A Novel Technique for Endovascular Detection and Removal of Radiographic Contrast during Angiography

ABSTRACT: Objectives. This study aims at in-vitro validation of the principles of endovascular detection of contrast medium and assessing the feasibility of in-vivo detection and removal of contrast during angiography. Background. Contrast-induced nephropathy is a growing concern in current percutaneous interventions with increasing lesion complexity and patient comorbidity. To address this clinical problem, a novel method of endovascular detection and automatic removal of contrast has been developed, and is comprised of a catheter-based system with a reflectance-type optical sensor. Methods. Blood samples were obtained from ovine subjects to characterize the optical response of blood by measuring the reflectance spectrum at varying levels of hematocrit diluted by a contrast agent. The results from the in-vitro test were implemented into an in-vivo system. An aspiration catheter equipped with a fiberoptic sensor was inserted into the coronary sinus (CS) of 5 canines. Contrast was administered through the coronary artery and reflectance signals were recorded at the CS. The removal rate was analyzed through 20 specimen collections. Results. A proportional relationship was found between hematocrit and reflectance intensity in in-vitro test. Upon in-vivo detection of contrast, the sensor signal showed a 79.5 ± 9.9% (n = 33) drop from the pre-injection baseline. This was highly reproducible and beyond the noise level of baseline, (2.5 ± 0.9%), enabling automatic activation of the aspiration system. The signal duration was 12.2 ± 3.7 seconds. The removal rate of contrast was 59.3 ± 11%. Conclusion. The present study validated the principles of endovascular contrast detection and demonstrated the feasibility of an in-vivo, catheter-based removal of contrast using reflectance technology. Keywords: contrast-induced nephropathy; contrast detection; intravascular optic sensor J INVASIVE CARDIOL 2009;21:314–318 The rising number of cardiac catheterization procedures, which increased by 390% between 1979 and 2002 in the U.S., has led to increased administration of radiographic contrast agents.1 A serious complication associated with exposure to iodinated contrast media is contrast-induced nephropathy (CIN), which is commonly defined as acute renal dysfunction occurring within 24–72 hours of exposure to radiographic contrast media. CIN is associated with increased in-hospital mortality, poor long-term survival rates and increased costs.2–5 Current treatment for CIN is mainly symptomatic, involving systemic administration of pharmacologic agents, often resulting in limited success.6 The concept of removal of contrast media from the coronary sinus before it escapes the injected site has been suggested.7–9 Previous attempts to remove contrast from the coronary sinus required lumen occlusion and lacked automaticity as well as contrast detection. Recently, a novel intravascular sensing system was developed that provides active and continuous monitoring of the contrast media in the coronary sinus, coupled with an aspiration mechanism.10 We hypothesized that intracoronary administration of contrast media dilutes local blood (i.e., local hematocrit [HCT]), leading to transient changes in optical signals. As a result, the presence of contrast can be detected by endovascular reflectance. This study was performed to assess the mechanism underlying endovascular reflectance and to evaluate the feasibility of the sensing modality for catheter-based, in-vivo detection and removal of contrast media at the coronary sinus following coronary angiography. Materials and Methods The present study involved in-vitro and in-vivo experiments. The in-vitro study was performed to investigate the mechanisms underlying endovascular reflectance of blood, which validates the applicability of reflectance sensing for detecting contrast media and to provide technical reference for the sensing system. Based on the in-vitro study results, the in-vivo sensing system was implemented and the in-vivo study tested the feasibility of real-time detection of contrast media using the endovascular reflectance in the canine coronary sinus. In-Vitro Study Various blood samples from porcine, ovine and canine subjects were examined for in-vitro tests in our preliminary tests, which showed no significant difference in overall optical characteristics between them except amplitude. Thus, blood samples were obtained from ovine subjects because of their availability in the animal lab. Prior to scheduled euthanasia under supervision of the Veterinary Authority of Stanford University, blood samples were collected. Once collected, the blood samples were heparinized and stored under refrigerated conditions. In-vitro reflectance was measured with a fiberoptic-based probe (Ocean Optics R400-7-UV-VIS, Dunedin, Florida). This spectral value represents the wavelength dependency of light backscattered from blood due to hemoglobin absorption. The value is provided for approximately every 0.24 nm within the range of 500–1000 nm. Spectral signals were recorded by a spectrometer (Horiba Jobin Yvon VS140, Edison, New Jersey) and analyzed by the software provided by the spectrometer manufacturer (Horiba Jobin Yvon). A representative contrast agent (Visipaque®, GE Healthcare, Chalfont St. Giles, United Kingdom) was used to evaluate reflectance spectra of different concentrations of the contrast agent mixed with ovine whole blood, which will provide background information regarding reflectance signal changes as HCT varies. In-Vivo Study Optical catheter. A catheter equipped with a fiber-optic probe (SENTINEL Catheter, Catharos Medical Systems, Campbell, California) shown in Figure 1 was used for the in-vivo contrast detection and removal test. The Sentinel Catheter is an aspiration catheter (10 Fr) having a central aspiration lumen, integrated fiberoptics, and an expanding basket tip. The aspiration line and fiberoptics extend extracorporeally to a control module. The active sensor using red light of 627 nm wavelength is displayed in Figure 1B. Animal models. Five mongrel dogs were used for in-vivo testing of endovascular reflectance. The animals were anesthetized according the existing