Open Access
Add to BibSonomy
Abstract
Acute ischemic stroke caused by large vessel occlusion is treated with endovascular thrombectomy, but treatment failure may occur when clot composition and thrombectomy technique mismatch. In this proof-of-concept study, diffuse reflectance spectroscopy (DRS) is evaluated for identification of clot composition ex vivo. DRS spectra and histology were acquired from 45 clot units retrieved from 29 stroke patients. DRS spectra correlated to clot RBC content, R= 81, p < .001, and could discriminate between RBC-rich and fibrin-rich clots, p < 0.001. Sensitivity and specificity for detection of RBC-rich clots were 0.722 and 0.846 respectively. Applied in an intravascular device, DRS could potentially provide intraprocedural information on clot composition that could increase endovascular thrombectomy efficiency.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
Corrections
18 May 2022: A correction was made to the citation of the supplemental document.
1. Introduction
The introduction of endovascular treatment (EVT) for acute ischemic stroke (AIS) caused by large vessel occlusions was a paradigm shift in stroke management and resulted in significantly improved outcomes [1–10]. EVT can be performed using a self-expanding stent retriever, through contact aspiration of the clot using a large bore catheter, or with a combination of the two techniques [11,12]. Regardless of the method used, achieving reperfusion at the first attempt has been proven to decrease mortality and result in better clinical outcomes [13,14].
However, in up to 20% of the cases reperfusion is not achieved at all, and only about 50% of the patients reach functional independence (modified Rankin scale 0–2) at 90 days after treatment [6,15]. Thrombectomy failures or prolonged procedures with several failed attempts before clot removal may play important roles.
A key determinant for successful revascularization is the composition of the blood clot. Clots mainly contain red blood cells (RBCs) and fibrin, but also other constituents like leukocytes, platelets, von Willebrand factor, neutrophil extracellular traps and extracellular DNA [16–19]. Clots rich in fibrin are often more difficult to treat with EVT due to a firmer consistency and greater friction against the vessel wall, resulting in lower revascularization rates [20–22]. Clots rich in RBC, on the other hand, are associated with faster and less complicated EVT procedures with better patient outcomes [23–26]. Large bore contact aspiration catheters and stent retrievers vary in their efficacy depending on the clot type and devices targeted at specific clot types have been developed [27–29]. Consequently, adapting the EVT technique to the composition of the occluding clot may improve EVT success rate. However, this requires an accurate method for detection of clot composition before thrombectomy is attempted.
To date, the clinically available methods for clot composition characterization have been restricted to indirect methods like conventional CT and MRI examinations. The increasing recognition of clot composition as an important prognostic factor in AIS has incited several attempts to develop more accurate, intravascular techniques for clot composition characterization. Spectral Optical Coherence Tomography (OCT) [30], Optical Coherence Elastography (OCE), ultrasound Shear Wave Elastography (SWE) [31] and Electrochemical Impedance Spectroscopy (EIS) [32,33] are all examples of such attempts. The first method to be successfully applied for in vivo characterization of clots in AIS was however Diffuse Reflectance Spectroscopy (DRS) [34]. In DRS, tissues are illuminated with white light that is scattered and absorbed within the tissue. The reflected light is collected and spectrally analyzed. The intensity and shape of the spectra are determined by the properties of the examined tissues. Since hemoglobin and its derivates have strong optical absorption features, the method is particularly suitable for recognition of RBC content [35,36].
In our previous work, we have shown that DRS can identify the composition of laboratory manufactured blood clot analogues in vivo, and that the technology can be incorporated in a fully working micro guide wire-like device [34]. The aim of the current study was to validate the ability of DRS to determine the RBC content in human blood clots extracted from AIS patients treated with EVT.
2. Methods
2.1 Patient selection
Adult patients ($\ge $ 18 years old) treated for AIS with EVT at the study center were eligible for inclusion. The Department of Neurosurgery and Interventional Radiology at the study center is the only provider of neurointerventional care in the region, serving a population of roughly 2.5 million.
Cases where a blood clot of >5 mm could be extracted and immediately measured with DRS were included irrespective of prior treatment with rtPA. Patients with suspected or confirmed infection with COVID-19, were excluded. Patient enrollment started on December 20, 2019 and was closed on November 17, 2020, when 29 patients had been included. The study was approved by the National Ethical Review Authority (Dnr 2018/2488-31/2 and 2019-00344).
2.2 Mechanical thrombectomy procedure
Mechanical recanalization was performed according to the department’s standard procedure. Clots were removed with a stent retriever, often in combination with an aspiration catheter. The choice of device was made by the treating neurointerventionalist. Directly after retrieval, clots were gently removed from the device, put in a test tube with 0.9 mg/ml NaCl and stored at 4°C for a maximum of 4 hours before examination with DRS. The visual appearance of the extracted clot was annotated by the treating interventionist as “red”, “mixed” or “white”.
Fig. 1. The DRS system 1a) A tungsten halogen broadband light source (LS) with light ranging from 360 to 2500 nm is coupled to a 200 µm core diameter optical fiber, that is integrated in a handheld test probe. The light exits the fiber at the tip of the probe and reflected light from the examined tissue is collected by another optical fiber. The collected light is led via a Y-junction (fiber splitter) to two optical spectrometers. One spectrometer resolves light in the visible and near-infrared wavelength range of 450 to 1100 nm (VIS) and one covers the near infrared and infrared wavelengths from 900 to 1600 nm (NIR). The spectra of the collected light act as a fingerprint of the composition of the examined tissue, allowing identification of the examined tissue. 1b) Absorption spectra of typical blood clot constituents, including deoxygenated hemoglobin (Hb) and oxygenated hemoglobin (HbO2).
