Successful recanalization increases the chance of favorable outcome 4-fold compared to patients without recanalization and reduces mortality rate 4-fold . The importance of recanalization is even more pronounced in basilar artery occlusion, where the chance of an independent life is extremely low in patients without recanalization . Therefore, the current treatment options for acute stroke aim at fast and effective flow restoration to the cerebral tissue. Using intravenous thrombolysis with recombinant tissue plasminogen activator (rtPA), the rate of recanalization in severe stroke due to large artery occlusion (LAO) is low and depends on the length of the thrombus, with nearly no potential to recanalize occluded vessels if thrombus length exceeds 8 mm . In contrast, modern image-guided, endovascular mechanical recanalization, based on advanced neuroimaging, can achieve rapid recanalization in 80–90% of cases, independent from the length and localization of the thrombus [4, 5] with a number needed to treat of less than three patients for improved clinical outcome .
This chapter summarizes current aspects of cerebral digital subtraction angiography (DSA) in the context of endovascular stroke treatment and the working principles of techniques and instruments for endovascular catheter-based treatment, including local, intra-arterial thrombolysis with different thrombolytic agents and various mechanical recanalization techniques in complex stroke cases. To avoid redundancy, this chapter refers to the previous chapter, where the design and outcome of the recent prospective randomized trials are described, which confirmed the concept of image-guided, endovasvcular microsurgical treatment of acute stroke.
Cerebral Digital Subtraction Angiography
Almost 100 years after the introduction of angiography by the Portuguese neurologist Edgaz Moniz , catheter angiography is still considered the gold standard for imaging of the cerebral vasculature. Indications include diagnostic imaging of primary neurovascular diseases, such as intracranial aneurysms, arteriovenous malformations (AVM), dural arteriovenous fistulas (dAVF), atherosclerotic stenosis, vasculopathy (e.g. vasculitis, vasospasm), and acute ischemic stroke, as well as imaging follow-up after treatment (e.g. after aneurysm coiling/clipping, treatment of AVMs or dAVF). Diagnostic angiography is performed as the first step prior to endovascular procedures. Vascular access is usually gained by femoral puncture, alternatively by radial, brachial, axillary, or rarely by direct carotid or aortic puncture. Endovascular catheters of various sizes and shapes depending on the vascular anatomy and target vessel are navigated under fluoroscopic control within the arterial system using steerable wires. Imaging acquisition is based on the use of X-ray technology and intra-arterial application of iodinated contrast material. The established technical principle for image acquisition is DSA. Using DSA, a plain X-ray image is acquired as a so-called “mask image,” which is then after contrast injection subtracted from the subsequently acquired contrasted X-ray images. The resulting run of subtracted images consequently depicts only the vasculature without the surrounding bony or soft-tissue anatomy. The advantage of DSA is the combination of high spatial resolution of up to 100–150 μm and the high temporal resolution of up to 7.5 frames per second, which has not been reached yet by other non-invasive imaging modalities. Disadvantages are mainly due to the invasiveness of the procedure. Neurologic complications due to thromboembolism from catheters and wires, disruption of atherosclerotic plaques, and dissections overall occur in around 1.3% of cases, where approximately 0.9% account for transient or reversible and 0.5% for permanent neurological deficits . Less common neurological complications include transient cortical blindness and amnesia. Non-neurological complications of cerebral angiography include hematoma and pseudoaneurysm formation at the puncture site, including retroperitoneal hematoma, allergic reactions, and nephropathy due to the application of iodinated contrast media and pulmonary thromboembolism or thromboembolism to the lower extremity. The rate of complications is strongly related to the underlying pathology and patient condition, such as atherosclerotic disease, recent cerebral ischemic event, advanced age, procedure time, hypertension, diabetes, and renal insufficiency. DSA in the context of endovascular acute stroke treatment is able to differentiate between occlusion and pseudo-occlusion of a brain-supplying vessel (e.g. ICA) and to visualize exactly and in detail the occlusion site and access vessels for planning of the interventional procedure. Furthermore, the characterization and dynamic evaluation of leptomeningeal collateral circulation has been shown to be a crucial imaging information in regard to clinical outcome . Therefore, a complete diagnostic cerebral angiography including the right and left common carotid artery and the dominant vertebral artery territories is essential prior to endovascular treatment approaches and should be performed in less than 10 minutes.
