A previously described modular neurovascular simulation model (HANNES) [15] with a commercially acquired silicone iliac-thoracoabdominal vasculature (United Biologics, Santa Ana, CA) and custom-made cervical and intracranial vessels were used. Patient-specific hollow vessel models based on anonymized patient data (ethic approval was waived by the local ethics committee) were produced by additive manufacturing using a commercially available 3D printer (Form 2, Formlabs, Somerville, MA, USA). The simulation system is integrated on a monoplane angiography system (AlluraClarity FD 20, Philips Healthcare, Best, The Netherlands). Three different anatomical variants (straight, elongated, and looped) of the cervical internal carotid artery (Fig. 1) were fabricated using a flexible resin (Flexible 80A, Formlabs, Somerville, MA) and attached to a skull base prototype. A physiological environment was simulated by an integrated fluid pump, equipped with a pulsatile valve and a heating system. The standard system configuration produced a flow rate of 0.4 L/min, a pulse rate of 70 bpm and a system temperature of 37 °C. Blood was substituted with a solution of water and small amount of commercially available soap for friction reduction. Synthetic clots [16] were placed in the M1-segment of the middle cerebral artery. Their position was confirmed by a single angiographic run using iodinated contrast medium. A custom-made distal emboli detection system was attached to the outflow channels.

Custom-made patient-specific, 3D-fabricated anatomical variants of the ICA: straight (a), elongated (b), and looped (c)

The Walrus™ 087 BGC was compared with a commercially available large bore new generation FlowGate2™ (FG2) balloon guide catheter (Stryker, Kalamazoo, Michigan, USA). The FG2 used has an 8F/0.106in outer diameter (OD), an inner diameter of 0.084in (ID) and a length of 95 cm. The ID of the Walrus BGC is 0.087in, the OD 8F+/0.110in and equal length of 95 cm for comparison purposes. The diameters of the inflated balloons were 10 mm for the FG2 and 11.1 mm for the Walrus. Due to their compatibilities, different distal access catheters were used. With the FG2, a 5F Sofia™ (Microvention, Aliso Viejo, CA, USA) with an OD of 0.068in, ID 0.055in, and a working length of 125 cm was used. The Walrus™ could accommodate a 6F SofiaPlus™ with a proximal OD of 0.0825in, ID 0.070in, and length of 131 cm (Fig. 2). For aiding intracranial access, a Headway™ 0.021in ID microcatheter (Microvention, CA, USA) and a Traxcess™ 0.014in microwire (Microvention, CA, USA) were available when necessary.

The figure illustrates a 0.087in ID, 95 cm long Walrus™ BGC with inflated balloon placed in the petrous segment of the ICA with a large bore AC (6F SofiaPlus™) placed in the MCA for contact aspiration. Courtesy of Q’apel Medical Inc.

The balloon guide catheters were advanced through a short 9F femoral sheath in the right femoral artery and advanced to the common carotid artery over a 5F SIM2 125 cm catheter (Cordis, Miami Lakes, FL, USA) for navigability. Seven interventional neuroradiologists with an average experience of 9.9 years in mechanical thrombectomy (BE 15, CB 13, FW 12, TI 10, MB 10, FF 7, HG 2) performed one recanalization attempt per anatomical variant by direct aspiration with the two different catheter combinations: Walrus/Sofia6F and FG2/Sofia5F. Operators were instructed to place the BGCs as distal as comfortably possible. The segmental level of the ICA [17] reached by the BGC, access to occlusion site, time to delivery, induced anatomical change and catheter kick-back were registered.

The participants responded to a questionnaire composed of 8 questions and an ordinal classification scale with a 5-point system (1 to 5). Responses were collected with the survey online platform Forms.App (Tallin, Estonia). Further data collection was performed on Microsoft Excel for Mac (Redmond, WA, USA). All statistical analyses were performed with SPSS (IBM, Chicago, IL, USA). Continuous variables were analyzed using Student’s t-test for two samples; comparison across operators and ICA models was performed with ANOVA. Nonparametric variables were analyzed using chi-square independence test, the Mann–Whitney U-test for two independent variables, and Kruskal–Wallis test for more than two independent groups.