In vitro testing of a funnel-tip catheter model to decrease clot migration and to evaluate clot effects in mechanical thrombectomy

Clot-retrieval failure and embolism in new territories (ENT) are complications reducing the ecacy of mechanical thrombectomy (MT). Reasons for these problems are clot abruption during stent retriever retraction into the cylindrical tip of a standard distal access catheter (DAC), and antegrade blood ow during retrieval. To overcome these complications, a funnel-shaped tip catheter model was developed and tested. Using an experimental in vitro setup, this study compared the ecacy of a funnel-shaped tip with a standard cylindrical tip in combination with different clot compositions. Mechanical thrombectomy was performed 80 times for each tip, using two stent retrievers (Trevo XP ProVue 3/20 mm and 4/20 mm) and four different clot types (mechanical vs. static preparation, 0-24h vs. 72h clot aging times). Signicantly higher rst-pass recanalization rates were observed for the funnel-shaped tip, which reached 70.0% vs. 30.0% for the standard tip (absolute difference, 32; relative difference 20.0%, P < 0.05), regardless of the clot type and stent retriever. Recanalization could be increased using brin-rich mechanical clots vs. clots prepared under static conditions, as well as 0-24h vs. 72h aged clots, respectively. The extended in vitro-experiments veried the usability of the recently patented funnel-shaped DAC, hereby implementing next-level in vivo-experiments.


Introduction
Mechanical thrombectomy (MT) is a recommended treatment for acute ischemic stroke (AIS) patients with large vessel occlusion (LVO), within 24 hours of symptom onset 1 . A pooled meta-analysis (HERMES collaboration) from the so-called "big ve" multicenter randomized controlled trials reported a successful recanalization rate of 71% (mTICI Scores 2b): ranging from a minimum of 58.7% (MR CLEAN) to a maximum of 88.0% (SWIFT PRIME), and improved functional outcome of 46% (mRS score after 90 days) for mechanical thrombectomy in LVO of the anterior circulation (M1, M2 and ICA). Despite high recanalization rates, embolism in new territories (ENT) rates of 1 to 8.6% have been reported for most randomized controlled trials 2−5 . Although the use of proximal balloon occlusion during thrombectomy has been suggested to signi cantly reduce this problem, clot migration into distal territories still remains unresolved. This might be caused by clot abruption at the wall of the cylindrical tip during stent retriever retraction into standard distal access catheters (DAC) 2 . Due to an overlap of thrombus material during retraction into the narrow opening tip, the thrombus is compressed and elongated, increasing the risk of fragmentation 6,7 .
A funnel-shaped tip catheter model was tested to decrease the risk of clot migration and to increase the rate of rst pass TICI 2b/3 results. We compared the funnel-shaped tip featuring a larger inner-opening diameter (d = 2.5mm) to a standard tip (d = 1.5mm) model, in terms of functionality and e cacy for mechanical thrombectomy in an in vitro setup. In our rst in vitro study, we observed higher recanalization rates of up to 90% using the modi ed funnel-shaped tip compared to 50 to 60% for the standard tip 2 . Consequently, the proof-of-concept for a newly designed funnel-shaped tip was deemed viable for continuing further in vitro experiments, this time including a greater case number (N = 160) and ≥ ≤ 2 employing different clot types. As other recent studies 6,8,9 have indicated, thrombus-vessel and thrombus-device interaction are in uenced by the viscosity and composition of clot types. Successful MT is associated with reduced friction and adhesion during retraction. Certain clot types may be more prone to fragmentation and may increase the risk of ENT 6,8  Welfare Act (Tierschutzgesetz).
Venous whole blood was extracted before sacri cing the pigs, in accordance with the above mentioned animal welfare regulations and procedures. Polyvinyl chloride (PVC) tubes (clear PVC tubing, inner diameter 8.0 mm, outer diameter 12.0 mm; Thermo Scienti c Fischer, Waltham, Massachusetts, USA) were half lled with blood, and the ends of the tubes were connected using a silicone cuff 2 .
Two methods were employed for generating thrombi: (I) mechanical preparation using a Chandler loop device (Chandler loop System, Neuffen, Germany), and (II) clot preparation under static conditions. The Chandler loop simulated physiological blood circulation under dynamic conditions in a standardized setting. For this purpose, a temperature-controlled water basin was preheated to 38.5°C. The PVC tubes containing porcine blood were placed in the loop cradle and rotated at 15 rpm. The "moving column of blood" 10,11 coagulated after 20 to 30 minutes. Simultaneously, the thrombi processed under static conditions were prepared by hanging the PVC tubes vertically from a rod for 20 to 30 minutes at room temperature.
After blood coagulation within the PVC tubes, the clots from both preparation methods were removed from the tubes and refrigerated in saline solution at 10°C. MT was performed either 0 to 24 hours or 72 hours after preparation of thrombi. Prior to each experiment, the thrombi were cut equally into 20 mm long segments.

