In vivo imaging of rat coronary arteries using bi-plane digital subtraction angiography

Abstract

Introduction: X-ray based digital subtraction angiography (DSA) is a common clinical imaging method for vascular morphology and function. Coronary artery characterization is one of its most important applications. We show that bi-plane DSA of rat coronary arteries can provide a powerful imaging tool for translational safety assessment in drug discovery.

Methods: A novel, dual tube/detector system, constructed explicitly for preclinical imaging, supports image acquisition at 10 frames/s with 88-micron spatial resolution. Ventilation, x-ray exposure, and contrast injection are all precisely synchronized using a biological sequence controller implemented as a LabVIEW application. A set of experiments were performed to test and optimize the sampling and image quality. We applied the DSA imaging protocol to record changes in the visualization of coronaries and myocardial perfusion induced by a vasodilator drug, nitroprusside. The drug was infused into a tail vein catheter using a peristaltic infusion pump at a rate of 0.07 mL/h for 3 min (dose: 0.0875 mg). Multiple DSA sequences were acquired before, during, and up to 25 min after drug infusion. Perfusion maps of the heart were generated in MATLAB to compare the drug effects over time.

Results: The best trade-off between the injection time, pressure, and image quality was achieved at 60 PSI, with the injection of 150 ms occurring early in diastole (60 ms delay) and resulting in the delivery of 113 μL of contrast agent. DSA images clearly show the main branches of the coronary arteries in an intact, beating heart. The drug test demonstrated that DSA can detect relative changes in coronary circulation via perfusion maps.

Conclusions: The methodology for DSA imaging of rat coronary arteries can serve as a template for future translational studies to assist in safety evaluation of new pharmaceuticals. Although x-ray imaging involves radiation, the associated dose (0.4 Gy) is not a major limitation.

Figure 1

For coronary artery DSA in the rat, an imaging sequence starts by bringing the x-ray anode rotor up to speed and suspending ventilation at end-expiration. The interval between each R wave is ~ 300 ms. Each 10-ms exposure and subsequent camera readout is initiated at the same delay after the peak R wave. A series of non-contrast images are acquired to create the mask. A 150-ms injection of Isovue is triggered followed by the acquisition of post-contrast injection images, again at the same point in the cardiac cycle. The whole sequence lasts about 8-10 seconds yielding a stack of 30 different frames.

Figure 2

A dedicated image pipeline has been developed in MATLAB to automatically process the serial DSA images. We generate the average mask from pre-injection frames and subtract it from the post-contrast images to produce the vascular only DSA images.

Figure 3

The effect of catheter tip position on visualizing coronary arteries is illustrated for two different animals. (A) and (C) are ventral/dorsal projections, and (B) and (D) are lateral projections. In first animal (A, B), the catheter tip appears to be touching the cranial surface of the aortic valve (red arrow) and apparently obstructing the orifice of the left coronary artery, whereas in second animal (C, D), the tip of the catheter (blue arrow) is well above the valve, and both arteries and branches are clearly seen.

Figure 4

DSA images corresponding to the heartbeat with the maximum enhancement of the coronary arteries selected for tests where either the injector pressure or the time for injection were varied. Note that the best trade-off between the injection time, pressure, and image quality is achieved at 60 PSI, 150 ms injection (see green square).

Figure 5

The injected volume is linearly related to the pressure driving the injector and the injection duration. A visual inspection of Figure 4 favors the image corresponding to 60 PSI, 150 ms corresponding to a volume of injected contrast agent of 113 microL. This volume is sufficient.

Figure 6

The tests involved multiple injections (60 PSI, 150 ms) at different delay times from the R peak. The RR-interval was approximately 200 ms. Note that the visualization is slightly improved when the injection occurs early in diastole, 60 ms delay.

Figure 7

These DSA images illustrate the effects of nitroprusside (NP) infusion on coronary arteries. Row A shows selected images before NP infusion (−5 minutes), during NP infusion (2 to?and 3 minutes), and after infusion has stopped (5 to 20 minutes). Note that dilation of the coronary arteries occurs within minutes of infusion and persists for at least 7 minutes after cessation of NP infusion, and after 17 minutes, the coronary arteries have returned to pre-infusion caliber. Row B shows the first 7 frames of the pre-infusion sequence (−5 minutes) and row C shows the first 7 images of the 8-minute sequence. In row C, after the first frame, note the evidence for myocardial perfusion as the ventricular walls are darkened.

Figure 8

Perfusion maps from the study shown in Figure 7. Note the maximum perfusion occurs in the 8-minute sequence, 5 minutes after cessation of NP infusion. 25 minutes later, the perfusion levels have returned to pre-infusion levels.