protocols of the Stanford University Animal Care & Experimental Cardiology Department. Coronary sinus catheterization procedure. Standard techniques to obtain access to the canine coronary sinus were used. A 10 Fr introducer sheath (Cordis Corp., Miami Lakes, Florida) was inserted into the canine left internal jugular vein (IJ). Subsequently, a 5 Fr Swan-Ganz Catheter (Edwards Life Sciences, Irvine, California) was introduced through the venous introducer sheath into the IJ, the superior cava vein and subsequently into the right atrium. Correct placement of the guidewire was monitored based on the expected curve shape of the coronary sinus. A 0.035 inch hydrophilic GlideWire (Terumo Corp., Japan) was introduced through the Swan-Ganz catheter into the canine coronary sinus all the way through the great cardiac vein. Once the GlideWire was in place in the CS, the Swan-Ganz catheter was withdrawn and the Catharos Sentinel catheter was prepared as follows: The Sentinel catheter was connected to the aspiration line, which was primed with physiological saline (0.9%) and introduced into the canine coronary sinus over a previously deployed guidewire. With the catheter prepared and its tip retracted, it was advanced over the guidewire. Once the tip was positioned at the desired site of contrast removal, the basket can be deployed by holding the main body of the handle and slowly pulling back the front part of the handle. The catheter tip (basket) was introduced over the GlideWire, and, subsequently, the catheter was introduced into the proximal coronary sinus over the guidewire. Fluoroscopic guidance provided confirmation that the basket was satisfactorily expanded in the proximal coronary sinus. At this point, the Sentinel system was considered deployed and ready for use. The ideal deployment for the Sentinel system is illustrated in Figure 2A, and the actual deployment is shown in Figure 2B by fluoroscopic image. Coronary arterial catheterization and contrast injections. Catheterization of the canine coronary arterial system was carried out in accordance to standard procedures. An introducer sheath was placed in the canine left carotid artery, and an 8 Fr Judkins left coronary guiding catheter (Cordis) was introduced retrograde into the aortic root, where the canine left main coronary artery was accessed. A standard contrast agent (Omnipaque® 300 mg/mL, GE Healthcare) was used throughout the procedure. The undiluted, pre-warmed contrast agent was injected using a 10 Fr syringe/manifold system through the coronary guiding catheter into the coronary arteries. Due to the small size of the canine subjects, the injection volumes were limited to 5 cc per injection. In 2 canines, the ex-vivo hearts were formaldehyde-fixed and sent for tissue analysis by an independent histopathology laboratory. Quantification of contrast contents (assay). A spectrophotometric absorbance assay was utilized to determine the contrast removal rate, which is the amount of contrast media in the blood-contrast aspirant solution as described by Michishita et al.7 They conducted an iodine concentration study using spectrophotometry with diluted blood samples and found a high correlation (R2 = 0.993) between this method and the original concentrations. We have performed similar spectrophotometry using ultraviolet light to measure optical absorption versus contrast concentration in diluted contrast-blood mixture samples. The optical absorption is proportional to the contrast concentration and we have been able to confirm our test results which showed a high correlation (R2 = 0.980) between our method and the original concentrations. Our variability tests have determined the accuracy of the assay to be > 95%. Results In-vitro reflectance spectra of blood as a function of blood dilution (HCT). Figure 3A shows reflectance spectra under varying dilution levels of ovine arterial blood diluted with the contrast agent (Visipaque). There was a proportional relationship between reflectance intensity and degree of dilution, which explains the variations in reflectance amplitude as the HCT changes. Figure 3B shows the specific reflectance variations at the wavelength of 627 nm to provide reference data of the wavelength used in the SENTINEL system. No change was observed in the overall pattern of reflectance spectrum. In-vivo contrast detection. The Catharos Sentinel system was able to register changes in reflectance signal upon fluoroscopic contrast appearance in the coronary sinus following arterial contrast injection. An example of a contrast signal as registered by the Sentinel detection system is shown in Figure 4. It indicates that if blood is diluted by contrast in-vivo, the optical signal varies due to HCT changes as shown in in-vitro test (Figure 3B). The overall in-vivo signal pattern (Figure 4) is close to an inverted bell shape and was reproducible at most injections and in different canine subjects. Upon detection, the signal rapidly decreased by an average of 79.5 ± 9.9% (range 54.1–92.9%; n = 33) from the baseline voltage. Comparatively, the baseline noise level averaged 2.5 ± 0.9% (range 0.9–4.3%; n = 33). The duration of the detection signal was 12.2 ± 3.7 sec (range 6.5–24.3 sec; n = 33). The results of signal analysis are summarized in Table 1. Contrast removal and recovery analysis. Evacuation of the canine coronary sinus was initiated at a signal drop beyond 15% of baseline and was terminated when the signal returned to within 10% of baseline. Contrast removal attempts were performed during four canine studies. Analysis of the collected samples revealed contrast removal rates of 59.3 ± 11% (range 42–76%; n = 20) of total injection volume (77.8 ± 34.6 ml, range 33–124 ml), depending on the system settings from the series of 20 samples. Total extracted volume of blood for each animal was 668.7 ± 196.7 ml (range 383–880 ml), depending on the experimental nature of the procedures, where varying settings of aspiration flow rates, pump triggers and basket positions were tested. Histological study. Analysis involved gross pathology and microscopic inspection of all heart tissues, with special focus on the integrity of the coronary sinus wall structures and its major tributaries (i.e., great cardiac vein, anterior interventricular vein and the