Background Current studies on clot characterization in acute ischemic stroke focus on fibrin and red blood cell composition. Few studies have examined platelet composition in acute ischemic stroke clots. We characterize clot composition using the Martius Scarlet Blue stain and assess associations between platelet density and CT density.
Materials and method Histopathological analysis of the clots collected as part of the multi-institutional STRIP registry was performed using Martius Scarlet Blue stain and the composition of the clots was quantified using Orbit Image Analysis (www.orbit.bio) machine learning software. Prior to endovascular treatment, each patient underwent non-contrast CT (NCCT) and the CT density of each clot was measured. Correlations between clot components and clinical information were assessed using the χ2 test.
Results Eighty-five patients were included in the study. The mean platelet density of the clots was 15.7% (2.5–72.5%). There was a significant correlation between platelet-rich clots and the absence of hyperdensity on NCCT, (ρ=0.321, p=0.003*, n=85). Similarly, there was a significant inverse correlation between the percentage of platelets and the mean Hounsfield Units on NCCT (ρ=−0.243, p=0.025*, n=85).
Conclusion Martius Scarlet Blue stain can identify patients who have platelet-rich clots. Platelet-rich clots are isodense on NCCT.
Statistics from Altmetric.com
Mechanical thrombectomy is standard of care for treatment of acute ischemic stroke (AIS) secondary to large vessel occlusion.1 The composition of the occlusive clot has been shown to significantly influence the outcome for patients treated with both recombinant tissue plasminogen activator (rt-PA) and mechanical thrombectomy devices.2–7 Retrieval of occlusive clots has afforded researchers the opportunity to study their histological composition, thereby increasing our understanding of their inherent characteristics and in the future potentially helping the interventional community to individualize patient care based on the suspected composition of the occlusive clot. In order for this to have an impact in the clinical setting, accurate characterization of all clot components as well as correlations between clot composition and diagnostic imaging findings must first be identified.
Initial studies assessing correlations between clot histological composition and diagnostic imaging have typically used the standard hematoxylin and eosin (H&E) histopathological stain. Red blood cell-rich clots have been shown to be associated with a hyperdense artery sign (HAS) on non-contrast CT (NCCT) and a positive susceptibility vessel sign (SVS) on MRI, both of which are associated with improved outcome after treatment.8 9 However, the H&E stain fails to accurately distinguish between fibrin and platelets, and thus studies using H&E have tended to refer to clots in terms of fibrin/platelet or fibrin/other content.10 11 The quantity of platelets can vary greatly between clots and it is well established that certain patients with AIS benefit significantly from treatment with antiplatelet therapy.12 We present a histological stain, Martius Scarlet Blue (MSB), capable of identifying platelet-rich regions of AIS clots and assess associations between platelet content and diagnostic imaging findings.
Materials and methods
Patient selection and clinical data
This investigational study was performed as part of the multi-institutional Stroke Thromboembolism Registry of Imaging and Pathology (STRIP) registry. The study was institutional review board approved and HIPAA compliant. Patients were included in the study if they were aged >18 years, had undergone mechanical thrombectomy treatment for AIS, and had a NCCT scan prior to endovascular treatment. Patients treated with rt-PA alone, patients without an available NCCT scan prior to endovascular treatment, and those with incomplete data were excluded from the study (figure 1). An example of the data abstraction form is provided in the supplementary material (see online supplementary file 1).
Prior to endovascular treatment, each patient had a NCCT scan performed and expert readers at each site evaluated the mean and maximum clot attenuation on NCCT as measured by the placement of regions of interest (ROIs) along the clot. A positive HAS was defined as ≥50 Hounsfield Units (HU). Because this is a multi-institutional study, there was variability in the CT scanners and techniques used. All imaging findings were reported at the site where the thrombectomy was performed. There was no core laboratory for imaging.
