Article Text
Abstract
Background Endovascular technological advances have revolutionized the field of neurovascular surgery and have become the mainstay of treatment for many cerebrovascular pathologies. Digital subtraction angiography (DSA) is the ’gold standard' for visualization of the vasculature and deployment of endovascular devices. Nonetheless, with recent technological advances in optics, angioscopy has emerged as a potentially important adjunct to DSA. Angioscopy can offer direct visualization of the intracranial vasculature, and direct observation and inspection of device deployment. However, previous iterations of this technology have not been sufficiently miniaturized or practical for modern neurointerventional practice.
Objective To describe the evolution, development, and design of a microangioscope that offers both high-quality direct visualization and the miniaturization necessary to navigate in the small intracranial vessels and provide examples of its potential applications in the diagnosis and treatment of cerebrovascular pathologies using an in vivo porcine model.
Methods In this proof-of-concept study we introduce a novel microangioscope, designed from coherent fiber bundle technology. The microangioscope is smaller than any previously described angioscope, at 1.7 F, while maintaining high-resolution images. A porcine model is used to demonstrate the resolution of the images in vivo.
Results Video recordings of the microangioscope show the versatility of the camera mounted on different microcatheters and its ability to navigate external carotid artery branches. The microangioscope is also shown to be able to resolve the subtle differences between red and white thrombi in a porcine model.
Conclusion A new microangioscope, based on miniaturized fiber optic technology, offers a potentially revolutionary way to visualize the intracranial vascular space.
- angiography
- endoscopy
- technology
- technique
Statistics from Altmetric.com
Introduction
Endovascular technological advances have revolutionized the field of cerebrovascular surgery. From studies of the first detachable coil in the early 1990s by Guglielmi et al to the FDA approval of the first flow diverter in 2011, technology has advanced the field of neurointerventional surgery.1–3 Each innovation provides a new treatment option for complex pathologies and also requires a parallel advancement in imaging to allow safe and accurate deployment.
All modern neuroendovascular techniques rely on fluoroscopy and iodinated contrast to guide the positioning and deployment of devices. Radiation-induced complications of fluoroscopy include skin burns and hair loss, which can occur at doses as low as 3 Gy.4 5 Furthermore, contrast-related nephropathy has been reported to occur in approximately 20–30% of patients with pre-existing renal disease and in up to 5% of individuals at low risk.5 6 In addition, operator radiation, even with proper shielding and clothing, is not trivial, with reports of up to 254 μGy of radiation to the operator’s hands and eyes for each case; the dose is cumulative throughout the operator’s career.7 Lastly, indirect visualization in cases with difficult anatomy may contribute to malpositioned devices, leading to thromboembolic and hemorrhagic complications.5 8 9
Because of these limitations, other new methods of intravascular visualization have been studied. Intravascular ultrasound (IVUS) employs an ultrasound microcatheter to provide real-time cross-sectional images of the catheterized artery and has been used to study the progression of atherosclerotic disease and complications after extracranial stenting in vivo.10 11 In one clinical application, IVUS was shown to detect in-stent plaque protrusion after carotid stenting with high sensitivity relative to standard digital subtraction angiography (DSA).12–15
Another emerging method of intravascular visualization is optical coherence tomography (OCT), which involves using a fiber optic wire that emits near-infrared light producing a signal based on the scattered or reflected light off the surrounding tissue, and can be used clinically to evaluate atherosclerotic disease. In one case report, OCT was used to assess the optimal placement of a carotid stent and was found to have an even higher resolution of the arterial wall than IVUS.16 In a separate report, Griessenauer et al attempted to use OCT to evaluate endothelialization of a Pipeline embolization device in follow-up but were unable to navigate the device through the carotid siphon.17 Nonetheless, both IVUS and OCT fall short as they offer neither direct visualization of intravascular pathology nor the miniaturization necessary to be compatible with the intracranial vasculature.
Adapted endoscopic techniques have been reported that allow direct visualization of the extracranial neurovasculature. Since the early 1990s, direct endoluminal optical visualization, or angioscopy, has been proposed as the future of adjuvant endovascular therapy, with in vivo angioscopic studies of the aorta,18 coronary arteries,19 and carotid arteries.20 Most recent studies use scanning fiber endoscopes (SFEs), which are now becoming increasingly more miniaturized in an effort to view more distal vasculature. This technology is based on reflectance and laser-induced fluorescence emission of intrinsic fluorescent constituents in the vascular tissue.21 22 In two proof-of-concept studies, Savastano et al showed that SFE angioscopy generates high-quality spectral images in human cadaveric and live porcine models of the carotid arteries.22 However, they failed to navigate beyond the petrous carotid artery in the cadaveric study. Coherent fibre bundle (CFB) is an alternative angioscopic technology to SFE; it involves a densely packed array of small-diameter glass fibers that relay images from the vessel lumen.23 Previous CFB technology also had limited applications in the distal vasculature. Indeed, the two aforementioned studies describe angioscopes mounted on 7 and 5 French (F) guide catheters, respectively, too large for intracranial vessels.
