Our in vitro vascular model revealed the behavior of NLI231 in combination with microcoils and with or without flow control. There were no migrations in scenarios with both microcoil and flow control. This suggests that NLI231 in combination with sparse coil placement and flow control may be feasible for embolization of medium-sized arteries. NLI231 and a mesh of sparse microcoil are both fragile alone. NLI231 is soft immediately after injection and migrates easily under fast flow conditions because it becomes harder slowly in the process of polymerization [4, 16]. Microcoil of 5 or 10% density is not enough to embolize the artery, considering the recommended density of 25% in the embolization of aneurysms [17]. However, similar to the relationship between cement concrete and reinforcement steel, the combination of NLI231 and microcoil with flow control led to successful embolization by trapping the NLI231 complex within a mesh of microcoil, acting as an anchor at the embolization site. Additionally, flow control prevented the migration of the NLI231 complex by inducing temporary stasis during the polymerization process, particularly during propagation from the surface toward the center of the NLI231 complex, as demonstrated by Li et al. [16].

Our results also suggest that microcoils with densities of both 5% and 10% are suitable as a scaffold for the NLI231 complex when flow control is applied and a distal anchor is present. Both densities are sparse, allowing the catheter tip to be easily visualized under fluoroscopy. Moreover, using these sparse microcoils can potentially reduce the cost and shorten the procedure time for extensive embolization.

NLI231 with microcoils without flow control can embolize the vessel, but partial migration occurred, and the distal distance of the NLI231 complex from the embolization site was longer. Under fast flow conditions without flow control, NLI231 migrates easily. While rapid injection of NLI231 into the embolization site may allow most of NLI231 to remain at the embolization site, it is impossible to prevent a partial migration. Therefore, the combined use of NLI231 and microcoils without flow control should be limited to cases where a certain degree of distal embolization is acceptable or where sufficient modification of distal blood flow is possible.

In scenarios that achieved embolization (distal flow = 0 ml/min), the systolic pressure on the proximal side of the embolization site was approximately 300 mmHg. This appears to be higher than the pressure observed in vivo, likely due to the presence of additional collaterals in actual vessels. Consequently, the proximal pressure would likely be even lower in vivo compared to our in vitro experimental setup.

There was no adhesion of NLI231 to the cannula after injection in any scenario, consistent with findings from previous studies [3,4,5,6,7,8]. Therefore, balloon catheters can be used for flow control without concern for adhesion. Non-adhesive liquid embolic materials such as LAVA ® [18] and Obsidio ® [19] may be used instead of NLI231, but they are expensive and not widely available.

The glue-in-plug technique has been reported by Ikoma et al. [5]. The glue-in-plug technique is superior to the microcoil–NLI231 combination in that it does not require an anchor and allows embolization over a short distance. However, the glue-in-plug technique cannot be used in cases where the vascular plug cannot be delivered to the target site due to strong tortuosity, or when the diameter of the vessel varies widely with extensive vascular injury. In these cases, the combination of NLI231 and microcoils would be a better option. While the plug has a fine and uniform mesh on its envelope, there are vacant areas where the NLI231 does not touch the mesh in the core. On the other hand, microcoils have an uneven overall gap; unless the center is very sparse, it may tangle with NLI231 from the beginning of injection, and it is unclear as to which is effective for the embolization with NLI231. It appears that this weak point of the plug could be compensated for by injecting the NLI231 after the microcoil is sparsely placed in the core of the plug, but further study is needed.

This study had several limitations. Although we simulated conditions similar to those in vivo, such as temperature, serum composition, and the hydrophilic surface of the vessels, which are important for the polymerization of n-butyl cyanoacrylate as demonstrated by Wang et al. [20], factors such as blood coagulation, arterial elasticity, blood viscosity, and variety of blood velocity were not taken into account. In particular, we used donor horse serum instead of human blood. This choice may result in slower polymerization and increased migration compared to a clinical setting. The absence of various factors, including the effects of clot formation that aid in coagulation and polymerization, contributes to these differences [21]. Therefore, further in vivo studies using blood and actual vessels are needed. Additionally, the number of trials was small, and variations in coil specifications, such as stock wire/primary/secondary diameter, shape, and material, could not be adequately assessed due to budget limitations. In this study, needles were used to anchor the embolized coils, assuming that the coil did not move with the flow. More coils will likely be needed in clinical practice to prevent coil migration than in this experiment. Finally, the long-term outcomes of the embolization procedures remain unknown.