Histopathological Profile of Thrombi Differs Depending on Pathogenesis of Embolic Cerebral Infarction and Evidence of Intracardiac Thrombus

Histopathological Profile of Thrombi Differs Depending on Pathogenesis of Embolic Cerebral Infarction and Evidence of Intracardiac Thrombus


Introduction
Determining the etiopathological source of stroke is a cornerstone for correct secondary prophylaxis. The majority of cerebral ischemic strokes are thromboembolic [1]. Although, their etiologic un pathogenetic mechanisms can differ significantly. The main embolic cerebral ischemic stroke causes are arterioarterial embolism from atherosclerotic plaque and cardioembolism [2]. The proportion of thromboembolic infarctions in Latvia is reaching 75% (Pauls Stradiņš Clinical University Hospital local stroke registry). Nevertheless, in up to 40% the precise source of embolism cannot be determined even after extensive evaluation [3]. For clinicians cryptogenic strokes are a major problem and are an important research focus [4,5]. Several studies have shown high rates of recurrent stroke episodes of ≤30% in the first year after suffered cryptogenic stroke [6].
With the development of endovascular therapy, it became possible to analyze the thrombus material histologically, thus providing additional diagnostic opportunities [7]. Some authors have hypothesized that it is histologically possible to differentiate between cardioembolus and arterioarterial embolus using the classical hematoxylin-eosin staining method [2]. The pattern of each thrombus component distribution often differs between patients with cardiac thrombi and those with arterial thrombi, and the efficacy of endovascular thrombectomy is different according to the thrombus composition [8]. Thus, it can be possible to clarify cryptogenic stroke etiology basing on thrombus histology. Although it may be possible, that histopathology of retrieved thrombi may be affected in patients who received intravenous thrombolysis procedure before endovascular mechanical thrombectomy was performed.
There is also little data on the correlation of radiological results with the histological qualitative structure of the thrombus. It is proposed that the presence of hyperdense artery sign (HAS) indicates more red blood cells (RBC`s) components and better revascularization outcomes [9]. The presence of thrombus in the heart and its impact on histopathology of cardioembolic thrombi has never been investigated, although it can provide an important clinical information about the thrombus composition on acute pretreatment stage.
However, only a few studies have analyzed the outcome of endovascular therapy recanalization and the neurological clinical outcome in correlation with the histological composition of the thrombus. Successful reperfusion is routinely achieved in 70-80% of cases [10]. That's why a major and unanswered question in vascular neurology is about the poor clinical outcome of stroke patients with endovascular therapy. Of course, there are several already known factors such as time, collaterals, and vascular anatomy, however, it cannot be ruled out that thrombus structure is also important. For instance, it has been proposed that thrombi with a high white blood cells (WBC`s) fraction are related to more organized thrombi of cardioembolic origin associated with less favorable recanalization and clinical outcome in acute ischemic anterior circulation stroke [11]. A better understanding of thrombus type and properties can significantly improve stroke treatment outcomes.
The aim of this study was to investigate histopathological features of different stroke emboli subtypes retrieved during mechanical thrombectomy and compare radiological findings and stroke outcomes with histological profile of thrombi.

Materials and Methods
The study is classified as a single-center analytical observational study. Forty patients with cerebral infarction and large magistral artery occlusion, during a period of three months (13 January 2021 -1 July 2021), have received acute reperfusion therapy with endovascular thrombectomy in Pauls Stradiņš Clinical University Hospital (Riga, Latvia). Immediately after clot retrieval, thrombus material was fixed in phosphate-buffered (10%) formalin. The total of 24 included patients were treated with intravenous recombinant tissue-type plasminogen activator (IV tPA) before endovascular thrombectomy.

Study Population
Patient data (age, gender, stroke subtype) were extracted from the medical records. Major clinical characteristics of patients included in the study could be found in Table S1 (see supplementary materials).

