

IR-786 as a perfusion indicator in vivo. NIR autofluorescence (top panels), 1 min after injection of IR-786 (second panels), after ligation of a branch of the LAD artery with a suture (S) and injection of IR-786 (third panels), reperfusion after 5 min of occlusion (bottom panels). A perfusion defect is seen in the NIR fluorescence image and appears as the absence of green pseudo-color in the merged image (white arrows). Data are representative of six independent experiments.

Figure 5.

Intraoperative imaging of prolonged ischemia. Normally perfused heart assessed with IRDye78-CA (top panels), after ligation of a branch of the LAD artery with a suture (S) and reinjection of IRDye78-CA (second panels). Occluded vessels (arrowheads), a blood flow defect (white arrows), and vascular engorgement (V) are now visible in the NIR fluorescence and pseudo-colored merged images. After a total of 30 min of ligation, removal of the suture, and reinjection of IRDye78-CA (third panels), hyperemia of the previously ligated vessels is seen (white arrowheads). After injection of IR-786 (bottom panels), a large perfusion defect persists (white arrows). Data are representative of four independent experiments.

In Vivo Assessment of Area at Risk versus Area of Infarction

To address the sensitivity of the technology for assessing injury typical of ischemia/reperfusion, the rat heart was subjected to a 30-min LAD artery occlusion. During the occlusion, 10 μm red fluorescent microspheres were injected to mark the vascular territory at risk. At 24 hr postocclusion, the beating heart was imaged (Figure 6A). Red fluorescence imaging of microspheres, obtained simply by changing filter sets on the system, revealed a large, irregularly shaped territory at risk. Although not obvious from the color image, IRDye78-CA reveals marked dilation of arteries feeding this area, and IR-786 reveals an approximately 1.6-fold increase in perfusion from these dilated coronary arteries. Both fluorophores identify a 1.4-mm–wide, crescent-shaped band of presumed infarction extending towards the apex. The band is most obvious in the IR-786 image and accounts for 43.9% of the area at risk. When the same heart is sectioned and stained with TTC, a crescent-shaped area comprised of subendocardial and transmural infarction is confirmed (Figure 6B). Using these transverse sections, the infarcted area was 49.5% of the area at risk as measured by TTC staining, and 47.1% of the area at risk as measured by IR-786. These data suggest that penetration depth of NIR light in the rat is adequate to assess the full thickness of the myocardium.

In Vivo Assessment of Myocardial Vascularity and Perfusion after Epicardial Necrosis

We next examined the sensitivity of NIR fluorescence imaging for visualizing epicardial necrosis (Figure 7A). The vasculature of a normally perfused, resting heart was mapped using intravenous injection of IRDye78-CA. A localized area of necrosis, 2.5 mm in diameter, was then induced using cryofreezing. In the color image, this area is seen as a deep red circle with a bull's eye appearance. When IRDye78-CA was injected, the vascular defect in the area of necrosis was visualized, and vessels were seen to end abruptly at the boundary of necrosis. When IR-786 was injected, there was homogenous uptake throughout the normal myocardium, however, the area of necrosis appeared as a darkened area on the NIR fluorescence image and as a “transparent” area on the pseudo-colored merged image. After sectioning of the heart transversely, it was determined that the depth of necrosis was approximately 1.2 mm, with co-localization of NIR fluorescence and normal myocardium (Figure 7B). This depth of injury is consistent with the 30–40% decrease in emission intensity seen with the ventral view, and our prior modeling of NIR fluorescent light penetration depth [11].

Discussion

In this report, we describe a novel surgical imaging system that complements visible light imaging by exploiting otherwise invisible NIR fluorescence. Similar systems have been described for endoscopy [15] and microscopy [11]. By keeping visible and NIR light channels completely separate during data acquisition, one is able to choose whether anatomical information, functional information, or both are visualized in real time. Pseudo-coloring the functional image and merging it with the visible image is a particularly attractive way to present the data since it provides an “overlay” of anatomy and function. Although we present this system in the context of cardiac surgery and guided placement of cardiac gene therapy, its application in other types of surgery, especially oncologic surgery, in pathology, and in machine vision applications is apparent.

In previous work [11], we have estimated that the maximal imaging depth for detecting an embedded structure of low micromolar concentration by reflective NIR fluorescence imaging is approximately 3–5 mm due to the combined effects of tissue absorption and scatter. Specifically, fluorescence emission signal decreases exponentially from the surface of the target being imaged, and scatter reduces resolution. Despite this limitation, the subendocardial infarction data presented above suggest that, at least in the rat heart, this technology is capable of assessing the entire myocardium. When planar imaging (this study) is combined with angioscopic imaging of the endocardial surface (future studies), it may be possible to assess the full thickness of hearts as large as humans. It may also be possible to apply new and exciting frequency domain [16,17] or diffuse tomographic methods [18–20] for improving imaging depth.