2.3 DRS system
The general principles of DRS, instrumentation and calibration have been described previously [34,37] and the experimental setup is schematically illustrated in Fig. 1. In short, light from a tungsten halogen broadband light source (360 to 2500 nm) is coupled to an optical multimode fiber (200 $\mu $m diameter) that is integrated in a handheld probe. The light exits the fiber at the tip of the probe and reflected light from the examined tissue is collected by another optical fiber. The distance between the fibers at the probe tip is 0.45 mm. The collected light is led via a Y-junction (fiber splitter) to two optical spectrometers. One spectrometer (OceanOptics Maya 2000PRO, Ocean Insight, Orlando, FL, USA), resolves light in the visible and near-infrared wavelength range of 450 to 1100 nm, while the other (OceanOptics NIRQuest512, Ocean Insight, Orlando, FL, USA) covers the infrared wavelengths from 900 to 1600 nm. For data acquisition and controlling of the spectrometers, a custom software developed in-house was used (Labview).
2.4 DRS measurements
DRS measurements of the clots were performed 15 min to 4 hours after clot retrieval in a dark room with the clots positioned on a black paper sheet. Measurements were obtained with the probe directed both perpendicular and parallel to the black surface. The number of measurement positions per clot was determined by clot size. Multiple measurements were acquired along the length of the clot. In addition, one measurement was acquired at each end of the clot. The acquisition time for each position was 4.5 seconds. Measurement positions were documented on photographs of the clots. Before each measurement session the setup was optically calibrated with a Spectralon sample [43].
2.5 Histological analysis of clots
After DRS-measurements, the clot material was immediately placed in 10% formalin for fixation at room temperature for 24 hours and then embedded in paraffin. The clots were cut into 5-µm slices, and for every three slices one was stained with Hematoxylin & Eosin (H&E), one with Martius Scarlet Blue (MSB) and one with anti-CD42b (Invitrogen MA511642, clone 42C01). All slides were scanned at the resolution of 0.23 µm/pixel using a Hamamatsu NanoZoomer S360 Digital slide scanner (Product number C13220-01).
Clot specimen slides were evaluated by a pathologist experienced with characterization of blood clot compositions (O.A.). Quantification of the clot constituents were performed by training a random forest machine learning pixel classifier for each clot using QuPath image software (v0.2.0) [38]. Six different classes (RBC, fibrin, platelets, leucocytes, calcifications and “ignore”, the latter used to exclude the white background) were annotated on H&E-stained slides only. Quantification results were reported as percentage content of RBC, fibrin, platelets, leucocytes and calcifications. A detailed description of the pixel classifier is provided as Supplement 1.
2.6 Data analysis
2.6.1 Clot units and averaging of measurements
Blood clots causing AIS are heterogeneous in composition but may have distinct areas with a dominant structural component [18,19], and are often fragmented during extraction. To match DRS measurement data with clot histology with the highest possible granularity of the data, clot fragments as well as demarcated clot areas with a distinct appearance on both photographs and histology were defined as separate clot units. The measurements within a clot unit were averaged.
2.6.2 Diffuse reflectance spectral data analysis
Spectral analysis was performed with a custom software using Matlab (MathWorks Inc., Natick, MA). DRS data acquired from the ex vivo measurements were analyzed in the wavelength range of 450 to 1600 nm. Two principal methods were employed for analysis of the spectra: fitting of standard spectra from known clot constituents to the acquired DRS spectra, and analysis of the slope at different wavelength of the DRS spectra.
The measured spectra were fitted from 450 to 1600 nm with the model of Farrell et al., which is derived from diffusion theory using a Levenberg–Marquardt nonlinear inversion algorithm with the absorption spectra of known absorbers (Fig. 1(B)) as input parameters [35,36,39]. The main blood absorption region from 500 to 800 nm was overweighted to achieve a good blood fit. The exact fitting procedure has been described in detail by Nachabé et al. [43]. The absorption bands corresponding to oxygenated hemoglobin (450–600 nm), deoxygenated hemoglobin (≈760 nm) and methemoglobin (≈630 nm) [36], were tested for correlation to histological RBC clot content. Since RBCs are often aggregated in distinct regions of a clot [18], a correction factor named Rves that corresponds to regional clustering of strong absorbers (“vessel packing”), previously described by Rajaram et al. [39], was also analyzed for correlation to clot RBC content. Absorption coefficients for water and collagen, as well as parameters describing the wavelength-dependent scattering contributions from Mie and Rayleigh scattering, discussed in detail by Hahn et al. [40], were tested for correlation to histology. Absorption spectra of known blood clot components including deoxygenated and oxygenated hemoglobin, methemoglobin, water and collagen are presented in Fig. 1(b) [35,36,41,42].
The slope of the DRS spectra, i.e. the direction and steepness of the spectral curve, was also analyzed at different wavelengths and compared with the histologically determined content of RBCs, platelets, WBCs and fibrin. Grid Search was used to find the wavelength combination yielding the best correlation.
