Stent Retriever Thrombectomy of Small Caliber Intracranial Vessels Using pREset LITE: Safety and Efficacy

Triggered by the positive results of the recently published prospective randomized trials, mechanical thrombectomy will become an integral part of stroke therapy and will contribute to improved clinical outcome of patients. However, the trials providing evidence for the effectiveness of endovascular stroke therapy mainly focused on proximal large artery occlusions of either the terminal carotid artery or the M1 segment of the MCA. Despite the fact that some trials allowed M2 occlusions to be included, only a few patients with an occlusion beyond M1 were randomized. This was possibly driven by the assumption that peripheral artery occlusions cause less severe clinical syndromes and respond well to intravenous thrombolysis. Earlier studies proved a relative safety of M2 compared with M1 thrombectomy even with the MERCI device, but still some safety concerns remain [7]. Several curves have to be overcome to access a peripheral target and will be stretched during device retraction. Thus, thrombectomy in small caliber peripheral vessels may harbor an increased procedural risk. The failure to prove a clinical benefit of early M2 reperfusion in a pooled analysis of the PROACT II, IMS, and IMSII study may have supported the reluctance to include these patients into a prospective trial [8].

On the other hand distal vessel occlusions may lead to serious and permanent morbidity. Despite the fact that patients with an M2 occlusion have a better prognosis compared with patients with a more proximal occlusion, unfavorable outcome can be expected in up to 45 % if left untreated [9]. Intravenous thrombolysis improves outcome, but still up to 30 % of patients will die or be dependent at 90 days [10, 11]. Similar considerations hold true for posterior cerebral artery occlusions, which can be associated with severe somatosensory and behavioral as well as cognitive deficits [12]. Even after intravenous thrombolysis the rate of unfavorable outcome after posterior cerebral artery occlusion may be as high as 40 % [13]. These data support the search for more effective treatment strategies for distal vessel occlusions.

Occlusion of distal vessels may also be encountered while treating more proximal targets. Emboli to multiple territories or thrombus extension to small branches can be a primary finding. In addition, thrombus fragmentation and embolization to distal branches or new territories are known side effects of mechanical recanalization. Emboli into new territories can be expected in 0.7–11 % of recanalization procedures depending on the technique applied [14‐16]. Recanalization attempts seem reasonable if eloquent brain tissue is jeopardized and recanalization techniques with a good safety profile are available.

Recently some multi- and mono-centric studies emerged addressing thrombectomy of M2 branches as well as anterior cerebral artery occlusions in one series [17‐20]. Treatment in these studies was performed with regular stent retrievers and demonstrated good feasibility and safety. Successful recanalization of M2 was associated with improved clinical outcome and smaller infarct size in one investigation [17]. However, in some distal branch occlusions, treatment with already available devices may be considered to be potentially harmful due to the small diameter of the affected vessel, location of the occlusion in a curved vessel segment, or retraction pathway. Retrievers with a reduced radial force and diameter fitting through small caliber microcatheters may increase the spectrum of potentially treatable vessel occlusions.

pREset LITE 4–20 and 3–20 were designed and approved to meet these demands. In this manuscript we report our clinical experience with the two devices regarding recanalization results and safety data. pREset LITE was selected in situations in which the operator felt treatment with regular devices to be associated with an increased risk. In our analysis we included target branches with a diameter equal to or less than 2 mm to focus on vessels, which cannot be treated with the well-described and established regular devices according to their specification.

Accessibility of peripheral branches was excellent, with only 2.2 % failed attempts and successful recanalization was achieved in 70.0 % of targets. For the treatment of proximal vessels with the regular pREset retriever a recanalization rate as high as 85 % was previously reported [21]. The small gap of successful recanalization between proximal and distal vessels could possibly be explained by the reluctance to act aggressively in small branches. This is underlined by the finding that more than two thrombectomy passes were rarely performed in a distal vessel. Recanalization rates reported in the recently published randomized trials addressing thrombectomy of large vessel occlusion ranged between 59 and 88 %, which compares well with the 70 % achieved using pREset LITE in small vessels [1‐5].

In terms of adverse events, significant vasospasm after retraction of the device was observed in 5.6 % but always resolved with the infusion of vasodilators and did not affect outcome in any case. Cerebral vasospasm occurs frequently after mechanical thrombectomy. A detailed analysis of procedural complications in the “Solitaire With the Intention For Thrombectomy” trial revealed some vasospasm in 22.5 % after treatment with Solitaire but none was symptomatic [12]. Haussen et al. reported their experience with Trevo XP 3 × 20 for the treatment of distal vessels and described relevant vasospasm in five of eight patients, and vasospasm always responded to vasodilators [22]. Thus, vasospasm after thrombectomy seems to be a benign phenomenon even when treating smaller distal branches.