In this study, we introduce a novel microangioscope that offers both high-quality direct visualization and the miniaturization necessary to navigate in the small intracranial vessels. We will describe the evolution, development, and design of the microangioscope and provide examples of its potential applications in the diagnosis and treatment of cerebrovascular pathologies using an in vivo porcine model.
Methods
Technological gap and evolution
Angioscopy has had limited neuroendovascular use owing to its poor ability to visualize through blood and the size restrictions of previous technologies. After recent technological breakthroughs, we have found that CFB-based angioscopes provide better miniaturization than those based on complementary metal oxide semiconductors (CMOSs) or SFEs. A CMOS is chip-on-tip technology that features a small camera at the end of the imaging probe, and is limited by the size of the physical sensor at the distal tip of the imaging probe. SFEs use leached fiber bundles that continuously oscillate, relying on complex image processing software to generate interpretable images, and involving stiff ferrules at the distal tip of the devices to collimate the individual fibers. These ferrules make the distal tip rigid (difficult for intravascular use) and also further restrict the size. In contrast, CFB uses fixed optics and an external charge-coupled device (CCD) camera to reliably generate images while making it possible to minimize the diameter of the interventional imaging probe owing to the small individual fibers.
Past attempts at miniaturization of CFB fiberscopes almost universally failed to meet the size and flexibility requirements needed to be useful in small, tortuous blood vessels. Prior fiberscopes that began to achieve suitably small diameters and other favorable physical properties were also plagued by poor resolution and overall image quality. This inhibited their usefulness and clinical adoption. In general, CFB fiberscopes can be manufactured at lower cost than CMOS and SFE technology, which makes them better candidates to be disposable—essential for clinical endovascular use. Real-world sterilization has poor adherence and detracts from standard clinical workflow. Thus, the next evolution of CFB angioscopy is one with greater miniaturization, flexibility, and cost effectiveness—the microangioscope.
Development and design of a novel microangioscope
Recent advances in telecommunication fiber optic cables allow for thinner coherent image bundles than previous attempts at intravascular endoscopes. The fiberscope used in this study uses CFB technology and features a 285 µm outer diameter coherent image bundle with 3000 silica glass fibers, representing 3000 individual image pixels. The fibers are approximately 2.5 µm in diameter and have a 3.3 µm center-to-center spacing. The image fibers all share a common cladding, a jacket and a coating. The image bundle is surrounded by uncladded, silica illumination fibers that are 25–50 µm to increase flexibility and reduce the overall bend radius. A 250 µm outer diameter GRIN lens with a 70° field of view in air and 50° field of view in water is used to provide acceptable viewing angles while maintaining a low profile. The lens is designed to be as short as possible at 500 µm to ensure the distal tip of the imaging device is not prohibitively stiff. Typical angioscopes make use of polymer barrels at the distal tip to house the distal end of the image bundle and lens. This was forgone to achieve a minimal distal profile with maximum flexibility (figure 1).
The proximal end of the microangioscope is connected to a CCD camera via a coupler fitted with relay lenses to allow the image to be focused onto the CCD imaging array. As the number and size of both the image and illumination fibers were kept as small as possible, it was crucial to optimize the CCD array, CCD camera parameters, and the light source, to allow as much light as possible to be transmitted. Frame rate is limited to approximately 30 Hz to allow for adequate CCD sensor exposure time while maintaining acceptable intravascular video quality.
Compared with prior CFB designs, the current microangioscope has fewer image fibers (while maintaining an adequate image resolution) but offers a smaller distal tip, a shorter rigid distal tip, and a smaller bend radius (improved flexibility). The device is also made to be disposable and can connect with standard endoscopy carts.
Resolution of the microangioscope
The microangioscope has a relatively high pixel density for its size, consisting of approximately 113 000 pixels/mm2, with 3000 pixels in total. Thus, the active imaging area is 27 225 µm2. The lens has a resolution of 400 line pairs/mm at the center, and 150 line pairs/mm at 0.8 times the radius.