2.7 Statistical analysis
The DRS data showed non-normal distribution. Spearman’s rank test and linear regression analysis were used to evaluate correlation of DRS data to histological data. The significance level was set to 5% (p < 0.05). Log transformation of the DRS data was performed without achieving normally distributed data, and the linear regression analysis must be interpreted with caution. Univariable distribution of metric variables is reported by median and min-max range.
Fig. 2. Histology results Consecutive 5 $\mathrm{\mu} $m sections of paraffin embedded clots were stained with (a) Martius Scarlet Blue (MSB), that stains fibrin red and RBC yellow, (b) conventional hematoxylin and eosin (H&E) that visualize the overall clot composition and organization, and (c) anti-CD42b antibodies (Invitrogen MA511642, clone 42C01) that show the presence of platelets. Figure 2(d) presents the result of the software classification model of clot constituents, where RBC are coded yellow, WBC blue, fibrin red and platelets green. Figure 2(e) shows the histological composition of all clot units ordered by RBC content. Clot constituents are color coded according to the colors of the histological classification model presented in Fig. 2(d). Case 11 was found to be strongly calcified.
For classification, clots with fibrin > RBC according to histology were defined as fibrin-rich, and clots with RBC > fibrin were defined as RBC-rich, as previously described by Maekawa et al. [24]. A classification model based on the sum of the fitted DRS parameters for deoxygenated and oxygenated hemoglobin (Hb + HbO2), previously shown to correlate to RBC content [34], was created using logistic regression. Kruskal-Wallis test was used for statistical analysis. Analyses were performed in SPSS version 24 (IBM Corporation, Armonk, NY, USA).
3. Results
3.1 Baseline characteristics
Twenty-nine patients treated with endovascular thrombectomy due to AIS were included in the study. Median clot length was 13 mm (5–39) and median clot diameter was 2 mm (1–4) mm. The gross appearance was red in 7 clots, white in 7 clots and mixed in 15 clots. From the retrieved clot material 45 separate clot units were identified, whereof 16 were whole clots, 27 were clot fragments and 2 were within the same clot.
3.2 Histological clot composition
The clot composition was in general heterogenous, with regional aggregation of the constituents (Fig. 2(a)-(c)). The software classification model was trained on the histology results until reliable identification of the constituents was achieved (Fig. 2(d)). One case (number 11) was found to be strongly (66%) calcified. The classification resulted in the following distribution of clot constituents: RBC 28.7% [0.0–71.9]; fibrin 35.0% [8.8–92.3]; platelets 30.0% [4.3–81.9] and WBC 3% [1.0–12.7], (Fig. 2(e)).
3.3 DRS recordings
DRS was recorded at a total of 247 separate positions. Median measurement positions per clot unit was 5 (1–20). The averaged DRS spectra for each clot unit are presented in Fig. 3(a). The histologically strongly calcified clot case 11 (marked in magenta in Fig. 3(a)) was considered an aberrant case that was excluded from further analysis. After averaging the DRS data over each RBC content group (3b), the spectra showed typical features explained by the absorptive properties of the clot constituents (3c).
Fig. 3. DRS results 3a showing the averaged optical DRS spectra for each clot unit, colour coded according to histological RBC content, where blue represents RBC content < 15%, cyan 15 −30%, green 30–45% and red > 45%. The strongly calcified case 11 (magenta)was excluded from further analysis. The averaged DRS spectra for each RBC content group (Fig. 3(b)) expose the variation in spectral shape explained by different RBC content. In Fig. 3(c), the spectral differences between the groups are correlated to the absorption bands of individual chromophores and optical scattering properties of key clot constituents. With increasing RBC content, increased light absorption at 450–600 nm as well as around 630–760 nm were seen, corresponding to the main absorption bands of hemoglobin (450–600 nm), minor absorption peaks of methemoglobin (≈630 nm) and deoxygenated hemoglobin (≈760 nm). Differences in the overall shape of the spectra can be explained by the individual clot constituents having differently sized scattering structures. The overall slope of the spectra from visible to near-infrared was steeper for clots with low RBC content than for RBC-rich clots, probably because the large RBCs are predominantly Mie scatterers while the smaller Fibrin strands or Platelets are predominantly Rayleigh scatterers. This is mostly visible in the different slopes between 700 nm and 1400 nm.
3.4 Correlation of DRS and histological data
Deoxygenated and oxygenated hemoglobin as measured by DRS (Hb + HbO2) were positively correlated to the histological RBC clot content r = 0.78, p <0.001, R2 = 0.50 p <0.001 (Fig. 4(a)). Samples of (Hb + HbO2)-fits are presented in Supplement 1. Methemoglobin correlated more weakly to RBC content, r = 0.61, p < 0.001, R2 = 0.32, p < 0.001 (Fig. 4(b)), while the pigment clustering correction coefficient, Rves correlated clearly to RBC content r = 0.80, p < 0.001, R2 = 0.55 (Fig. 4(c)). As expected, the parameters with a positive correlation to RBC content had a weak negative correlation to fibrin and platelet content. No significant positive correlations between DRS parameters and fibrin, platelet or WBC were identified.