Technical developments of modern angiography systems, especially so-called flat panel detector (FPD) technology, have further improved diagnostic accuracy. FPD-mounted C-arm angiography systems have been widely introduced into neurinterventional suites. These detector systems permit the acquisition of high-quality three-dimensional vascular imaging, so-called 3D rotational angiography. Post-processing of such a data set allows detailed three-dimensional reconstruction of the intracranial vessel anatomy, which is especially helpful for the analysis, treatment planning, peri-interventional monitoring, and follow-up imaging of complex intracranial vascular anatomy, such as in endovascular aneurysm treatment. Furthermore, FPD imaging allows to obtain CT-like cross-sectional soft-tissue imaging of the brain parenchyma. Applications include assessment of the extent of subarachnoid or intracerebral hemorrhage and the width of the ventricles, as well as immediate control of coil and intra- and extracranial stent placement. FPD technology permits fast imaging without the need for time-consuming patient transfer to a CT facility and, therefore, has become a helpful tool for immediate “on the table” evaluation of treatment results and intra-procedural complications. Recently, the possibility of performing whole-brain parenchymal cerebral blood volume (CBV) measurements by means of intravenous contrast media injection have been added to the spectrum of FPD imaging [10, 11]. We use FPD technology at the moment mainly for stroke patients being tranfered with or without bridging i.v.-thrombolysis to our stroke center for mechanical thrombectomy and a long period of time between the initial imaging in the referring stroke unit, especially if the clinical symptoms changed or if the acute neurological deficits are not compatible with the initial images, which are available in the angio suite via teleradiology (Figure 20.1). Current promising developments aim at further extending this technology to dynamic perfusion measurements such as cerebral blood flow (CBF) and further improving the quality of CT-like imaging of the brain parenchyma . Further improvements of FPD CT, FPD angiography, and FPD perfusion imaging will enable one-stop diagnostic and endovascular treatment in the angio suite, at least for the majority of patients who suffer from unambiguous symptoms of acute stroke caused by large-vessel occlusion of the anterior circulation .
Figure 20.1 Flat panel detector-computed tomography (FPD-CT). A 64-year-old woman with improving symptoms of acute stroke and the diagnosis of tandem occlusion in a stroke unit 70 km away from our stroke center. Deterioration following bridging i.v. thrombolysis and transfer to our stroke center, where she arrived 2 hours later with NIHSS of 20 in the angio suite.
A: Pre-interventional FDP-CT excludes an intracerebral hemorrhage; angio-FDP (not shown) confirms persistent tandem occlusion.
B: FDP-perfusion measurement confirms reduced left hemispheric CBV (cerebral blood volume) with reactive hyperperfusion immediately after recanalization of the tandem ICA-M1 occlusion (C).
Digital substraction angiography is the gold standard for high-resolution imaging of the cerebral vasculature with a spatial resolution of 100 µm and enables minimal invasive, 3D-image-guided endovascular microneurosurgery. Recent developments of FPD technology generate CT-like soft-tissue and quantitative perfusion images of the brain and may enable one-stop diagnostic and endovascular treatment of acute ischemic stroke in the modern angio suite.