One study by Lee et al reviewed SFE versus CMOS and CFB devices for high quality, miniaturized, and flexible clinical application.24 They determined the pixel density of their tested SFE device to be 345 000 pixels/mm2. For reference, the CFB device from this study has approximately 113 000 pixels/mm2, a resolution approximately a third of an SFE and half that of CMOSs, at 238 000 pixels/mm2. However, they also found that the rigid tip of their SFE device is 9 mm long, while the shortest rigid tip of a CMOS device is estimated to be 5 mm at its distal tip. This is compared with the 0.5 mm tip of the CFB microangioscope. In other words, the smallest diameter of SFE-based devices is approximately 3.7 F, and the smallest CMOS is also >3 F in diameter, compared with the 1.7 F tip of the CFB microangioscope described in this study (online supplementary table 1). Ultimately, with longer tip lengths, SFE and CMOS devices would be unlikely to reach beyond the tortuous, distal segments of the internal carotid artery (ICA), severely limiting their usefulness in intracranial neuroendovascular practice.
Supplemental material
Implementation and catheter selection
To allow the imaging probe to be used effectively in neurointerventional procedures, and for the purposes of this exploratory study, the fiber optics were embedded inside the working lumen of existing neurointerventional microcatheters. This allowed these microangioscopes to be pushable, trackable, and flexible, offering neurointerventionalists a device that feels familiar. Prototypes were created with the SL-10 microcatheter (Stryker Neurovascular, Fremont, California, USA), Scepter C balloon microcatheter (Terumo MicroVention, Terumo, California, USA), and the Swift Ninja microcatheter (Merit Medical, Houston, Texas, USA).
The SL-10 microcatheter was selected because it is commonly used and has an adequate 0.0165" working lumen to house the fiber optics and a low-profile distal tip of 1.7 F (figure 2A). The goal was to provide imaging inside a familiar catheter for cerebrovascular interventions. Note that the fiber optics were completely embedded inside all of the microcatheters and could not move longitudinally or rotate relative to the microcatheter lumen.
The Scepter C balloon microcatheter was chosen to explore the possibility of stabilizing the scope within the center of an arterial lumen by inflating the compliant balloon. Its dual-lumen design with a second, separate inflation lumen allows the fiber optics to be embedded in the working lumen. In particular, the Scepter was selected for its working lumen of 0.0165" inner diameter that comfortably housed the fiber optics while maintaining a favorably small distal tip of 2.1 F (Figure 2B).
Lastly, prototypes were built using the Swift Ninja because it has a deflectable distal tip that may be controlled proximally through a dial. Although designed for peripheral (non-cerebrovascular) use, the Swift Ninja has a reasonably small distal tip of 2.4 F, which made it an acceptable candidate. The primary objective of the deflectable tip was to allow the clinician to control the viewing direction inside the blood vessel (figure 3A–C). The deflectable tip compensated for the problem created by embedding the fiber optics in the microcatheters' guidewire lumen: mainly that they were no longer over-the-wire devices and could not be steered or directed effectively in the vasculature without the aid of a close guide or intermediate catheter.
To clear the surrounding blood from the tip of the microangioscope, continuous irrigation was used, initially through a balloon guide catheter (BGC) under inflation in close proximity to the lens. Later, in anticipation of intracranial application, an intermediate catheter, the Sofia Plus (MicroVention, Terumo, California, USA), was fitted with an occlusion balloon at its distal tip to allow for even further distal access using this technique (figure 4).
In vivo imaging
All animal procedures were carried out with institutional approval. The pigs (n=3, female, 35–45 kg) were sedated (Telazol 4.4 mg/kg and xylazine 2.2 mg/kg IM), induced, and intubated under general anesthesia. Mechanical ventilation was given with oxygen mixed with isofluorane (1–3%). Routine physiological monitoring was performed. Right femoral artery access was obtained percutaneously under ultrasound guidance, and a 9 F sheath was placed and connected to a continuous heparinized saline flush. Under standard fluoroscopic guidance, either a BGC or a balloon-equipped distal access catheter (bDAC) was navigated to the common carotid artery in the case of the BGC and distal external carotid artery in the case of the bDAC. We specifically chose the distal external carotid artery to mimic the vessel size of human intracranial circulation. At this point, the microangioscope mounted on the microcatheter was advanced through the guiding catheter under balloon insufflation and continuous saline irrigation. This created a clear field-of-view of about 3 cm distal to the balloon occlusion during exploration of the proximal ICA and external carotid branches. On completion of the study the animals were euthanized under anesthesia.
Red and white thrombi were prepared using autologous blood. Red thrombi were prepared by mixing autologous blood with bovine thrombin. The mixture was injected into a silicone tube and then incubated at room temperature for 60 min. They were then cut into size before introduction through the BGC or bDAC. White thrombi were obtained after centrifuge separation of autologous whole blood. After 20 min, the fibrin-rich white thrombi were then separated and introduced in a similar fashion through the BGC or bDAC.25
Results
Illustrative examples
Online supplementary video 1: Microangioscope embedded in the working lumen of the Scepter C balloon navigated 3 cm beyond the distal end of a Flowgate BGC (Stryker Neurovascular, Fremont, California, USA). This is an example of navigation through a bloodless field with balloon inflation and continuous irrigation. Arterial branching points can clearly be seen.