The analysis of spectral slopes revealed a robust correlation to RBC content. The slope parameter with the strongest correlation to histological RBC content is presented in Fig. 4(d), and was defined as the intensity at wavelength 1098 nm minus the intensity at 938 nm, divided by the intensity at 1100 nm, r = 0.81, p < 0.001, R2 =0.57 (p < 0.001).
Fig. 4. Correlation between DRS and histology 4a Correlation between the histological RBC content and a) the sum of the fitted DRS parameters deoxygenated and oxygenated hemoglobin, b) the fitted DRS parameter of methemoglobin, c) the pigment clustering correction factor Rves and d), a slope parameter derived from the optical spectra. Each dot represents the averaged result from histology and DRS for a single clot unit. The line shows the resulting linear regression. A.u. represents arbitrary units.
3.5 Classification of clots
Of the 44 clot units included in the analysis, 26 were defined as fibrin-rich and 18 as RBC-rich. The classification model based on the sum of the DRS parameters deoxygenated and oxygenated hemoglobin (Hb + HbO2) was able to significantly discriminate between RBC rich and fibrin rich clots (p < 0.001; Fig. 5(a)). AUC was 0.872, sensitivity and specificity for detection of RBC-rich clots were 0.722 and 0.846 respectively (Fig. 5(b)).
Fig. 5. Clot classification results 5a: Boxplot of the sum of DRS parameters deoxygenated and oxygenated hemoglobin (Y-axis, arbitrary units) as discriminator between fibrin-rich clots, defined as clots with histological fibrin content > RBC content, and RBC-rich clots, defined as clots with histological RBC content > fibrin content (p < .001). Figure 5(b): The result of the classification model expressed in a confusion matrix. The rows show the result of the histological classification of clots based on RBC content (ground truth), and the columns show the prediction of clot type based on the sum of DRS parameters deoxygenated and oxygenated hemoglobin. From the matrix, the sensitivity and specificity of the model can be calculated.
4. Discussion
This study demonstrates that DRS can characterize the composition of blood clots extracted from AIS patients. DRS spectra and clot RBC were clearly correlated, and the clot composition classification model performed with a sensitivity and specificity for detection of RBC-rich clots of 0.722 and 0.846, respectively. The differences in the optical spectra were mostly due to oxygenated and de-oxygenated hemoglobin and methemoglobin, which have strong optical absorption bands in the visible spectral range, and to differences in the light scattering properties. DRS was also able to identify a strongly calcified clot and could scriminate between fibrin/platelet-rich and RBC-rich parts of individual clots (Fig. 6).
Fig. 6. Presentation of a clot case Gross photograph of case number 8 and a millimeter marked ruler with annotated DRS measurement positions (a). The clot had a macroscopic heterogenous appearance, with a distal white and a proximal red portion. In Fig. 6(b), the result of the histological classification is presented, where RBC are coded yellow, fibrin red, platelets green and WBC blue. The DRS measurement positions 1 and 2 were located in the RBC-rich proximal part, and positions 3–5 in the fibrin-rich distal part. Figure 6(c) shows the DRS spectra from each of the annotated measurement positions, and 6d the averaged spectra from each clot unit (distal vs proximal). The spectra from the RBC-rich locations 1 and 2 had strong hemoglobin absorption around 760 nm and methemoglobin absorption around 630 nm and differed significantly from those at the fibrin-rich locations 3, 4, and 5.
The DRS technique has previously been successfully applied in an intravascular catheter for in vivo clot composition characterization in a porcine stroke model [34], and the results of this ex vivo study on human clots indicate that DRS has the potential to detect clot composition also in a clinical setting.
The composition of the clot is a determinant for thrombectomy success in AIS patients [20–22,25], and clot RBC content has been identified as a key feature. Clots retrieved from successful recanalization exhibit a higher RBC content than those retrieved from attempts in which full recanalization could not be achieved [23,25]. RBC-rich clots are associated with better patient outcomes, shorter puncture-to-recanalization time, fewer recanalization maneuvers [24] and shorter procedure times [26]. A biomechanical explanation for this phenomenon has been proposed. In addition to being firmer and more difficult to catch with a stent retriever, clots with a low RBC content have a significantly higher static friction against the vessel wall than RBC rich clots. Friction increases dramatically at RBC content below 20 percent [20]. These findings have led to the development of specific catheters and endovascular techniques targeted towards clots with low RBC and high fibrin conten t [29]. However, to tailor the treatment to a specific clot type, an accurate pre-procedural method to assess clot composition is needed. Currently, no such method is in clinical use.
To date, the available methods for characterization of clot composition in AIS patients have been limited to conventional neuroradiology. Indirect signs such as hyperdense middle cerebral artery sign (HMCAS) on non-contrast CT and blooming artifact (BA)/susceptibility vessel sign on gradient-echo MRI have been correlated to RBC-rich clots [43,44] and successful revascularization [15,27,45]. In vitro studies indicate that dual-energy CT may be used for clot composition differentiation [46]. However, clot heterogeneity and secondary clot formation due to stasis proximal and distal to the occlusion have been identified as factors that impede the accuracy of indirect methods [43]. To increase reliability in clot composition characterization, intravascular imaging methods have been developed. Spectral Optical Coherence Tomography (OCT) [30], the closely related Optical Coherence Elastography (OCE) and ultrasound Shear Wave Elastography (SWE) [31] have been studied for determination of mechanical properties and RBC content of clot analogues in vitro. Even though these results are promising, the techniques struggle with difficulties of miniaturization of the imaging probes, poor probe flexibility and lack of distal rotational control. Further development is needed before they can overcome the specific challenges posed by the often tortuous neurovasculature [47]. Electrochemical impedance spectroscopy (EIS), which uses an impedance sensor for measurement of the electrophysiological characteristics of clots, has also been suggested as a possible technique, and a first clinical trial on human subjects has been initiated [32,33]. The DRS technique presented in this paper can easily be applied through a thin micro guide wire-like probe and has been proven successful in an in vivo stroke model [34]. Other benefits with DRS include that there is no need to flush the clot area free from blood before the test, as opposed to OCT, and that the properties of the spectral data allow for development of algorithms for automated interpretation [30].