middle cardiac vein). In summary, the histologic study showed the outflow venous channels of the heart, with no significant mural injury to the coronary sinus, great cardiac vein or the anterior interventricular vein of the heart. Their lumens were patent, with no evidence of thrombosis or hemorrhage. There were patchy areas of superficial endothelial cell loss, with no disruption or break in their underlying stromal framework. No significant mural inflammation was present. Discussion The present in-vitro findings established the mechanism underlying blood reflectance measurements in that it demonstrated that reflectance is proportional to HCT. The in-vivo study demonstrated feasibility of an endovascular catheter-based reflectance system in detecting radiographic contrast agents, and the practicality of removing it from a venous site following angiographic injections. Applicability of endovascular reflectance in endovascular detection of contrast. The in-vitro study revealed that the maximum reflectance of blood is registered in the range of 600–700 nm. This would be the ideal wavelength for in-vivo endovascular reflectance. The analysis also showed that blood reflectance is affected by local properties of the blood such as HCT variations. These findings are compatible with previous studies.11,12 In-vivo, where a defined anatomical location exhibits constant oxygenation conditions, transient changes in local HCT (following local, injections of agents) directly translate into changes in blood reflectance signals. The in-vivo study also demonstrated a proportional relationship between HCT and reflectance intensity. Furthermore, the in-vivo study revealed a U-shaped contrast signal in the coronary sinus, which potentially represents, to a large extent, the arterially injected bolus, with maximum contrast concentration (hence maximum reflectance signal drop) in the middle segments of the signal. Thus, the sensitivity of this mechanism can be employed in the detection and removal of locally injected contrast agents. The most important rationale in using different contrast agents and different animal blood samples in in-vitro and in-vivo studies is to show the applicability of our technology to various situations. Feasibility of in-vivo contrast detection and removal. During contrast injection, the blood is locally diluted producing changes in local HCT that can be registered with reflectance intensity as demonstrated in the present in-vivo study. The signal-to-noise ratio is favorable and sufficient to serve as a trigger for aspiration of the contrast column at the time of detection. This was also shown during the in-vivo segment of the present study. Several previous attempts were aimed at removing contrast from the coronary sinus, however, none of them utilized endovascular detection for triggering aspiration. Additionally, the present method provides a convenient, active system for contrast removal that does not require the physician’s attention during the procedure. Optimization of detection/removal performance. As mentioned previously, Figure 4 shows typical contrast detection signals observed throughout our in-vivo studies. The reflectance signal was somewhat noisy, but the detection could be clearly distinguished from baseline signals. The aspiration threshold, at this moment, was set to down-crossing 4 volts, but some portion of contrast prior to crossing the threshold might have already passed and escaped the collection site. Thus, it is important to design an algorithm to automatically catch the falling edge caused by dilution, and not by noise, to maximize the removal rate and minimize blood loss. The first approach would be the optimization of baseline signals to reduce the noise level as much as possible. This would be a prerequisite for optimal detection of actual falling edges of contrast injections. The performance of aspiration would benefit from these optimized signals and would improve the removal rate more consistently. Another approach would involve design optimization of the flow converging mechanism by flow dynamic analysis to identify flow patterns at the deployment site. It is necessary to redirect the total volume toward the sensor so as not to miss any small portion of the injected contrast that could pass through the outer lumen area. More detailed investigations involving other possible improvements are indicated. Potential impact/benefit of the system. As previously reported, the volume of arterially injected contrast is a well-established risk factor in the development of CIN.3,4 Additionally, the CIN Task Force recommends limiting the volume of contrast during angiography and percutaneous coronary intervention in high-risk patients to Conclusion This present study shows the feasibility of the active monitoring and removal of contrast media injected during angiography based on a fiberoptic catheter-based endovascular reflectance system. Despite being in the pre-clinical stage, this study showed possibility for a more efficient intravascular optical sensing system to reduce the risk of renal complications caused by contrast agents for patients undergoing angiography or PCI. Acknowledgment. We would like to express our gratitude to Jennifer Lyons, RN, and Fumiaki Ikeno, MD, at the Interventional Cardiovascular Research Laboratory, Division of Cardiovascular Medicine, Stanford University Medical Center for their generous help in providing assistance during our in-vitro and in-vivo studies. From the *Center for Cardiovascular Technology, Stanford University Medical Center, Stanford, California, §Catharos Medical Systems, Inc., Campbell, California, £Weldon School of Biomedical Engineering, Purdue University, West Lafayette, Indiana, and the ∞Feinberg School of Medicine, Northwestern University, Chicago, Illinois. Disclosures: Dr. Hassan is an equity holder in and Mr. Lloyd and Drs. Davidson and Fitzgerald are consultants to Catharos Medical Systems. Manuscript submitted December 2, 2008, provisional acceptance given January 13, 2009, final version accepted March 2, 2009. Address for correspondence: Peter J. Fitzgerald, MD, PhD, Center for Cardiovascular Technology, Stanford University Medical Center, 300 Pasteur Drive, Room H3554, Stanford, CA 94305. E-mail: crci-cvmed@stanford.edu