Online supplementary video 2 and 3: Swift Ninja (embedded with the microangioscope) deflectable tip manipulation. (A) This illustrates the range of viewing angles that can be obtained by deflecting and rotating the tip of the microcatheter. The ICA origin and external carotid artery (ECA) along with its branches can clearly be seen. A red autologous thrombus was also visualized. (B) This example demonstrates that we can steer the catheter and guide it towards the ICA origin by deflecting and rotating the tip of the catheter.
Online supplementary video 3: Visualization of a white autologous thrombus and distinction of the white thrombus from the surrounding white vessel wall.
Supplemental material
Supplemental material
Supplemental material
Discussion
Angiography (DSA) is the gold standard for visualization of the cerebrovasculature. Radiation dose, contrast usage, and poor visualization over bony anatomy are its main disadvantages. While alternative methods of intravascular visualization exist, such as IVUS and OCT, they have the drawbacks of indirect visualization, metal artifact, and size restrictions for the intracranial vasculature.
Historically, angioscopy has been limited owing to the inability to visualize through blood, the size restrictions of previous technologies, and poor image quality. In this study, we have developed for the first time a high-resolution microangioscope designed to travel into the distal intracranial vasculature (eg, M2 segment of the middle cerebral artery), and an accompanying bDAC that would travel into the intracranial vasculature and assist visualization. With continuous irrigation and balloon inflation, we could obtain clear direct visualization of the common carotid artery, ICA, ECA, and distal ECA branches, with a 3 cm working space distal to the BGC or bDAC using a porcine model. To translate this to clinical practice, it is possible to place the DAC in the proximal middle cerebral artery for visualization of the M2, or even M3, branches with this 3 cm clear field of view beyond the guiding catheter.
Our experiments also showed that the resolution of the optics is excellent. Although the absolute pixel density is greater in other technologies, such as SFE, we could still clearly visualize arterial branching points and differentiate various types of artificially introduced thrombi. This might have important therapeutic implications as certain types of thrombus (eg, white thrombus) are known to be more recalcitrant.26 With direct visualization of the distal cerebrovasculature using this novel microangioscope, new models could emerge for diagnosis of cerebrovascular disease, treatments, postinterventional monitoring, and management of complications.
As an adjunct to standard DSA, microangioscopy theoretically may reduce radiation and contrast exposure during treatments. Furthermore, it could potentially reduce thromboembolic risk, since many thromboembolic events can, at least in part, be attributed to technically suboptimal device deployment secondary to poor visualization.8 9 It could be used in real time to deploy devices under direct visualization and help to identify intraprocedural complications, such as device malposition, malapposition, migration, and thrombus formation. Lastly, microangioscopy offers a new way to visualize treatment follow-up—for example, direct assessment of endothelialization over the device and direct visualization of intimal hyperplasia.
Given the potential of neuroendovascular endoscopy with this microangioscope, more in vivo experimentation is needed to explore its applications. A survey of the intravascular space should first be undertaken to characterize and report the anatomy and pathology using the technique described above. Another avenue of investigation is to compare this modality with DSA and other methods of vessel wall imaging, such as IVUS and OCT. Furthermore, in vivo experiments should be designed to study how neurointerventional devices (stents, coils, liquid embolic agents, etc) can be visualized during deployment and at follow-up.
Conclusion
Neuroendovascular techniques rely on fluoroscopy-based angiography. Alternative technologies such as optical coherence tomography and intravascular ultrasound offer indirect views of the endoluminal space but are too large for intracranial use. A new microangioscope, based on miniaturized fiber optic technology, offers a potentially revolutionary way of providing the neurointerventionalist with a new set of eyes in the cerebrovascular space.
Supplemental material
References
Footnotes
Contributors Conception and design: VMS, PK. Acquisition of data: TL, VMS, PC, MP, RG, SRC, JJ, DEC, PK. Analysis and interpretation of data: TL, VMS, PC, MP, RG, SRC, JJ, DEC, PK. Drafting the article: TL, VMS, PC, MP, DEC, PK. Critically revising the article: all authors. Reviewed the submitted version of the manuscript: all authors. Study supervision: PK. Guarantor: TL.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests PC and MP are employees of VenaMed, the company that developed the technology and main device discussed in this work. PK owns common stock in VenaMed.
Ethics approval Institutional Animal Care and Use Committee at Baylor College of Medicine.
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent for publication Not required.