Mechanical Thrombectomy Procedure
A thrombus segment was placed into the MCA model under aspiration for each MT experiment. Then, one of the catheter models (funnel-shaped tip; standard tip) was positioned at the proximal end of the thrombus. After inserting the stent retriever (Trevo XP Pro Vue 3/20 mm; Trevo XP Pro Vue 4/20 mm; Stryker, Kalamazoo, Michigan, USA) into the MCA model, it was released over the thrombus for 3 minutes.
The thrombus was retracted into the catheter model under aspiration, using a vacuum pressure syringe (60 ml, VacLok, Vacuum pressure syringes; Merit Medical System) set in advance to a negative pressure at a volume of 40 ml. MT was conducted using the primary combined approach (PCA), for both aspiration and stent retriever devices. Experiments were performed under visual control using uoroscopy and video monitoring.

Histopathological Analyses
Histopathological analyses of fresh (0h) and matured (72h) thrombi, prepared with and without the use of a Chandler loop system, were performed after xation of thrombi in 4% buffered formalin (Formaldehyde solution 4% buffered (PH 6.9); Merck KGaA, Darmstadt, Germany). The thrombi were cut at 3 mm intervals, sliced by a microtome in 2 µm thick sections onto microscopic slides, and stained with masson trichrome. For enhanced visualization, the microscopic slides were scanned with Nanozoomer 2.0 HT (Hamamatsu), magni ed 200-fold, and exported as TIFF les. To differentiate the exact brin distribution, pixel-by-pixel analysis was performed using Trainable Weka Segmentation Fiji plugin v3.2.27. Each of the different clot types was graduated according to three different categories: brin pattern, distribution and outer brin rim.
The category " brin pattern" was further speci ed as insular, reticular or no pattern. The distribution of brin was classi ed as peripheral, central, both peripheral and central (regular), or irregular distribution, depending on the localization. The outer rim of brin (de ned as circumferential layer of brin surrounding the thrombus) was divided into the subcategories total circumference of brine, focal or no rim.

Statistical Analyses
For the MT experimental series, statistical analyses were performed with GraphPad Prism 9.