New emboli after thrombectomy with pREset LITE occurred only in two occasions. Theoretically, the risk of losing thrombus material using small, low radial force retrievers loaded with small thrombi could be increased during the passage of larger proximal vessel segments. Since all patients in this series were treated with intermediate catheters and distal aspiration, we did not observe emboli into new but only the same territory. Emboli into the same vascular territory have been described to occur in approximately 4.4 % of cases [12]. Using this as a benchmark, a frequency of 2.2 % after treatment with pREset LITE is acceptable.

In two targets (2.2 %) dissection was suspected because the vessel remained occluded after thrombectomy and reentry seemed more difficult compared with prior catheterizations. No further thrombectomy was performed to avoid harm and, instead, self-expanding stents were permanently implanted to restore flow. Target vessel dissection is rarely reported after stent-retriever thrombectomy with a maximum rate of 1.5 % [15]. A rate of 2.2 % after treatment of distal vessel branches is slightly higher, but both intracranial dissections occurred in patients with a cervical artery dissection and may reflect a higher vulnerability due to an underlying vessel wall disease. In addition, the diagnosis of dissection was only suspected based on the difficulty to reenter the target branch and was not proven by the typical aspect of an intimal flap. A different explanation of this phenomenon (e.g., change of the mechanical properties of the thrombus due to the previous recanalization attempt) cannot be excluded. Alternatively, dissections may be underreported in the literature and “hiding” in the rate of unsuccessful recanalization attempts.

With the treatment of distal branches the risk of vessel perforation and intraprocedural hemorrhage could be increased. Device-related intraprocedural contrast extravasation was only observed once; it was self-limiting and without impact on the final outcome. Analysis of posttreatment imaging revealed potentially device-related PHI or SAH in 2.2 and 13.3 %. Although none of the latter altered the clinical course, the frequency of SAH was in the upper range compared with literature references. SAH was reported to occur in 0.9–16.2 % in stent-retriever studies and was below 10 % in most publications [1‐5, 12, 15, 23]. Studies explicitly analyzing the frequency and impact of SAH or contrast extravasation after thrombectomy performed imaging immediately or a few hours after treatment, and SAH may disappear at later time points [22, 24]. Since our patients did not routinely receive immediate posttreatment imaging studies, the frequency of SAH may even be underestimated. The authors reporting the highest rate of SAH of 16.2 % found an association with rescue angioplasty, which cannot serve as an explanation for the SAH we found after thrombectomy of small branches. Therefore, the hemorrhage is most likely attributable to the distal location of the target vessel. Stretching and rupture of small perforating branches may serve as an explanation for these angiographically occult bleedings. Predictors of procedure-related SAH were analyzed in detail after treatment with the MERCI device, and the authors found distal MCA occlusion to be an independent predictor together with hypertension, rescue angioplasty, and intraprocedural perforation [25]. In the same study SAH was without impact on clinical outcome unless extensive or accompanied by severe parenchymal hematoma. The benign course of SAH due to angiographically occult bleeding was also confirmed in a retrospective case control study investigating this phenomenon after stent-retriever thrombectomy [26]. In summary, up to now there is no evidence of clinical impairment due to focal SAH after mechanical thrombectomy, but some uncertainty remains since larger studies are still lacking.

Statistical analysis revealed a significant relationship between recanalization and infarcts in the dependent territory. With successful recanalization the chance to develop none or only a partial infarct was higher compared with the risk of complete infarction. Haussen et al. presented their initial experience using “Baby Trevo” in distal intracranial vessels [16]. The number of patients treated in their series was limited (n = 8), but all targets were successfully recanalized. Only one patient developed a complete infarct, four had a partial infarct, and three no infarct. Thus, the results were qualitatively comparable with the findings in our study. Due to the fact that the majority of patients developed at least a partial infarct, the authors considered the effectiveness of thrombectomy in distal vessels to be unproven. Our study had a higher overall number of patients and a reasonable number of not successfully recanalized targets, which served as a comparator. Under these premises we were able to demonstrate that successful recanalization significantly reduces the risk to develop a complete infarction in the dependent territory, and partial infarction also reflects at least partial treatment success. In addition, we found a trend for less infarcts in patients recanalized for emboli into a new territory. This finding is most likely attributable to the shorter occlusion time compared with patients with primary occlusions or emboli into the already affected territory.

Our study is limited by its retrospective nature and the fact that efficacy was proven by infarct occurrence and size, which served as a surrogate for clinical relevance. Further studies to directly address the impact on clinical outcome are warranted and could potentially be conducted in cases with primary isolated small branch occlusions. With embolic occlusions of more proximal vessels extending into smaller branches or secondary emboli, the impact of recanalization of a single branch on clinical outcome will be difficult to determine.

In summary, recanalization of distal branches is double-edged with a good chance for successful recanalization and reduction of infarct size but an increased risk of focal SAH. Up to now there is no evidence for a relevant clinical impact of focal SAH after thrombectomy, but larger confirmative studies are lacking. With the remaining uncertainty, the most reasonable approach would be to confine thrombectomy of small vessels to branches supplying eloquent brain tissue, which increases the likelihood of a relevant clinical benefit.