In a previous study, we showed that DRS applied through a micro guidewire-like probe was able to discriminate between RBC-rich, fibrin-rich and mixed type clot analogs [34]. However, the clot analogs used in that study were created in a laboratory and homogenous in nature. The current study confirms that DRS can also be used to identify the composition of real AIS blood clots, with a heterogenous and more complex structure [19]. With further development of the DRS micro guidewire-like probe, used in the previous in vivo study, an investigational device for clinical trials may be developed to examine the potential role of DRS in stroke treatment. In such a device it could be considered to only include the visible range of the DRS spectra, since the near infrared spectra did not significantly contribute to the discrimination between clot types in this proof-of-concept study. Moreover, in this study a clot type classification based on previous work by Maekawa et al. [24] was used, to put the results in a clinical context. In a future clinical application, other types of classifications could be considered, including dividing clots into three categories (RBC-rich, mixed and fibrin-rich), or focusing on identifying clots with the lowest RBC content (e.g. RBC < 20%), since such clots are more difficult to grab and produce higher friction against the vessel wall [20] whereby they may benefit from treatment with specific devices [29].
5. Limitations
The mechanical trauma of EVT may deform and fragment clots, and clot material may be left in situ. Hence, an extracted clot may not fully represent the characteristics of the original occluding clot in the cerebral vessel. Minor changes of clot integrity may occur during the time between clot retrieval and DRS measurements [19]. However, considering the short time intervals presented in this study, significant changes are not expected. Since clot composition is heterogeneous and an individual DRS measurement only measures the composition within a limited volume of the clot, there was a considerable variation between the recorded spectra within the individual clots. Due to clot heterogeneity and loss of clot integrity during the histological preparation, matching of individual DRS measurements with a specific location of the clot were not considered reliable in most cases. Instead, DRS measurements and clot histology were averaged over the smallest identifiable clot unit. This limits the granularity of the data, but on the other hand better corresponds to the actual clinical scenario where a whole clot needs to be treated regardless of clot heterogeneity. Also, in a future clinical setting, this could be compensated for by performing continuous measurements along the full axis of a clot, to get a better characterization of the clot composition.
The fit algorithm used to derive clot constituent concentrations is based on diffusion theory and has not been validated for fiber distances of 1 mm or less. The algorithm has however proven able to accurately determine relative absorbers concentrations for classification purposes [48].
This was an ex vivo, proof-of-concept study on DRS as a suitable technique for clot composition characterization in AIS. However, the sample size was limited and further studies with larger sample sizes, as well as clinical trials, are warranted to obtain a more robust estimate of the accuracy of the technique.
6. Conclusions
DRS can be used for clot composition characterization on clots extracted from AIS patients. Further clinical studies with DRS applied in an endovascular device are warranted, to evaluate the technique’s potential role in the management of acute ischemic stroke.
Acknowledgements
We acknowledge study nurse Åke Holmberg for administrative help and study coordination.
Disclosures
T.A. holds equity for Ceroflo and is a consultant for Amnis Therapeutics, Anaconda, Cerenovus-Neuravi, Medtronic, Rapid Medical and Stryker. None of the authors who are affiliated with clinical institutions or universities (S.S, G.B, E.E, O.A, A.E.-T, F.A, T.A.) have financial interests in the subject matter, materials, or equipment or with any competing materials and did not receive any payments from Philips. Karolinska University hospital and Philips have a major collaboration agreement. The authors affiliated with Philips Research (M.M, G.L.) have financial interests in the subject matter, materials, and equipment, in the sense that they are employees of Philips. Philips provided support in the form of salaries but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. Authors without conflicts of interest had full control of all data labelling, data analysis and information submitted for publication and over all conclusions drawn in the manuscript.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. O. A. Berkhemer, P. S. Fransen, D. Beumer, L. A. Van Den Berg, H. F. Lingsma, A. J. Yoo, W. J. Schonewille, J. A. Vos, P. J. Nederkoorn, and M. J. Wermer, “A randomized trial of intraarterial treatment for acute ischemic stroke,” N Engl J Med 372(1), 11–20 (2015). [CrossRef]