Discussion
MT in combination with IVT is the standard treatment for acute large vessel occlusion 1 . Despite high recanalization rates and improved techniques, endovascular treatment still faces limitations, reporting ENT incidences of up to 10% for aspiration devices [12][13][14] . ENT can be caused by clot abruption and stripping along the wall of conventional DACs, due to elongation and compression of thrombus material through the narrow opening 6,7 . Aiming to solve this problem, a modi ed model featuring a funnel-shaped tip was tested in this study. The larger opening (d = 2.5 mm compared to approx. for 1.7 mm for actual 068-distal access catheters) of the funnel-shaped tip relates to the anatomical diameter of the M1-MCA segment, which is reported to be 2.7 ± 0.23 mm (mean ± 1 SD) 15,16 , hereby reducing unwanted effects like striping of the thrombus at the catheter tip using a 4.0 mm stent retriever. The proof-of-concept of this study was validated by achieving signi cantly higher recanalization rates for the wider, funnel-shaped tip, with 56/80 successful MTs (70.0%) at rst pass compared to 24/80 (30.0%) for the standard tip, regardless of the stent retriever and clot type.
The funnel-shaped tip aims to facilitate MT in three ways: 1) by inducing local ow arrest during retraction into the catheter tip model, turbulence and reverse ow may be reduced around the catheter model, mitigating the risk of clot fragmentation and migration; 2) by decreasing the pressure gradient during aspiration due to the larger opening, aspiration forces may be applied across the entire catheterthrombus contact area; 3) the funnel-shaped tip may decrease the resistance and shear forces, which the self-expanding stent retriever has to overcome during retraction. One problem affecting the standard tip is the disproportionally larger size of the self-expanding stent retriever in relation to the smaller opening of the standard tip during retraction. However, the larger opening of the funnel-shaped tip may guide the stent retriever into the catheter model, enabling a better retraction or interlock during aspiration 6 .
Evaluating the causes of MT failure is one advantage of an in vitro setting. An abruption of clot material at the wall of the tip was the main cause of failure for the standard tip. In contrast, half of the failures for the funnel-shaped tip occurred before reaching the wall of the tip when the stent retriever failed to grip the clot.
Recent literature indicates that other funnel-shaped catheters, such as the Lazarus Device (Lazarus Effect, Campbell, California, USA) 17 and the Advanced Thrombectomy System (ANCD; R&D, Anaconda Biomed, Barcelona, Sant Cugat del Vallès, Spain) 7 have been independently developed to our project. The Lazarus Device uses a different mechanism with a funnel-shaped nitinol mesh to enclose the stent retriever without inducing local ow arrest. The ANCD bares technical resemblance to our funnel-shaped tip, in addition to similarly initiating local ow arrest during procedure 17,18 . Both studies also reported promising results for funnel-shaped models, with recanalization rates of 94% and 80% for the ANCD and Lazarus Device, respectively 7,17,18 .
Evaluating the effect of clot compositions, Chandler loop thrombi showed a tendency towards improved recanalization compared to thrombi prepared under static conditions. Recent publications have investigated the effect of clot composition on MT. Fibrin determines the structural integrity of blood clots forming a three-dimensional scaffold. Mechanical forces as well as thrombolysis may cause brin bers to break when certain threshold levels are exceeded. Consequently, the structural integrity of the blood clot is impacted. Physical defects and cracks within the blood clots make the thrombus more susceptible to rupturing. Thus, clot composition and geometry play a pivotal role in the pathogenesis of ENT 20 .
According to Duffy et al. 19 , clots processed under static conditions, similar to one group of clots from this study, displayed an erythrocyte-dominant (RBC-dominant) and low brin content. Chandler loop clots, showed a brin-predominant pattern and were characterized by their rmness and elasticity 19 . In correlation to the histopathological ndings of this study, clots prepared under static conditions were distinguished by lower levels of brin, absence of a brin pattern, as well as irregular brin distribution. Due to the coagulation of blood under static conditions, erythrocytes were unevenly distributed and more densely concentrated. Controversially, some studies classify Chandler loop clots as erythrocyte-rich 21 . In this study, Chandler loop clots are classi ed as brin-rich because they contain lower amounts of erythrocytes and higher amounts of brin relative to clots developed under static conditions. Research literature states that white brin-rich thrombi are associated with reduced recanalization rates, in which the stent retriever is unable to grip the clot. The high brin content could account for an increased adhesion and resistance along the vessel wall, thereby impeding MT [6][7][8] . However, this study registered higher recanalization rates for the Chandler loop clots ( brin-rich) than for the "static" clots. An explanation for this is provided by Madjidyar et al. 21 , stating that MT success not only depends on the "right" clot type, but also on the "correct" technique [21][22][23][24] . Clot removal for brin-rich clots could be improved by applying distal aspiration, as performed in our study. Moreover, the Direct Aspiration First Pass Technique (ADAPT) was considered more effective using RBC-rich clots, whilst stent retrievers and balloon-guided catheters (BGC) were more successful with brin-rich clots 21 . Since our research project relied on PCA, using aspiration in combination with stent retrieval, different MT techniques should be included in future studies to evaluate the e cacy of our modi ed prototype. Numerous studies indicated that RBC-rich thrombi were related to higher recanalization rates, but showed a higher tendency towards fragmentation 6-8 . Accordingly, "static" thrombi from this study (RBC-rich) were found to be more susceptible to fragmentation, resulting in higher rates of clot abruption at the wall of the tip. This can be explained by a lack of structural integrity due to the absence of a brin pattern. Furthermore, the disparity of recanalization rates between the Chandler loop 0-24h and 72h thrombi was found to be greater than between the "static" 0-24h and 72h clots. Since brin may dissolve over time and become lytic, the brinrich Chandler loop thrombi may be affected more by the aging time.
Evaluating how the clot aging times may affect MT, 0-24h aged clots achieved higher recanalization than 72h clots 25 . Histopathologically, the difference between aging times could be explained by an outer brin rim (predominant in the 72h aged thrombi group), which may cause increased friction during retraction. Aged thrombus material may restructure and degrade, becoming more condensed due to uid loss.
Additionally, the higher density could increase friction during retraction and impede stent retrieval 8,25−27 .
The simpli ed experimental setup of this study can be considered a limiting factor. In addition, the MCAmodel did not take the tortuosity of the MCA-vasculature into account. Also, a simpli ed modi ed tip model was used. The actual funnel-shaped prototype, which was still being developed at the time of the experiments, was recently patented. It features a self-expanding exible structure optimized to induce local ow arrest. Because it is unclear how the funnel-shaped tip could affect the vessel endothelium, further in vitro and in vivo research is required. Testing homogenous clots under standardized laboratory conditions is another limitation of this study, since clinical clots have irregular heterogeneous compositions that are still di cult to determine 27 . Additional studies should consider examining the impact of brinolytic agents on MT procedure and clot structure. A performance bias in this study cannot be excluded, since the interventionalist could not be blinded during MT procedure.

Conclusion
The underlying study demonstrates signi cantly higher rst-pass recanalization rates for the modi ed funnel-shaped tip versus the standard tip model, regardless of the clot type and aging time. Clot abruption at the wall of the tip was the leading cause of failure for the standard tip. Furthermore, results indicate that MT success rates were affected by different clot types and compositions. Higher recanalization rates were achieved using mechanically-prepared brin-rich Chandler loop clots, instead of erythrocyte-rich prepared clots under static conditions. In addition, 0-24h aged thrombi achieved higher recanalization rates than 72h aged thrombi. This follow-up study underscores the feasibility of the modi ed tip for further developing the funnel-shaped DAC prototype.