Clot collection, processing, and histology
On retrieval, each clot was immediately fixed in 10% phosphate-buffered formalin. Clots were shipped to the histology core facility. On arrival at the core facility, gross photos were taken of each clot. All clots were then processed using a standard tissue processing protocol and embedded in paraffin. The formalin-fixed paraffin-embedded clot material was cut into 3–5 µm sections. Representative slides from each clot were stained with H&E and MSB stains as per their protocols (see online supplementary files 2 and 3). Following staining, a representative MSB-stained slide was sent for whole slide scanning (Aperio ScansScope AT Turbo, Leica Biosystems). Histologic quantification was performed on the digital slide using Orbit Image Analysis Software (Orbit Image Analysis, www.orbit.bio), as described previously.13
The slides were dried by heating to 56˚C for 2 hours in an oven, deparaffinized in xylene (2×10 min), followed by rehydration through alcohols and in distilled water. Sections were pretreated with 0.1 mol/L citric acid buffer in a microwave for 15 min. Slides were left to cool at room temperature for 30 min before being rinsed in Tris-buffered saline (TBS). The slides were incubated with 4% normal donkey serum in TBS buffer for 30 min at 37°C, followed by incubation with primary antibodies CD42b (mouse monoclonal antibody, pre-diluted; Abcam) and fibrinogen (rabbit polyclonal antibody, 1:200; Dako) in TBS buffer for 1 hour at 37°C, then 4°C overnight. After incubation, the sections were rinsed in TBS buffer, followed by incubation with secondary antibodies (Cy3 conjugated donkey anti-mouse IgG (1:200); Alexa-Flour 488–conjugated donkey anti-rabbit IgG, (1:100); Jackson ImmunoResearch) for 2 hours at room temperature. The sections were rinsed in TBS (4×5 min) and counterstained with DAPI (1:250). Finally, the sections were rinsed in TBS and then dehydrated through alcohols, cleared in xylene, and mounted with EZ-Mount. Negative controls were performed with non-immune normal serum used instead of the primary antibody. The sections were viewed and imaged with a fluorescence confocal microscope.
All statistical correlations were assessed using IBM SPSS Statistics 22. Correlations between clot composition and HU density on NCCT imaging were assessed using the χ2 test. The Mann–Whitney U test for non-parametric data was used to assess correlations between continuous and categorical variables.
In total, 85 patients with a diagnosis of AIS and treated with mechanical thrombectomy met the inclusion criteria and were included in the study. Table 1 shows the clinical demographics of the patient cohort. The median age of the patients was 66 years (range 20–91 years). Fifty-one percent of patients had been treated with rt-PA prior to mechanical intervention. The majority of cases had an internal carotid artery (ICA) or middle cerebral artery (MCA) occlusion (40% and 84%, respectively) and 25 cases (30%) had occlusions that spanned two or more locations (data not shown). Stentriever devices were used in 61% of patients while aspiration alone was used to treat the remaining 39% of patients. Thrombolysis in Cerebral Infarction (TICI) 2b/3 was achieved in 96% of patients treated, with a mean number of passes of 2.1±1.4.
MSB stain identifies platelet-rich regions in clots
The MSB stain allows for a significantly better differentiation of the major components of clots than the H&E stain, as can be seen in figure 2. The MSB stain can accurately identify the presence of platelets as distinct from fibrin strands (figure 2A and C). Platelets cannot be accurately distinguished from fibrin in the H&E stained image, even at high magnification (figure 2B and D). The ability of the MSB stain to identify platelet-rich regions is confirmed by immunofluorescence staining, which demonstrates the presence of platelets, as detected using an antiplatelet (CD42b) antibody, in the areas identified as being platelet-rich by the MSB stain (figure 3).
AIS clot compositions
The composition of the major clot components was heterogeneous among the patient cohort, as can be seen in figure 4. Red blood cells and fibrin were typically the dominant components of AIS clots, with their mean compositions being 39.4% and 41.8% respectively, while the average white blood cell composition was 3.9%. The composition of platelets/other components varied from 2.5% to 72.5% of the total area with a mean value of 15.7%.
Treatment with rt-PA prior to mechanical thrombectomy resulted in significantly reduced platelets (12.2% vs 19.4%, p=0.018) and white blood cells (2.6% vs 3.5%, p=0.024) but did not significantly affect red blood cells (42.9% vs 35.8%, p=0.173) or fibrin (42.3% vs 41.3%, p=0.937). There was no significant difference in clot composition between cases treated with stentriever devices and aspiration. Clot composition also did not significantly affect the final TICI score nor the number of procedural passes.
Platelet content versus clot attenuation on NCCT
There was a significant correlation between platelet-rich clots (≥15.7%) and the absence of a HAS (<50 HU) on NCCT, (ρ=0.321, ρ=0.003*, n=85). In addition, there was a significant inverse correlation between the percentage of platelets and the mean HU on NCCT, as shown in figure 5 (ρ=−0.243, p=0.025*, n=85). An example of a patient with an isodense vessel on NCCT prior to treatment who was subsequently found to have had platelet-rich thrombi is shown in figure 6. No significant correlations between the other major clot components and a HAS were observed.