Intra-Arterial Pharmacological Thrombolysis
Intravenous administration of thrombolytic drugs has been evaluated in several randomized stroke trials using recombinant tissue-type plasminogen activator (rtPA) demonstrating efficiency up to 3–4.5 hours after symptom onset. However, studies have shown a limited effect in patients with severe stroke and proximal vessel occlusion, such as M1 or ICA, with a high thrombus burden compared to more peripherally located occlusions [3, 13]. In addition, due to the limited time window only a minority of patients admitted to stroke centers are eligible for intravenous stroke treatment. There are in addition many contraindications for intravenous thrombolysis, e.g. pregnancy, coagulopathy, thrombocytopenia, septicemia, recent trauma, surgery, and gastrointestinal or intracranial hemorrhage, which could be recanalized with local intra-arterial thrombolysis, or, even better, with mechanical thrombectomy. The advantages of intra-arterial application of thrombolytic drugs at the occlusion site is the delivery of a higher effective concentration of the thrombolytic agent directly to the thrombus. This increases the chance of clot dissolution, especially in occlusions with larger thrombus burden, resulting in higher recanalization rates. Furthermore, reduction of systemic exposure to the thrombolytic drug potentially reduces systemic side-effects and the risk of spontaneous intracerebral hemorrhage expanding the time window for treatment. Technically, for intra-arterial thrombolysis a microcatheter is navigated over a microwire at the occlusion site and placed proximally to  or at the site of the occlusion .
The landmark study for intra-arterial pharmacological thrombolysis is the Prolyse in Acute Cerebral Thrombembolysm II (PROACT II) . One hundred and eighty patients with angiographically confirmed M1 and M2 occlusions were randomized within 6 hours after symptom onset to be treated with 9 mg intra-arterial prourokinase and heparin or heparin only (control group). According to the study protocol prourokinase had to be administered in front of the proximal surface of the thrombus. Any mechanical manipulation of the thrombus, e.g. probing with the microwire, was prohibited. Favorable clinical outcome (mRS ≤2) was achieved in 40% of the endovascularly treated patients and in 25% of the control group with a mortality rate of 25% and 27%, respectively. Recanalization rate defined as TIMI 2–3 was significantly higher in the prourokinase group (66%) than in the control group (18%). SICH was lower in the control group (2%) compared to the patients treated with prourokinase (10%). The PROACT study was able to demonstrate that despite the higher incidence of SICH, intra-arterial application of prourokinase within 6 hours after symptom onset improves significantly clinical outcome in patients with M1 and M2 occlusions. The Japanese MELT study (Middle cerebral artery Embolism Local fibrinolytic intervention Trial) increased the rate of recanalization to 74% by injection of urokinase in or distally to the thrombus . However, despite the positive results of PROACT II and MELT, prourokinase and urokinase have not been approved by the US Food and Drug Administration (FDA). Nevertheless, local intra-arterial thrombolysis is still an option, if navigation of the stent retriever or aspiration catheter to or beyond the site of occlusion may be difficult or risky.
Local intra-arterial thrombolysis has the advantage of delivering the thrombolytic agent directly to the thrombus while reducing the systemic exposure. This technique is nowadays rarely used if the approach to the thrombus is difficult or risky with instruments for thrombectomy and aspiration.
Although pharmacological intravenous and intra-arterial thrombolysis have been shown to be effective in some patients, the limited recanalization rate and time window for treatment with an increasing risk for intracranial hemorrhage over time leaves room for further improvement. Mechanical recanalization approaches aim at fast blood flow restoration by rapid direct thrombus removal, especially in case of large thrombotic burden. Therefore, various endovascular mechanical treatment approaches have been advocated to overcome the limitations of pharmacological thrombolysis, i.e. to increase recanalization rate and accelerate time-to-recanalization. Furthermore, reduction or even waiving of thrombolytic drugs may reduce the risk of intracranial hemorrhage and may further enhance the time window of opportunity for recanalization and re-establishing of cerebral blood flow. Recent endovascular mechanical treatment approaches consist mainly of stent retriever thrombectomy and thromboaspiration.