Pre-interventional Imaging
Unenhanced brain Computed Tomography (CT) and contrast-enhanced Computed Tomography Angiography (CTA) of the brachiocephalic arteries were performed in all patients either in our center or in the regional hospital. Brain Computed Tomography Perfusion (CTP) imaging was performed, if time from symptoms onset to hospital arrival exceeded 4,5 hours (26 patients). Unenhanced brain CT retrospectively was analyzed for presence or absence of the HAS. HAS was determined by an increased radiodensity of an artery and Hounsfield (HU) value was not specifically used for measurement of clot density. 29 patients during the acute stroke CT protocol have received unenhanced thoracic CT examination. Unenhanced thoracic CT was retrospectively analyzed for the presence or absence of intracardiac thrombi.

Endovascular Techniques
All procedures were performed using arterial access under sedation or general anesthesia according to present clinical status and stroke severity. The procedures were performed in a dedicated angio-suite (biplane Artis Zee, Siemens). Femoral access was obtained with an 8F vascular sheath and 6F long vascular sheath (Neuro max 088, Penumbra Inc.) The guide was advanced distally, as far as safely possible, into the internal carotid or in the vertebral artery, according to previous imaging findings. The aspiration catheter (Ace 60 or 68, Penumbra) was advanced to the level of the thrombus over any microcatheter and microwire the operator chooses. Under road map guidance, the wire and microcatheter were navigated to or past the thrombus. Over this platform, the larger aspiration catheter was delivered and positioned immediately adjacent to the site of occlusion. The microcatheter and wire were removed, and aspiration was applied via the aspiration pump that is part of the Penumbra separator system. If aspiration failed, the Solumbra technique as rescue switching strategy was performed advancing a stent-retriever (Solitare, Covidien) through the 0.025-inch microcatheter distal to the clot with an aspiration catheter (Ace 60 or 68, Penumbra) at the clot face. The microcatheter was then removed till stent retrieval proximal markers. Intermediate catheter was connected to a negative pressure suction pump (Penumbra) for continuous aspiration. Three to five minutes later, the aspiration catheter was slowly withdrawn until the blood flow velocity returned to normal in the connecting pipe of the negative pressure pump.

Histological and Immunohistochemical Analysis of Retrieved Thrombi
Histopathologically 40 thrombi were investigated. The size of each thrombus was registered. The formalin-fixed specimens were embedded in paraffin, cross-sectioned at 4-µm thickness and stained with haematoxylin-and-eosin (H&E) with Dako Autostainer link 48.
Subsequently, based on H&E staining, the percentage of erythrocytes, leukocytes and fibrin was measured in investigated thrombi, as well as microscopically was measured the area of these cells in mm2. The area of the thrombus components was measured with Nikon's program Nis Elements D. From the H&E image of each thrombus, each component was outlined and the area obtained, which was then converted to a percentage of the total thrombus area. All thrombi were stained immunohistochemically, according to manufacturer`s (Dako) instructions, for CD3 positive T cells, CD68 positive monocytes, von Willebrand factor positive thrombocytes and CD15 positive leukocytes. Their amount was categorically measured ranging "+, "++" and "+++". Where "+" presence of cells occupying < 25% of the tissue, "++" presence of cells occupying 25 to 50% of the tissue and "+++" presence of cells occupying > 50% of the tissue.

Stroke Subtypes with Characteristic Histopathological Profile
Ischemic stroke patients, included in this study, were divided into three groups -cardioembolic, atheroembolic and cryptogenic group. In our study groups were defined as follows: cardioembolic group -patients with atrial fibrillation and/or proved thrombi in left atrium appendage. Atheroembolic group -patients with large artery atherosclerosis and/or artery dissection. Cryptogenic pathogenesis group -patients with ischemic stroke of unknown etiology -either fully investigated without any pathological findings or incompletely investigated. Histopathological findings of retrieved thrombi were compared between three mentioned groups.