2. S. Bracard, X. Ducrocq, J. L. Mas, M. Soudant, C. Oppenheim, T. Moulin, and F. Guillemin, “Mechanical thrombectomy after intravenous alteplase versus alteplase alone after stroke (THRACE): a randomised controlled trial,” The Lancet Neurology 15(11), 1138–1147 (2016). [CrossRef]
3. J. Broderick, Y. Palesch, A. Demchuk, S. Yeatts, P. Khatri, M. Hill, E. Jauch, T. Jovin, B. Yan, and F. Silver, Interventional Management of Stroke (IMS) III Investigators, “Endovascular therapy after intravenous t-PA versus t-PA alone for stroke,” N Engl J Med 368(10), 893–903 (2013). [CrossRef]
4. B. C. Campbell, P. J. Mitchell, T. J. Kleinig, H. M. Dewey, L. Churilov, N. Yassi, B. Yan, R. J. Dowling, M. W. Parsons, and T. J. Oxley, “Endovascular therapy for ischemic stroke with perfusion-imaging selection,” N. Engl. J. Med. 372(11), 1009–1018 (2015). [CrossRef]
5. M. Goyal, A. M. Demchuk, B. K. Menon, M. Eesa, J. L. Rempel, J. Thornton, D. Roy, T. G. Jovin, R. A. Willinsky, and B. L. Sapkota, “Randomized assessment of rapid endovascular treatment of ischemic stroke,” N. Engl. J. Med. 372(11), 1019–1030 (2015). [CrossRef]
6. M. Goyal, B. K. Menon, W. H. van Zwam, D. W. Dippel, P. J. Mitchell, A. M. Demchuk, A. Dávalos, C. B. Majoie, A. van der Lugt, and M. A. De Miquel, “Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials,” The Lancet 387(10029), 1723–1731 (2016). [CrossRef]
7. T. G. Jovin, A. Chamorro, E. Cobo, M. A. de Miquel, C. A. Molina, A. Rovira, L. San Román, J. Serena, S. Abilleira, and M. Ribó, “Thrombectomy within 8 hours after symptom onset in ischemic stroke,” N. Engl. J. Med. 372(24), 2296–2306 (2015). [CrossRef]
8. C. S. Kidwell, R. Jahan, J. Gornbein, J. R. Alger, V. Nenov, Z. Ajani, L. Feng, B. C. Meyer, S. Olson, and L. H. Schwamm, “A trial of imaging selection and endovascular treatment for ischemic stroke,” N. Engl. J. Med. 368(10), 914–923 (2013). [CrossRef]
9. J. L. Saver, M. Goyal, A. Bonafe, H.-C. Diener, E. I. Levy, V. M. Pereira, G. W. Albers, C. Cognard, D. J. Cohen, and W. Hacke, “Stent-retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke,” N. Engl. J. Med. 372(24), 2285–2295 (2015). [CrossRef]
10. J. L. Saver, R. Jahan, E. I. Levy, T. G. Jovin, B. Baxter, R. G. Nogueira, W. Clark, R. Budzik, and O. O. Zaidat, “Solitaire flow restoration device versus the Merci Retriever in patients with acute ischaemic stroke (SWIFT): a randomised, parallel-group, non-inferiority trial,” The Lancet 380(9849), 1241–1249 (2012). [CrossRef]
11. C. Maegerlein, M. T. Berndt, S. Mönch, K. Kreiser, T. Boeckh-Behrens, M. Lehm, S. Wunderlich, C. Zimmer, and B. Friedrich, “Further development of combined techniques using stent retrievers, aspiration catheters and BGC,” Clinical neuroradiology 30(1), 59–65 (2020). [CrossRef]
12. J. M. Ospel, R. McTaggart, N. Kashani, M. Psychogios, M. Almekhlafi, and M. Goyal, “Evolution of Stroke Thrombectomy Techniques to Optimize First-Pass Complete Reperfusion,” in Seminars in Interventional Radiology(Thieme Medical Publishers2020), pp. 119–131.
13. O. O. Zaidat, A. C. Castonguay, I. Linfante, R. Gupta, C. O. Martin, W. E. Holloway, N. Mueller-Kronast, J. D. English, G. Dabus, and T. W. Malisch, “First pass effect: a new measure for stroke thrombectomy devices,” Stroke 49(3), 660–666 (2018). [CrossRef]
14. O. Nikoubashman, S. Dekeyzer, A. Riabikin, A. Keulers, A. Reich, A. Mpotsaris, and M. Wiesmann, “True first-pass effect: first-pass complete reperfusion improves clinical outcome in thrombectomy stroke patients,” Stroke 50(8), 2140–2146 (2019). [CrossRef]
15. A. J. Yoo and T. Andersson, “Thrombectomy in acute ischemic stroke: challenges to procedural success,” J Stroke 19(2), 121–130 (2017). [CrossRef]
16. F. Denorme, F. Langhauser, L. Desender, A. Vandenbulcke, H. Rottensteiner, B. Plaimauer, O. François, T. Andersson, H. Deckmyn, and F. Scheiflinger, “ADAMTS13-mediated thrombolysis of t-PA–resistant occlusions in ischemic stroke in mice,” Blood The Journal of the American Society of Hematology 127, 2337–2345 (2016). [CrossRef]
17. E. Laridan, F. Denorme, L. Desender, O. François, T. Andersson, H. Deckmyn, K. Vanhoorelbeke, and S. F. De Meyer, “Neutrophil extracellular traps in ischemic stroke thrombi,” Ann Neurol. 82(2), 223–232 (2017). [CrossRef]