Different techniques for thrombus disruption and fragmentation have been reported. The most widely used is probing the thrombus with the microwire and/or advancing the microcatheter into or distally to the site of occlusion. This simple mechanical maneuver has been shown to be able to improve the rate of successful recanalization. Using this technique as an adjunct to intra-arterial thrombolysis successful recanalization of M1 occlusions could be achieved in 74% (MELT trial ) and 79% of cases  compared to 66% in the PROACT II trial , where mechanical clot disruption was not allowed. Using local injection of urokinase into the thrombus in addition to cautious mechanical clot disruption and thromboaspiration, good clinical outcome in patients with hyperdense middle cerebral artery sign (HMCAS) could be achieved in 53% compared to 23% in a control group treated with i.v. thrombolysis only, even though the time window for endovascular treatment has been 6 hours compared to 3 for i.v. thrombolysis .
Another method of mechanical clot disruption is the use of percutaneous transluminal balloon angioplasty (PTA). The largest study compared 34 patients with M1 occlusions undergoing PTA as the first-line treatment to 36 patients treated with intra-arterial thrombolysis alone . Sufficient recanalization was achieved in 91.2% in the PTA group as compared to 63.9% in the thrombolysis group, with a favorable clinical outcome (mRS at 90 days of ≤2) in 73.5% versus 50%, respectively. However, 64.5% of patients required additional intra-arterial thrombolysis after PTA due to distal emboli occurring during the procedure. Therefore, due to the risk of procedural complications and distal vessel occlusion caused by downstream embolization of thrombus fragments associated with PTA, this technique has not become accepted as a first-line treatment.
More elaborated thrombus disruption devices apply ultrasound or laser technology. The EKOS system (EKOS, Bothell, USA) consists of a 2.5 F microcatheter (MicroLysUS infusion catheter) with a 2.1 MHz piezoelectric sonography probe at its distal tip. Ultrasonic vibration applied at the thrombus is supposed to increase fluid penetration within the clot in order to ease and augment the effect of intra-arterial thrombolysis. Successful recanalization was reported in the initial clinical study of 14 patients in 57% with a favorable clinical outcome in 43% . However, a 14% rate of SICH was observed. The EPAR system (EPAR, Endovasix, Belmont, USA) applies laser technology to emulsify the thrombus by application of microcavitation bubbles at the microcatheter tip. The initial clinical study  enrolled 34 patients in whom the device could only be used as intended in 18 patients (53%) with a successful recanalization rate of 61.1% in this subset of patients. One severe adverse event due to a vessel rupture, a 5.9% rate of SICH, and a mortality rate of 38.2% were reported.
Mechanical disruption of the thrombus leads to local spasms, damage of the vessel wall and fragmentation and iatrogenic, distal, downstream arterio-arterial embolization with change for the worse for the collateral supply of the penumbra.
Distal Thrombectomy Devices
Distal devices are advanced by guide wire and microcatheter to pass the thrombus at the occlusion site and are unsheathed distally to it. These devices apply force at the distal aspect of the thrombus and include snare-like, basket-like, corkscrew-like, and brush-like devices.
Distal devices are successful in removing the thrombus; however, application of force on the distal base of the thrombus results in thrombus compression, which increases friction during retrieval and increases the risk of distal thromboembolic events and side branch occlusion [21, 22]. Therefore, the use of proximal balloon occlusion and aspiration during retrieval of a distal device is recommended. Thrombectomy approaches have been advocated alone or in combination with adjuvant intravenous or intra-arterial thrombolysis.
Compared to proximal thrombectomy approaches, distal thrombectomy is technically more challenging because the occlusion site has to be passed with a microcatheter in order to deliver the device distally to the thrombus. Proximal balloon occlusion using a balloon guide catheter placed in the cervical segment of the internal cerebral artery and aspiration during device retrieval is recommended for most devices to avoid thromboembolic complications. Several distal thrombectomy devices have been introduced into clinical practice (e.g. Catch, BALT, Montmorency, France; Phenox pCR/CRC/Bonnet, Phenox GmbH, Bochum, Germany). Large clinical studies have been performed and published using the MERCI device (Concentric Medical, Mountain View, USA), which was the first distal thrombectomy device to receive FDA approval in 2004.
