Sub-angiographic peripheral emboli in high resolution DWI after endovascular recanalization

Multiple randomized trials have demonstrated the benefit of mechanical thrombectomy (MT) for acute ischemic stroke due to large vessel occlusion (LVO) [1, 2]. Clinical outcome after MT can be improved by an increasing degree of reperfusion evaluated by the expanded thrombolysis in cerebral infarction (eTICI) scale [3, 4]. Accordingly, complete reperfusion (TICI 3) is leading to a markedly better outcome with the highest rate of excellent functional outcome (mRS ≤ 1) [5, 6].

Peripheral emboli as a complication after MT, supposedly caused by thrombus fragmentation, might worsen clinical outcomes [7]. In vitro data suggests that the majority of peripheral emboli released during MT are smaller fragments [8, 9]. Prevalence of distal emboli and infarcts in new territories are underestimated since most studies rely on computed tomography for follow-up [9]. Magnetic resonance imaging (MRI) in comparison is more sensitive in the detection of peripheral emboli that might not cause vessel obstruction detectable by digital subtraction angiography (DSA) [10]. An increase of spatial resolution of DWI termed high-resolution DWI has been shown to improve conspicuity and allows to detect more ischemic lesions than the standard-resolution DWI [11, 12]. This may give information of MT success beyond the eTICI scale. We aimed to analyze the incidence of peripheral emboli after successful reperfusion. We hypothesized that a substantial number of peripheral emboli lead to tissue ischemia within the vascular territory of an LVO and that these emboli can be detected on MRI more often than on DSA using high resolution diffusion-weighted imaging (DWI) and/or susceptibility-weighted imaging (SWI). This might be of importance for procedure-related safety features to prevent peripheral emboli during MT.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Twenty-eight patients were included in the study. Prior to MT, alteplase was administered in 11 patients (39.2%). Details on baseline characteristics, procedural, and functional outcome of the patients can be found in Table 1.

Twenty-six out of 28 patients showed peripheral punctuate lesions on high-resolution DWI suspicious for peripheral emboli. With a total of 324 lesions high resolution DWI detected significantly more peripheral emboli compared to standard resolution DWI (total 324 vs. 151; median 10 vs. 5; p < 0.001). 33 SWI lesions suspicious for peripheral emboli were found, all in close proximity to DWI lesions (33/324 = 10.2%). 8.0% (26/324) of DWI lesions were directly distal to occluded peripheral vessels on post-interventional DSA (Fig. 1). The majority of peripheral ischemic lesions was small with < 0.5 ml volume on DWI (online supplement Figure I).

The number of peripheral emboli was inversely correlated with eTICI scale (Spearman’s rho 0.611; p < 0.001). Eight (8/9) patients rated as eTICI 3 showed peripheral emboli. In 11 patients emboli into new territories were found in DWI (31/324 = 9.6%). The volume of peripheral lesions on high resolution DWI represented 2.1% of the cumulated ischemic stroke volume. The number of DWI and SWI lesions for each eTICI grade is shown in Table 2.

There was a tendency towards a lower number of peripheral emboli when MT was performed under flow arrest using a balloon-guide catheter compared to MT with a median number of emboli of 5 (IQR 0.5–9.5) with and 11 (IQR 4–21.5) without a balloon-guide catheter but the difference was not statistically significant (p = 0.057) (online supplement Figure II). The number of emboli did not differ between patients that received IV alteplase and those who did not with a median of 11 (IQR 4–22) versus 17 (IQR 2.5–16) (p = 0.944).

Distribution of peripheral emboli in relation to the core DWI lesion for each eTICI grade is visualized in Fig. 2.

Our analysis confirmed the hypothesis that high-resolution DWI detects peripheral emboli within the vascular territory of LVO even after complete reperfusion (eTICI 3). Only a small fraction of DWI lesions occurred directly distal to DSA proven vessel occlusions. These sub-angiographic emboli occur much more frequently in the vascular territory of the initially occluded vessel than in other territories.

In vitro studies suggested that every endovascular recanalization attempt leads to peripheral emboli [14]. Previously, it was assumed that small peripheral emboli may dissolve spontaneously, thus, not leading to tissue infarction [15]. This is the first study to examine the degree of brain injury by peripheral emboli after MT using high resolution DWI. We observed that a certain amount of these emboli indeed cause peripheral tissue ischemia. Peripheral emboli were more frequent in eTICI 2b than 2c but occurred also in patients with eTICI 3.

Previous studies comparing pre- and post-interventional MRI have shown that over 90% of intracranial vessel occlusions are caused by a single thrombus [7]. In consequence, most incomplete reperfusions after MT (less than TICI 3) would result from thrombus fragmentation as a complication of the procedure. Reperfusion success was found to be the most important modifiable predictor for outcome in previous studies [5]. Consequently, persisting vessel occlusions after MT worsen clinical outcomes. However, the question why some patients with successful reperfusion and small core infarcts do not reach functional independency after 90 days remains incompletely answered. Hypertension, hyperglycemia, and fever are modifiable factors that are associated with worse outcomes after stroke [16]. Another factor might be peripheral emboli. In our study, total peripheral emboli volume represented only 2.2% of the cumulated ischemic core volume. Nevertheless, if located within strategic brain regions such as the motor cortex or cortico spinal tract, even small lesions might have a relevant effect on clinical outcome. Thus, it has been found that infarct location in addition to core infarct volumes is a significant factor determining the clinical outcome [17].

It was already shown that pre-interventional thrombus fragmentation worsens outcome in patients [7]. However, the relation of peri-interventional thrombus fragmentation to functional outcome has to be elucidated in further studies.

Our findings might have important implications for safety features during MT and design of future stroke studies. One focus of future studies will be the question if intravenous thrombolysis before mechanical thrombectomy is associated with an additional benefit [18]. If peri-procedural intravenous thrombolysis could reduce the burden of peripheral emboli as detected with DWI it might improve outcome. Currently, recommendations for procedural details are based on hypothetical considerations, in vitro testing and porcine in vivo models that are difficult to prove in the clinical setting [14, 19]. DWI infarct growth was proposed as a surrogate endpoint for future stroke therapy trials to reduce sample sizes [20]. Acquiring DWI in high resolution to detect peripheral emboli might further add to this metric.

It is possible that not all peripheral emboli may have been related to MT but could have occurred independently. But, as previously mentioned, over 90% of intracranial vessel occlusions are caused by a single thrombus [7]. Another argument for the association with MT is the clear inverse correlation of the number of emboli with the eTICI scale. Furthermore, a primary embolic source would have led to more emboli into new territories. Lastly, a peri-interventional occurrence of the emboli explains best the trend towards a lower number of emboli when using a balloon-guided catheter.

Another limitation is the limited sample size of the study. Future studies on larger samples using high-resolution DWI as follow-up will explore possible predictors for these emboli and their influence on clinical outcome.

Emboli after MT for acute ischemic stroke occur most frequently within the vascular territory of an LVO, even after complete reperfusion of eTICI 3. High resolution DWI to detect peripheral emboli after MT could be a promising surrogate marker for reperfusion success beyond the TICI scale.