18. S. Staessens and S. F. De Meyer, “Thrombus heterogeneity in ischemic stroke,” Platelets 32(3), 331–339 (2021). [CrossRef]
19. S. Staessens, F. Denorme, O. François, L. Desender, T. Dewaele, P. Vanacker, H. Deckmyn, K. Vanhoorelbeke, T. Andersson, and S. F. De Meyer, “Structural analysis of ischemic stroke thrombi: histological indications for therapy resistance,” Haematologica 105(2), 498–507 (2020). [CrossRef]
20. G. M. Gunning, K. McArdle, M. Mirza, S. Duffy, M. Gilvarry, and P. A. Brouwer, “Clot friction variation with fibrin content; implications for resistance to thrombectomy,” Journal of neurointerventional surgery 10(1), 34–38 (2018). [CrossRef]
21. W. Brinjikji, S. Duffy, A. Burrows, W. Hacke, D. Liebeskind, C. B. Majoie, D. W. Dippel, A. H. Siddiqui, P. Khatri, and B. Baxter, “Correlation of imaging and histopathology of thrombi in acute ischemic stroke with etiology and outcome: a systematic review,” Journal of NeuroInterventional Surgery 9(6), 529–534 (2017). [CrossRef]
22. B. Cline, J. Vos, J. Carpenter, and A. Rai, “O-027 Pathological analysis of extracted clots in embolectomy patients with acute ischaemic stroke,” Journal of NeuroInterventional Surgery 5(Suppl 2), A15.2–A16 (2013). [CrossRef]
23. T. Hashimoto, M. Hayakawa, N. Funatsu, H. Yamagami, T. Satow, J. C. Takahashi, K. Nagatsuka, H. Ishibashi-Ueda, J.-I. Kira, and K. Toyoda, “Histopathologic analysis of retrieved thrombi associated with successful reperfusion after acute stroke thrombectomy,” Stroke 47(12), 3035–3037 (2016). [CrossRef]
24. K. Maekawa, M. Shibata, H. Nakajima, A. Mizutani, Y. Kitano, M. Seguchi, M. Yamasaki, K. Kobayashi, T. Sano, and G. Mori, “Erythrocyte-rich thrombus is associated with reduced number of maneuvers and procedure time in patients with acute ischemic stroke undergoing mechanical thrombectomy,” Cerebrovascular Diseases Extra 8(1), 39–49 (2018). [CrossRef]
25. J. W. Shin, H. S. Jeong, H.-J. Kwon, K. S. Song, and J. Kim, “High red blood cell composition in clots is associated with successful recanalization during intra-arterial thrombectomy,” PLoS ONE 13(5), e0197492 (2018). [CrossRef]
26. P. B. Sporns, U. Hanning, W. Schwindt, A. Velasco, B. Buerke, C. Cnyrim, J. Minnerup, W. Heindel, A. Jeibmann, and T. Niederstadt, “Ischemic stroke: histological thrombus composition and pre-interventional CT attenuation are associated with intervention time and rate of secondary embolism,” Cerebrovasc. Dis. 44(5-6), 344–350 (2017). [CrossRef]
27. R. Bourcier, M. Mazighi, J. Labreuche, R. Fahed, R. Blanc, B. Gory, A. Duhamel, G. Marnat, S. Saleme, and V. Costalat, “Susceptibility vessel sign in the ASTER trial: higher recanalization rate and more favourable clinical outcome after first line stent retriever compared to contact aspiration,” Journal of stroke 20(2), 268–276 (2018). [CrossRef]
28. J. Madjidyar, L. P. Vidal, N. Larsen, and O. Jansen, “Influence of thrombus composition on thrombectomy: ADAPT vs. balloon guide catheter and stent retriever in a flow model,” in Influence of thrombus composition on thrombectomy: ADAPT vs. balloon guide catheter and stent retriever in a flow model. RöFo-Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren(Georg Thieme Verlag KG, 2020), pp. 257–263.
29. V. S. Fennell, S. V. S. Nagesh, K. M. Meess, L. Gutierrez, R. H. James, M. E. Springer, and A. H. Siddiqui, “What to do about fibrin rich ‘tough clots’? Comparing the Solitaire stent retriever with a novel geometric clot extractor in an in vitro stroke model,” Journal of neurointerventional surgery 10(9), 907–910 (2018). [CrossRef]
30. Y. Ding, M. Abbasi, A. M. Eltanahy, D. R. Jakaitis, D. Dai, R. Kadirvel, D. F. Kallmes, and W. Brinjikji, “Assessment of blood clot composition by spectral optical coherence tomography: an in-vitro study,” Neurointervention 16(1), 29–33 (2021). [CrossRef]
31. H.-C. Liu, M. Abbasi, Y. H. Ding, T. Roy, M. Capriotti, Y. Liu, S. Fitzgerald, K. M. Doyle, M. Guddati, and M. W. Urban, “Characterizing blood clots using acoustic radiation force optical coherence elastography and ultrasound shear wave elastography,” Phys. Med. Biol. 66(3), 035013 (2021). [CrossRef]
32. Y. Xu, X. Xie, Y. Duan, L. Wang, Z. Cheng, and J. Cheng, “A review of impedance measurements of whole cells,” Biosens. Bioelectron. 77, 824–836 (2016). [CrossRef]
33. . “Clotild® Smart Guidewire System Evaluation in Endovascular Thrombectomy Procedure, ClinicalTrials.gov, Identifier: NCT04993079, Retrieved from https://clinicaltrials.gov/ct2/show/study/NCT04993079 19 September 2021,” (2021).