The MERCI retrieval system has been investigated in the MERCI trial (Mechanical Embolus Removal in Cerebral Ischemia) enrolling patients ineligible for intravenous treatment within 8 hours of symptom onset and Multi-MERCI trial enrolling patients eligible for intravenous and/or intra-arterial thrombolysis using a modified version of the device . Both trials showed recanalization rates of 46% and 69.5% with good clinical outcome in 27.7% and 36%, respectively. Complication rate was 7.1% and 5.5%, with a rate of SICH in 7.8% and 9.8%, respectively.
The MERCI and Multi-MERCI trials promoted the introduction of the MERCI device into wider clinical practice by providing clinical data from an early phase of the device introduction. Furthermore, both trials demonstrated a significant improvement in favorable clinical outcome in patients with recanalization compared to those without successful recanalization.
The IMS III  Synthesis  and MR Rescue trials  were the first randomized controlled trials comparing endovascular stroke treatment to standard intravenous rtPA administration. All three trials reported negative results for endovascular stroke treatment.
The IMS III trial  was a randomized clinical trial comparing a combined approach of intravenous thrombolysis followed by endovascular treatment within the 4.5-hour time window. Endovascular approaches approved were the MERCI device, the Penumbra device, and the EKOS Micro-Sonic SV infusion system. Only in the final phase of the study a stent retriever device, the Solitaire FR, was cleared. The IMS III trial has been suspended because of equipoise using mechanical approaches compared to intravenous thrombolysis (mRS ≤2: 40.8% endovascular, 38.7% intravenous treatment group). Mortality as well as SICH did not differ significantly (19.1 endovascular group vs. 21.6% intravenous treatment group and 6.2% vs. 5.9%, respectively). However, the study design has several flaws. First, during the relatively long inclusion period (2006–2012), diagnostic imaging modalities and endovascular techniques have significantly evolved. Especially the lack of multimodal diagnostic imaging including vessel imaging for demonstration of a major arterial occlusion and evaluation of the penumbra is a major weakness of the trial. Only 46% of patients underwent CT angiography, which was not used for inclusion, resulting in patients undergoing therapy without major arterial occlusion. Second, the approved endovascular techniques were heterogeneous and some already outdated. Only 1.5% of patients were treated using latest generation stent retrievers due to the late clearance for the study.
The Synthesis trial  was another randomized trial comparing standard intravenous rtPA treatment initiated within 4.5 hours after symptom onset to endovascular treatment within 6 hours after symptom onset. Patients undergoing endovascular treatment did not receive intravenous rtPA. Endovascular treatment included intra-arterial thrombolysis and guidewire thrombus fragmentation (66.1%) and endovascular devices (33.9%) (MERCI device and Penumbra device 8.5% and stent retrivers 13.9%). The primary endpoint defined as an mRS of 0–1 was similar in both groups (30.4% endovascular, 34.8% intravenous treatment group). There were no significant differences in the rate of SICH (6% each) and mortality (14.4% endovascular, 9.9% intravenous treatment group). The same criticism as for the IMS III trial also applies for the Synthesis trial, i.e. lack of appropriate diagnostic imaging protocols and use of outdated devices in the majority of endovascular treatment cases. Furthermore, patients with an NIHSS score as low as 2 were included, who are known to have a high chance of good outcome even without treatment.
