All angiographic procedures were performed at the Neurosurgical Education and Innovation Laboratories at Baylor College of Medicine (BCM) with a Siemens Artis Zeego monoplane system. The procedures described in this manuscript were reviewed and approved by the Institutional and Animal Care and Use Committee at BCM prior to initiation of experiments and were completed in accordance with federal policies and guidelines (approval number: IACUC protocol #6339). Three-dimensional (3D) angiographic images were acquired using a 10-second rotational DR acquisition program (syngo DynaCT, Siemens AG). This fast acquisition protocol consists of 250 projections acquired at an angular increment of 0.8 degrees/frame in 10 seconds with a 200-degree angular coverage around the subject [15]. For all DynaCT scans, 50% iodinated contrast was injected at a rate of 3mL/s using an automatic power injector, with an imaging acquisition delay of 2 seconds. Breathing was briefly suspended for all scans. All 3D reconstructions were performed on the Siemens syngo Workplace. Two-dimensional (2D) angiographic images were acquired using manual injections of 50% contrast and a frame rate of 7.5 frames per second. Measurements of the arterial conduits were performed using Siemens Artis Zeego workstation with standard calibration methods. Arterial conduit measurements were performed using the calibrated Siemens workplace. Flow characteristics of the cerebral circulation were determined by reviewing microangiogram runs from various points in the cervical and cerebral vasculature.

Male intact purpose-bred hounds (26 to 35 kg) with an average age of 9 months and 15 days, received at least a 5-day acclimation period prior to any procedure. All animals in this study (n=10) were deemed healthy on intake by the BCM veterinary staff. All primary operators are fellowship-trained physicians with over 10 years of experience with neuroendovascular procedures.

In preparation for the procedure, morphine (0.2 to 0.5 mg/kg) (Hospira, Lake Forest, IL, USA) in combination with midazolam (0.1 to 0.3 mg/kg) (Akorn Pharmaceuticals, Lake Forest, IL, USA) and atropine (0.04 mg/kg) (Vedco Inc., St. Joseph, MO, USA) were administered subcutaneously for initial sedation and prevention of bradycardia. Animals that underwent an excitatory or dysphoric phase also received acepromazine (0.02 to 0.04 mg/kg) (Henry Schein Animal Health, Dublin, OH, USA) intravenously. Propofol (2 to 4 mg/kg) (Propoflo; Zoetis, Parsipanny-Troy Hills, NJ, USA) was administered for anesthesia induction. All canines were intubated and anesthesia was maintained using isofluorane (1.5% to 3.5%) (Forane; Baxter, Deerfield, IL, USA) delivered in 100% oxygen. At the end of the study, all animals were euthanized with a commercial phenytoin/pentobartital veterinary euthanasia solution (1 mg/4.5 kg IV) (Beauthanasia-D; Patterson Veterinary, Greeley, CO, USA) followed by necropsy.

Using standard palpation and anatomic landmarks, a micropuncture was performed to access the right femoral artery and a 6-Fr sheath was introduced using a modified Seldinger technique. Heparin (1,000 U in 1 L saline) was continuously infused through all guide catheters and microcatheters during the procedures. For angiographic imaging, contrast (Omnipaque 300, Iohexol; GE Healthcare, Princeton, NJ, USA) was infused and a single plane digital subtraction angiography (DSA) (see description under Imaging Protocol) was performed. Under fluoroscopic and roadmap guidance, a 6-Fr angled guide catheter, Envoy (Cordis) or Chaperon (Microvention Inc., Aliso Viejo, CA, USA), was advanced over an 0.035-in Glidewire (Terumo IS, Somerset, NJ, USA) into the aortic arch and then into the right and left common carotid arteries (CCA) and imaging performed (Fig. 1). A 6-Fr pig tail catheter (Cordis) was used to perform angiography of the aortic arch, great vessels, distal aorta, iliac arteries and surrounding vessel. A 6-Fr guide catheter system was decided upon based on our experiences catheterizing this model in our preliminary studies.

Once in the cervical vessels, a coaxial system comprised of a microwire (Synchro-2 [Stryker, Fremont, CA, USA],Synchro-10 [Stryker] or mirage microwire [Medtronic, Minneapolis, MN, USA]) and microcatheter (Echelon-14 [Medtronic], Marathon [Medtronic], and Magic [Bard, Murray Hill, NJ, USA]) was then established. The coaxial system was used to catheterize the right and left internal carotid arteries (ICA) (Fig. 2), the basilar artery (Fig. 3), vertebral arteries (Fig. 4), and the posterior cerebral arteries (PCA) (Figs. 2, 3A, and 4). 2 and 3D angiography were performed. For cases of severe arterial vasospasm, additional Verapamil (5 mg; Henry Schein Animal Health) was given intra-arterially just proximal to the site of vasospasm. Microangiograms were performed in each vessel catheterized and studied for downstream flow patterns.