Statistical Analysis
Statistical analysis was performed using SPSS® (IBM SPSS Statistics for Windows, Version 25.0). All tests were statistically significant if p value was less than 0,05. As the data did not fit the normal distribution, only non-parametric statistical methods were used. For the analysis of dependencies between nominal and scale data, were applied Mann-Whitney U test and Kruskal-Willi's test. For finding correlations between nominal data was used Fisher`s exact test. Standard deviation as an additional descriptive parameter was not applied, as data did not fit a normal distribution. As additional descriptive parameters were chosen median with quartiles.

Impact of IV tPA on Thrombus Composition
The total of 24 patients have received IV tPA before mechanical thrombectomy. Only 16 patients underwent mechanical thrombectomy procedure alone. There were no statistically significant differences in thrombus composition comparing data in both groups (with and without IV tPA )-see  Although there was a tendency to be less T lymphocytes in the thrombi of patients, who did not receive IV tPA (p = 0,056), see Table 5.

Impact of Radiological Findings on Thrombus Histopathology
In total, 23 patients had HAS on non-enhanced head CT and 17 patients did not. Histopathological findings in both patients` groups, that is with and without HAS, is seen in Table S8 (see supplementary materials). The mentioned information shows, that HAS did not correlate with thrombus size, histological composition of thrombus, as well as immunohistochemical profile. There was no statistically significant correlation between HAS and outcomes.  In total, 29 patients have performed non-enhanced thoracic CT, and 11 patients did not receive this examination. From those, who had this investigation performed, 21 had thrombus in heart, whereas 8 had not.
Patients with visualised thrombus on CT in heart had larger erythrocyte area in thrombus, than those without visualised thrombus in heart (p=0,047). Other parameters did not differ statistically significantly and may be seen in

Impact of Thrombus Histopathology on Ischemic Stroke Outcome
Ischemic stroke outcome was measured with National Institutes of Health Stroke Scale (NIHSS) and modified Rankin Scale (mRS). NIHSS score was analyzed as scale data and mRS score as ordinal data with 4 groups: first group -mRS score 0-3, second group -mRS score 4, third group -mRS score 5, fourth group -mRS score 6. Analyzing ischemic stroke outcomes depending on each thrombus histopathology, there were not found any statistically significant differences in histopathological thrombus composition.

Impact of IV tPA on Composition of Thrombi
Analyzing our study results we have found, that IV tPA did not affect the results significantly, but there was a tendency to be less T lymphocytes in the thrombi of patients, who did not receive this treatment. It is known from previous research, that T cells are among the first cells to be recruited within the atheroma, and they are enriched in unstable plaques [12,13], it is known from cardiovascular studies that unstable plaques easier dissolve with IV tPA. Thrombi retrieved after IV tPA showed decreased number of T lymphocytes on histopathological analysis which confirms beforementioned hypothesis. It should be mentioned that there is lack of similar evidence in the stroke medicine. Other research has shown that RBC-rich thrombi respond better to r-tPA than plateletrich or white thrombi [14]. Nevertheless, in our study investigated thrombi with and without received IV tPA did not had statistically significant differencies concerning other thrombus components.