In this study we demonstrate that the MSB stain can identify platelet-rich regions in AIS clots. These findings are important as we are now able to characterize the histological composition of AIS clots more comprehensively than ever before by including one of the most essential components to thrombosis—namely, platelets. We also demonstrate that platelet-rich clots, as determined by the MSB stain, are isodense on NCCT scans. These findings are important as they (1) provide additional insight into the composition of clot in large vessel occlusion, (2) allow for more accurate radiological-pathological correlation between clot histology and density on CT, and (3) could potentially provide insights into treatment strategies for secondary stroke prevention.
Previous studies investigating the composition of AIS clots using basic histological stains have failed to accurately distinguish between fibrin and platelets.2 8 10 11 14 15 Activation of platelets has long been known to initiate the coagulation cascade and, logically, platelets and platelet-related factors are key components of clots.16 The ability to identify platelet-rich regions as distinct from fibrin now allows us to investigate correlations between platelet content and clinical and procedural parameters, in addition to more accurately representing the ‘true’ fibrin content of the clots.
In order for this to have an impact on the acute treatment of stroke, correlations between histological and mechanical clot characteristics and diagnostic imaging modalities must first be identified.17 18 Accurate identification of clot composition on diagnostic imaging would provide valuable information on the degree and ease of revascularization using mechanical thrombectomy techniques as the mechanical properties of clots are known to change significantly depending on composition.19 We demonstrate, for the first time, that platelet-rich clots are isodense on CT. This suggests that patients whose clots appear isodense on NCCT are more likely to be platelet-rich. Platelet-rich clots have long been known to be more resistant to standard thrombolytic therapy,20 21 and therefore those patients might benefit significantly from treatment with novel antiplatelet therapy that has previously been shown to improve outcomes in certain patients.12
The ability to quantify platelet composition might also play a significant role in the medical management of patients with AIS in order to prevent a recurrent stroke, as the cause of a secondary stroke is typically directly related to the original stroke. Accurate knowledge of the platelet content of the occlusive clot would allow clinicians to better select patients who might derive a clinical benefit from dual antiplatelet therapy, which would be of particular importance in patients with cryptogenic stroke.
Our study has limitations. First, TICI scores were self-reported at each site and not measured using a central core laboratory which may have resulted in some site-to-site variability. Second, while the MSB stain is more accurate than the H&E stain in identifying major clot components, it still does not specifically identify other potentially key clot components such as von Willebrand factor and calcification. Therefore, the authors represent this subgroup as ’platelet and other' components as we acknowledge that there are potentially other components in addition to platelets in these regions. Immunohistochemical analysis using specific antibodies is the only way to accurately distinguish between platelets and platelet-related factors such as von Willebrand factor. An additional histological stain such as Von Kossa staining should be used if calcification is suspected. There was variability in the make and model of CT scanners used between different sites and all imaging findings were reported at the site where the thrombectomy was performed, not at a core imaging laboratory, which might result in some slight intra-site variability when measuring HU density.
MSB stain can identify patients who have platelet-rich clots. These platelet-rich clots are isodense on NCCT. This correlation may inform acute treatment approach and thereby lead to better patient outcomes.
The authors would like to gratefully acknowledge the invaluable contributions made by the Interventional, Nursing and Clinical coordination teams at each of the sites included in the STRIP registry. The authors also wish to thank our industrial partners Cerenovus.
STF and SW contributed equally.
Contributors STF, DFK, DK, RK, and WB were all involved in all stages of the manuscript from concept design to drafting the manuscript. SW, DD, and AD contributed to the histological staining and quantification of the cases. MJG, JC, and ASP collected clots and extracted corresponding clinical data at University of Massachusetts Medical School. KFL and ICT collected clots and extracted corresponding clinical data at Baylor University Medical Center. RH, ES, and AA collected clots and extracted corresponding clinical data at Lyerly Neurosurgery/Baptist Neurological Institute. MAA and AMD collected clots and extracted corresponding clinical data at University of Calgary. RN collected clots and extracted corresponding clinical data at Grady Memorial Hospital and Emory University. VMP collected clots and extracted corresponding clinical data at Toronto Western Hospital. PK collected clots and extracted corresponding clinical data at University of Tennessee Medical Center. JEDA and YK collected clots and extracted corresponding clinical data at Abbott Northwestern Hospital. AJY collected clots and extracted corresponding clinical data at Texas Stroke Institute. All other authors reviewed, edited, and approved the final manuscript prior to submission.
Funding This work was supported by the National Institutes of Health grant number (R01 NS105853) and the European Regional Development Fund and Science Foundation Ireland grant number (13/RC/2073).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Deidentified participant data and corresponding histological data will be made available upon reasonable request.
Patient consent for publication Not required.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.