For closure of the recovery animals, a 6-F vascular closure device (Angio-Seal; Terumo IS) was used to seal the right femoral artery and the skin was sutured closed. Manual pressure was then applied to the puncture site for 10 to 15 minutes and followed by application of a local 2% lidocaine block. Minor complications for recovery animals included non-progressive, minor hematoma formation and slight stridor on recovery due to an elongated soft palate, which were treated with cold compress and supplemental oxygen, respectively. No other significant abnormalities were noted.

Selective and superselective angiography was successful in all subjects with no major IA complications. Precise measurements of key arteries were made from the 2D calibrated angiograms and are listed in Table 1. The mean animal weight was 30.3 kg.

Similar to humans, the canine common femoral artery was found to be very superficial and access was easily achieved by palpation and direct puncture in all subjects (n=10). There was no need for a femoral cut-down or ultrasound guidance, which are often required for other large animal models such as rabbits and swine [15-17].

Catheter-related vasospasm was seen in all 10 subjects while catheterizing the CCA and vertebral arteries. Vasospasm ranged in severity from mild to severe and was alleviated with a continuous IA infusion of Verapamil (10 mg in 1 L saline) through the guide catheter. Extended manipulation and catheter positioning in smaller distal vessels resulted in increased spasm. To ensure the accuracy of vessel measurements, spasms were alleviated prior to measuring vessels.

We were unable to advance the microcatheter into the intracranial circulation via the trans-cervical ICA route in any subject (n=7) due to the presence of the severe tortuosity of the carotid artery (Figs. 1, 2, 3B) and reactive arterial vasospasm. This tight turn in the ICA as it passes through the skull base has been noted in other studies as well [18]. Injections that were performed via a microcatheter positioned in the distal ICA still showed poor antegrade intracranial middle cerebral artery (MCA) and PCA opacification due to competitive flow from the posterior circulation via Circle of Willis interconnections. In all (n=3) subjects tested, injection via the vertebral or basilar arteries demonstrated strong opacification of the MCA and PCA, suggesting that in contrast to humans, the vertebrobasilar system in canines is the dominant source of cerebral large vessel flow (Fig. 3). In 3/3 subjects, we were able to pass the microcatheter via the vertebrobasilar system into the PCA. In the first subject in which posterior circulation was accessed using the larger diameter Echelon microcatheter, there was significant straightening of the vertebrobasilar arteries and we felt that it was unsafe to advance the microcatheter beyond the proximal PCA (Fig. 4). In the 2 subjects where flow-directed catheters were used, the catheter was successfully advanced through the vertebrobasilar system and into the posterior communicating artery (PCoA) (Figs. 5A-C). While the PCA and PCoA’s were accessible with flow directed catheters and the posterior circulation branches readily catheterized (Figs. 5D-F), intracranial tortuosity made advancement across the posterior circulation into the anterior circulation via the PCoA challenging. Since we were aiming access the Circle of Willis but avoid subarachnoid hemorrhage (SAH), we chose not to risk cerebral vessel injury by making the vertical turn into the ICA.

Post-mortem analyses demonstrated SAH in 2 of the 10 canines. In both cases, the perivascular hemorrhage surrounded the basilar artery and extended cranially along the PCA and PCoA. The SAH were experienced in 2/3 animals where the vertebrobasilar system was accessed with Echelon 10 catheters, and the bleed was most likely caused by intra-procedural vascular displacement during microwire and microcatheter passage through the curvature or the basilar artery to the intracranial posterior circulation (Fig. 4). In the animal where trans-vertebrobasilar cerebral circulation access was solely attempted with flow directed microcatheter (Magic), there was no SAH. In the canines with internal carotid access attempts only, there was no SAH. Cerebral blood flow patterns were documented based on microangiogram performed in each vessel catheterized (Table 2).

The imaging demonstrated that the posterior circulation is dominant and fills the PCAs, MCAs, and anterior cerebral arteries (ACAs) bilaterally with even a modest injection rate. Injection in the cervical ICA filled ipsilateral MCA and ACA. When the ipsilateral PCoA was injected, the ipsilateral PCA, MCA and the ACA filled.

Angiographic arterial conduits that are relevant for neuroendovascular access and treatment planning were measured in all animals with suitable injections. The femoral artery was found to have a mean diameter of 4.64 mm, which was large enough to accommodate 6-F sheaths that are used in humans and an Angio-Seal closure device without difficulty or complication. The iliac artery and mid-thoracic aorta measured 6.65 mm and 9.38 mm, respectively (Table 1). The external carotid artery was found to have a mean diameter 3.65 mm while the ICA diameter was measured to be 1.60 mm. Notable intracranial vessel diameters include the basilar artery, which had a comparable diameter to the ICA. The Circle of Willis was noted to be robust with 1.15 mm diameter PCoA’s leading to robust opacification of the MCAs (1.49 mm) via basilar artery injection.

This study has some important study limitations. Since the purposes of the parallel study did not require visualization of the femoral and iliac arteries, there is a low sample size (n=2) for measurements of these vessels. It is the nature of the primary study that led to a smaller sample size, not the lack of accessibility of the vessels. Similarly, with regards to the lower sample sizes (n=2) for measurements of both the aorta and vertebral arteries, as many measurements of these vessels were taken in the context of the primary study.