Importantly, NIR fluorophores with different functionalities are being actively developed. To date, in vivo imaging of somatostatin receptors [21,22], proteases [23,24], and hydroxylapatite [11] have been reported. Indeed, annexin-V cardiac imaging systems that now work with visible light [9] could be adapted for use in the NIR. The prototype heptamethine indocyanine, indocyanine green (ICG), has been FDA-approved since 1956 for clinical use as a blood flow agent, and is one of the safest drugs ever administered to humans. In preliminary experiments, ICG worked well with our imaging system, although its limited aqueous solubility and quantum yield of >2% [16,25] are not optimal. IRDye78-CA, as a vascular agent, and IR-786, as a perfusion agent, now add to this growing library of compounds. We expect this library to expand even more rapidly now that robust methods have been described for the conjugation of heptamethine indocyanines to targeting molecules such as antibodies and low-molecular-weight ligands [12].

Figure 6.

Intraoperative imaging of subendocardial infarction. (A) Color video, IRDye78-CA, red fluorescent microsphere, and IR-786 images from the beating rat heart 24 hr after transient LAD occlusion. Arrows mark coronary artery dilatation and increased myocardial perfusion as revealed by IRDye78-CA and IR-786, respectively. Arrowheads mark crescent-shaped band of presumed infarction. Red fluorescent microspheres reveal the territory placed at risk during LAD occlusion. Data are representative of four independent experiments. (B) Heart from (A) after transverse sectioning and TTC staining. Left to right is base to apex. Sections were visualized for color (top), red fluorescence (middle), or NIR fluorescence (bottom). Areas of infarction appear white by TTC staining.

Figure 7.

Intraoperative imaging of epicardial necrosis. (A) The top panels display a normally perfused, beating heart after intravenous injection of IRDye78-CA. The middle panels display the same field after cryonecrosis of a small area of myocardium, and reinjection of IRDye78-CA. A blood flow defect is now visible in the NIR fluorescence image and appears as the absence of green pseudo-color in the merged image (white arrows). The bottom panels display the same field after injection of IR-786. The perfusion defect seen using IR-786 (white arrows) overlaps the blood flow defect seen using IRDye78-CA. Data are representative of five independent experiments. (B) To correlate NIR fluorescent signal with the depth of injury, cryonecrosis was induced, IR-786 was injected intravenously, and the heart was imaged in vivo (left panel). After bisecting transversely (i.e., along the dotted line in left panel), the heart was imaged with visible (middle panel) and NIR fluorescence (right panel) light. The area of cryonecrosis is well defined in the visible image and appears black in the NIR fluorescence image (white arrows).

This study did not specifically address the in vivo toxicity of IRDye78-CA or IR-786, and to date, no long-term safety data on these compounds have been published. In our experience, there is no apparent acute toxicity for IRDye78-CA when injected at up to 10 times the concentration used in this study. However, when IR-786 is injected at concentrations four times or greater than those used in this study, cardiac arrythmias are frequent during periods of ischemia (data not shown). This toxicity at high dose is likely mediated by the known interference of mitochondrial enzymes by lipophilic cationic indocyanines [26], and warrants in vivo use of as low a concentration of IR-786 as possible.

With respect to animal models of cardiovascular disease, the system we describe may have several applications. IRDye78-CA can be used to finely map myocardial blood flow and to identify areas that either require revascularization or that have not been adequately revascularized. Similarly, IR-786 can be used to assess myocardial perfusion, and may be useful for intraoperative monitoring, including transmyocardial laser revascularization [27,28]. During our preliminary experiments with cryonecrosis, we were surprised to learn how resistant myocardial vessels and tissue were to freeze/thawing. Even after direct application of −78.5°C to the myocardium for up to 2.5 min, rapid thawing resulted in no obvious change to blood flow or myocardial viability as assessed by IRDye78-CA and IR-786. The technology we describe may aid in the development of more effective cryopreservation techniques for transplanted organs, since one now has the ability to assess blood flow and viability in situ and nonisotopically.

Over the last few years, tremendous progress has been made in somatic gene transfer to cardiac cells and in cell transplantation to the heart [29,30]. Although these strategies hold great promise for treating heart failure and cardiac ischemia, the procedures themselves require precise targeting. In fact, transfer of anti-apoptotic genes to rescue myocardial cells or calcium cycling proteins to improve contractility in dysfunctional cardiomyocytes depends on recognizing viable myocardium [29]. The success of creating new vessels to ischemic areas of the myocardium also relies heavily on identifying low perfusion areas [8,31,32]. Cell insertion into infracted regions of the heart requires the detection of nonviable myocardium while at the same time asserting that the area has adequate flow to sustain delivery of nutrients to the stem cells introduced [33]. For these reasons, intraoperative visual assessment of blood flow and perfusion may someday improve the success of gene- and cell-based therapies for the heart.

The acceptance of new technology by clinicians often requires an evolution rather than a paradigm shift. The technology presented herein should be viewed as an embellishment to the visible light imaging already familiar to surgeons, and because of this, should be adaptable to any procedure for which sensitive and specific detection of a target is required.