Histopathological Differences of Thrombi by Stroke Etiology
In our study, thrombi in cryptogenic strokes demonstrated slightly larger leukocytes area (median -4,0 mm2) comparing with cardioembolic thrombi (median -1,6 mm2; p=0,099), although this result was not statistically significant. No differences were found between cryptogenic and atheroembolic thrombi, as well as between cardioembolic and atheroembolic thrombi. Nevertheless, we have found, that cryptogenic thrombi showed statistically significantly lower fibrin/leukocytes ratio than atheroembolic thrombi (median -5,8 and 14,0 respectively; p=0,026), the same tendency, but not statistically significant, comparing cryptogenic thrombi and cardioembolic thrombi (median -5,8 and 11,5 respectively; P=0,067). No statistically significant differences were found comparing cardioembolic and atheroembolic thrombi (p=0,330). That means that fibrin area to leukocyte area ratio in cryptogenic and cardioembolic thrombi was almost similar and at the same time lower than in atheroembolic thrombi, which in turn means that as leukocytes area almost do not differ between groups, fibrin area in both cryptogenic and cardioembolic thrombi is lower than in atheroembolic thrombi. That is also proved by our study following results: atheroembolic thrombi had a higher percentage of the fibrin area versus total thrombi area than cryptogenic thrombi (median -85,0% and 79,0% respectively; P=0,038). The same tendency, but not statistically significant was found between atheroembolic and cardioembolic thrombi (median -85,0% and 82,5% respectively; P=0,099). There were no any statistically significant differences comparing both cryptogenic and cardioembolic thrombi. That means, we can propose fibrin/ leukocytes ratio as a new precise index for defining the fibrin area to total thrombus area in thrombi retrieved in stroke patients. One previous study came to the conclusion, that most large artery atherosclerosis thrombi are fibrin and platelet rich at plaque site and become fibrin and RBCs rich while occluding brain arteries [15]. Other study proposed, that cardioembolic thrombi are RBCs rich and have lower fibrin content compared with large artery atherosclerosis thrombi [16]. This corresponds to the traditional concept that cardioembolic thrombi forming in regions of stasis or slow flow are mainly composed of entrapped RBCs, and thrombi occurring in the context of atherosclerotic large arteries are mainly composed of fibrin and platelets [17,18]. In our study, making comparison between quantitative data of fibrin and erythrocytes content in thrombi, the conclusions were close to the traditional concept, but analysing retrieved material by fibrin or erythrocyte predominance in the thrombi, we had to admit, that not only atheroembolic, but also cardioembolic thrombi predominantly were fibrin rich. Some studies came to the similar conclusion -for instance, Niesten et al, 2014 [2] investigated 22 thrombi retrieved after mechanical thrombectomy in patients with acute stroke. They reported that there were no significant differences in the proportion of fi-brin and platelets between different stroke subtypes and even that clots originating from large-artery atherosclerosis (LAA) had the highest percentage of RBC`s compared with other stroke subtypes. But it is important to know, that in their study, the patient population included a small number of patients with LAA (n = 8) and cardioembolism (n = 6). This factor was the significant limitation. The same was about Maekawa et al, 2018 [19] research results, where they had found, that the proportion of fibrin in retrieved thrombi was higher in patients with cardioembolism than in those with no cardioembolism. But in this study there was only 5 patients with LAA, from whom 3 had erythrocyte rich and 2 fibrin rich thrombi. Sporns et al, 2017 [20] had larger study, where it was proved, that cardioembolic thrombi had higher proportions of fibrin/platelets and less erythrocytes. It is important to admit, that traditional concept about pathophysiology of cerebral thrombi has mainly been derived from coronary circulation studies, but in the era of mechanical thrombectomy, new results were shown. Our study had proved, that majority of cardioembolic thrombi were fibrin rich, although some of them had predominance of erythrocytes, as well as in quantitative measuring the amount of fibrin was higher in atheroembolic thrombi. Data from other authors` studies show, that the infiltration of thrombi with leukocytes increases as the thrombus matures [8].
In another study the WBC`s proportion was similar in thrombi of cryptogenic and non-cardioembolic stroke patients (7,1% and 6,5% accordingly; p=0.487), while cardioembolic had a higher WBC`s proportion (9,1%) [21]. Boeckh-Behrens et al also concluded, that a higher percentage of WBC`s in the thrombus was associated with a cardioembolic etiology [11]. Although aged forms of thrombus may be more common in cardiac thrombi [8], some other studies did not found any statistically significant differences in leukocyte amount among different stroke subtype groups [16,19]. Our study also did not find statistically significant differences in leukocyte amount among different stroke subtype groups, which could mean that studied thrombi were of similar age and stage of the organization.