1. Katzberg RW. Contrast-induced nephrotoxicity: Clinical landscape. Kidney Int 2006;69:S3.

2. McCullough PA, Wolyn R, Rocher LL, et al. Acute Renal Failure After Coronary Intervention: Incidence, Risk Factors, and Relationship to Mortality. Am J Med 1997;103:368–375.

3. Dangas G, Iakovou I, Nikolsky E, et al. Contrast-Induced nephropathy after percutaneous coronary interventions in relation to chronic kidney disease and hemodynamic variables. Am J Cardiol 2005;95:13–19.

4. Bartorelli AL, Marenzi G. Contrast-induced nephropathy. J Interv Cardiol 2008;21:74–85.

5. Tepel M, Aspelin P, Lameire N. Contrast-induced nephropathy: A clinical and evidence-based approach. Circulation 2006;113:1799–1806.

6. Kelly AM, Dwamena B, Cronin P, et al. Meta-analysis: Effectiveness of drugs for preventing contrast-induced nephropathy. Ann Intern Med 2008;148:284–294.

7. Michishita I, Fujii Z. A novel contrast removal system from the coronary sinus using an adsorbing column during coronary angiography in a porcine model. J Am Coll Cardiol 2006;47:1866–1870.

8. Movahed M-R, Wong J, Molloi S. Removal of iodine contrast from coronary sinus in swine during coronary angiography. J Am Coll Cardiol 2006;47:465–467.

9. Danenberg HD, Lotan C, Varshitski B, et al. Removal of contrast medium from the coronary sinus during coronary angiography: Feasibility of a simple and available technique for the prevention of nephropathy. Cardiovasc Revasc Med 2008;9:9–13.

10. Fitzgerald PJ, Courtney BK, Hassan A; Catharos Medical Systems, Inc., Los Gatos, CA (U.S.), assignee. Methods and devices for retrieval of a medical agent from a physiological efferent fluid collection site. U.S. patent 7,211,073 B2, 2007.

11. Steinke JM, Shepherd AP. Reflectance measurements of hematocrit and oxyhemoglobin saturation. Am J Physiol – Heart Circ Physiol 1987;253:H147–H153.

12. Reynolds L, Johnson C, Ishimaru A. Diffuse reflectance from a finite blood medium: Applications to the modeling of fiber optic catheters. Applied Optics 1976;15:2059.

13. McCullough PA, Adam A, Becker CR, et al. Risk prediction of contrast-induced nephropathy. Am J Cardiol 2006;98:27–36.