34. S. Skyrman, G. Burström, O. Aspegren, G. Lucassen, A. Elmi-Terander, E. Edström, F. Arnberg, M. Ohlsson, M. Mueller, and T. Andersson, “Identifying clot composition using intravascular diffuse reflectance spectroscopy in a porcine model of endovascular thrombectomy,” Journal of NeuroInterventional Surgery 14(3), 304–309 (2022). [CrossRef]
35. R. Nachabé, B. H. Hendriks, M. van der Voort, A. E. Desjardins, and H. J. Sterenborg, “Estimation of biological chromophores using diffuse optical spectroscopy: benefit of extending the UV-VIS wavelength range to include 1000 to 1600 nm,” Biomed. Opt. Express 1(5), 1432–1442 (2010). [CrossRef]
36. T. M. Bydlon, R. Nachabé, N. Ramanujam, H. J. Sterenborg, and B. H. Hendriks, “Chromophore based analyses of steady-state diffuse reflectance spectroscopy: current status and perspectives for clinical adoption,” J. Biophotonics 8(1-2), 9–24 (2015). [CrossRef]
37. E. Tanis, J. W. Spliethoff, D. J. Evers, G. C. Langhout, P. Snaebjornsson, W. Prevoo, B. H. Hendriks, T. J. J. E. J. o, and S. O. Ruers, “Real-time in vivo assessment of radiofrequency ablation of human colorectal liver metastases using diffuse reflectance spectroscopy,” 42, 251–259 (2016).
38. P. Bankhead, M. B. Loughrey, J. A. Fernández, Y. Dombrowski, D. G. McArt, P. D. Dunne, S. McQuaid, R. T. Gray, L. J. Murray, and H. G. Coleman, “QuPath: Open source software for digital pathology image analysis,” Sci. Rep. 7(1), 16878–7 (2017). [CrossRef]
39. N. Rajaram, A. Gopal, X. Zhang, and J. W. Tunnell, “Experimental validation of the effects of microvasculature pigment packaging on in vivo diffuse reflectance spectroscopy,” Lasers Surg. Med. 42(7), 680–688 (2010). [CrossRef]
40. D. W. Hahn, “Light Scattering Theory,” Department of Mechanical and Aerospace Engineering, University of Florida (2009).
41. W. Zijlstra and A. Buursma, “Spectrophotometry of hemoglobin: absorption spectra of bovine oxyhemoglobin, deoxyhemoglobin, carboxyhemoglobin, and methemoglobin,” Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 118(4), 743–749 (1997). [CrossRef]
42. R. Nachabé, B. H. Hendriks, G. W. Lucassen, M. van der Voort, D. J. Evers, E. J. Rutgers, M.-J. V. Peeters, J. A. Van der Hage, H. S. Oldenburg, and T. J. Ruers, “Diagnosis of breast cancer using diffuse optical spectroscopy from 500 to 1600 nm: comparison of classification methods,” J. Biomed. Opt. 16(8), 087010 (2011). [CrossRef]
43. D. S. Liebeskind, N. Sanossian, W. H. Yong, S. Starkman, M. P. Tsang, A. L. Moya, D. D. Zheng, A. M. Abolian, D. Kim, and L. K. Ali, “CT and MRI early vessel signs reflect clot composition in acute stroke,” Stroke 42(5), 1237–1243 (2011). [CrossRef]
44. S. T. Fitzgerald, S. Wang, D. Dai, A. Douglas, R. Kadirvel, M. J. Gounis, J. Chueh, A. S. Puri, K. F. Layton, and I. C. Thacker, “Platelet-rich clots as identified by Martius Scarlet Blue staining are isodense on NCCT,” Journal of Neurointerventional Surgery 11(11), 1145–1149 (2019). [CrossRef]
45. S. F. De Meyer, T. Andersson, B. Baxter, M. Bendszus, P. Brouwer, W. Brinjikji, B. C. Campbell, V. Costalat, A. Dávalos, and A. Demchuk, “Analyses of thrombi in acute ischemic stroke: a consensus statement on current knowledge and future directions,” International Journal of Stroke 12(6), 606–614 (2017). [CrossRef]
46. J. Borggrefe, J. Kottlors, M. Mirza, V.-F. Neuhaus, N. Abdullayev, V. Maus, C. Kabbasch, D. Maintz, and A. Mpotsaris, “Differentiation of clot composition using conventional and dual-energy computed tomography,” Clinical Neuroradiology 28(4), 515–522 (2018). [CrossRef]
47. M. J. Gounis and D. A. Steinman, “Up around the bend: progress and promise of intravascular imaging in neurointerventional surgery,” Journal of NeuroInterventional Surgery 13(6), 495–496 (2021). [CrossRef]
48. B. H. W. Hendricks, A. J. Balthasar, G. W. Lucassen, M. van der Voort, M. Mueller, V. V. Pully, T. M. Bydlon, C. Reich, T. M. H. van Keersop, J. Kortsmit, J. Langout, and G.-J. van Geffen, “Nerve detection with optical spectroscopy for regional anesthesia procedures,” J. Transl. Med. 13(1), 1–15 (2015). [CrossRef]