Impact of Radiological findings on Histopathology of Thrombi
In general, HAS corresponds to the thrombus composition to some degree [22] and is associated with RBC rich thrombi [9]. In addition, patients with a HAS are more likely to have a good angiographic outcome than those without [23]. Better results also depend not only on the composition of thrombi, but also on HU value -higher thrombus HU values are predictive of successful recanalization [24]. In our study HAS did not correlate neither with the histopathological composition of thrombi nor stroke outcomes. This may be due to the low number of included patients.
Our study has confirmed that patients with CT visualized thrombus in the left atrium appendage had larger erythrocyte area in thrombus than those without mentioned finding (p=0,047). This group is defined as cardioembolic group. It is widely accepted, that fresh red thrombi, containing mixtures of fibrin and RBCs, originate from low flow-regions while white thrombi, existing mainly out of platelets and fibrin, arise in regions of fast moving blood [25,26]. In cerebral stroke, thrombi from LAA are thought to originate from regions of high velocity blood stream, while most cardioembolic thrombi are assumed to develop from low flow-regions, such as the left atrial appendage [2]. The results of our study fully correspond to conventional point of view. But nevertheless, there may be absolutely opposite results -in other studies, patients with LAA had the highest percentage of RBCs [2,21,27]. Such different results may be due to some limitations -interpretations about the pathophysiology of cerebral thrombi have mainly been derived from coronary circulation studies [2], as well as usage of IV tPA before mechanical thrombectomy and following thrombi retrieval, that may have an impact on the composition of thrombi. Our study finding makes us consider including unenhanced thoracic CT examination in stroke patients` investigation protocol.

Impact of Histopathology of Thrombi on Stroke Outcome
Thrombi could roughly be divided into fibrin-and RBCsrich clots [28]. In literature, it is highlighted, that RBCs-rich clots are associated with significantly higher recanalization rates, reduced number of maneuvers and a shorter mean recanalization time than fibrin-rich clots [29]. That may be explained with the fact that fibrin-rich clots have a significantly higher coefficient of friction than RBC-rich clots and therefore have a stronger interaction with the vessel wall and are harder to remove from the vessel wall [30]. Higher percentage of WBC`s in the thrombus are associated with less favorable recanalization (TICI<3) and clinical outcome (NIHSS score at discharge and mRS scores up to 90 days) [11]. The amount of WBC`s represents a marker of the level of organization, this is the age of the thrombus. The age, as well as stage of the organization of the thrombus are important for the success of the mechanical thrombectomy, as the stability and strength of adherence to the vessel wall of the targeted clot are thought to increase with time, making the removal of the thrombus more difficult [11]. In our study, the ischemic stroke outcome in patients after performed endovascular mechanical thrombectomy did not have any correlations with thrombus histopathology. Such results may be due to limited study population sample.

Study Limitations
Part of the patients, that were included in the research, before mechanical thrombectomy procedure had treatment with IV tPA. Although there was not statistically significant impact of this procedure on the results, the distribution of patients by performed IV tPA was not homogenous.
The limited number of patients was included -only 40 patient stroke thrombi were investigated.
The spreading of patients by stroke etiology was heterogenous, making it difficult to compare the groups.
Although Mann-Whitney test has shown statistically significant differences between atheroembolic and cryptogenic thrombi groups by the following parameters -percentage of fibrin area in thrombus and fibrin area to leukocyte area ratio in thrombus, Kruskal-Wallis test had not shown statistically significant differerence comparing three groups simultaneously (although it had shown tendency to be statistically significant). Also Bonferroni correction was not applied.

Conclusions
There was found, that cryptogenic thrombi had partial histologic similarities to cardioembolic thrombi, i.e., lower fibrin percentage in thrombi area. Therefore, patients with cryptogenic stroke should probably undergo cardiac examination including long-term heart rhythm monitoring.
As to radiological findings, patients with CT visualized thrombus in the left atrium appendage had larger erythrocyte area in thrombus than those without a mentioned finding. It may be useful to consider the inclusion of unenhanced thoracic CT examination in stroke patients` investigation protocol.
Histopathology of thrombi was not affected by IV rt-PA therapy which may rise questions about usefulness of rt-PA in patients with large vessel occlusion.
The ischemic stroke outcome in patients after performed endovascular mechanical thrombectomy did not depend on thrombus histology or its immunohistochemical profile.