The MR Rescue trial  was a third randomized trial enrolling patients undergoing mechanical thrombectomy (MERCI device or Penumbra device) or standard medical care within 8 hours of symptom onset. All patients underwent pretreatment diagnostic multimodal CT or MR imaging allowing stratification according to a favorable penumbral pattern (substantial savageable tissue and small infarct core) and non-penumbral pattern (large core or small or absent penumbra). Fifty-eight percent of patients showed a favorable penumbral pattern. Successful recanalization was achieved in 67% of patients undergoing endovascular treatment. The patients were classified into four groups: endovascular therapy/penumbral pattern; standard care/penumbral pattern; endovascular therapy/non-penumbral pattern; and standard care/non-penumbral pattern. Clinical outcome measured as mean mRS did not significantly differ between the groups (3.9 vs. 3.9, p = 0.99). Endovascular therapy was not superior to standard care in patients with either favorable penumbral pattern (3.9 vs. 3.4, p = 0.23) or non-penumbral pattern (4.0 vs. 4.4, p = 0.32). The percentage of patients with good outcome and mortality was not significantly different between the groups (endovascular therapy/penumbral pattern – 21%/18%; standard care/penumbral pattern – 26%/21%; endovascular therapy/non-penumbral pattern – 17%/20%; and standard care/non-penumbral pattern – 10%/30%). Therefore, a favorable penumbral pattern was unable to identify patients who would benefit from endovascular therapy and mechanical thrombectomy was not superior to standard medical care. In contrast to the IMS III and Synthesis trials, pretreatment diagnostic imaging was more sophisticated, using multimodal imaging to evaluate major vessel occlusion and the penumbra. However, the assignment to a penumbral pattern using a dedicated software was not easy to handle and only 58% of patients could be processed in real time with a shift of pattern assignment in 8% after final core laboratory post-processing. Furthermore, subgroups were rather small and therefore with limited statistical power. Finally, similar to the IMS III and Synthesis trials, the latest generation of thrombectomy devices were not used.
The first randomized controlled trials comparing endovascular stroke treatment to standard intravenous rtPA administration reported negative results for endovascular stroke treatment, mainly due to the use of inappropriate endovascular techniques and instruments with subsequent high rates of complications, especially iatrogenic distal embolization of the fragments of the thrombus.
The limitations of pharmacological thrombolysis and mechanical thrombectomy have led to a further pursuit of devices to increase the recanalization rate. Placement of a permanent intracranial stent achieves immediate flow restoration and recanalization by compressing the thrombus against the vessel wall (Figure 20.2). Stenting allows fast and effective recanalization without the need of repetitive passing of the occlusion site and retrieval attempts compared to thrombectomy devices. However, this straightforward concept has some disadvantages in general and especially in the setting of acute stroke treatment. Thrombus compression may lead to permanent side branch or perforator occlusion. Moreover, permanent stent placement needs double platelet anti-aggregation medication in order to prevent in-stent thrombosis and re-occlusion. This preventive medication may increase the risk of symptomatic intracranial hemorrhage in the setting of acute stroke. Furthermore, an in-stent restenosis rate of bare metal stents has been reported in up to 32% of cases in the treatment of intracranial arteriosclerotic stenosis after a follow-up period of 9 months (Sylvia trial) using the Neurolink system, whereas no significant restenosis has been described 6 months later using the wingspan system .
Figure 20.2 An 81-year-old patient presenting with aphasia and right hemiplegia (NIHSS 18). MR examination showing diffusion restriction in the left basal ganglia (A) and perfusion deficit (prolonged mean transit time; MTT) in the left MCA territory (B). DSA demonstrating the left M1 occlusion (C). Passing of the occlusion site with the microwire and microcatheter in the cranial aspect of the occluding thrombus (D). Control angiogram after deployment of a permanent stent (Wingspan 3 × 15 mm) showing immediate and complete recanalization of the M1 segment. Note preservation of the lateral lenticulo-striate arteries due to compression of the thrombus against the contralateral vessel wall (arrows in D and E). Schematic illustration of the principle of stent recanalization depicting passing of the occlusion site with the microcatheter (F) between the thrombus and the upper wall of the artery and compression of the thrombus against the contralateral vessel wall after stent deployment (G).
The use of different stent systems has been reported in case reports and small case series. In general, self-expandable stents are preferentially used over balloon-mounted stents. Recanalization rates are reported to be between 79% and 92%, with moderate clinical outcome in 33–50% . The Stent-Assisted Recanalization in Acute Ischemic Stroke (SARIS) trial is the first FDA-approved prospective trial investigating stenting in acute stroke treatment. Twenty patients (mean NIHSS 14) were included within 6 hours after symptom onset. Recanalization rate was 100% with adjuvant therapies such as angioplasty, intravenous tPA, and intra-arterial thrombolysis applied in 63% of patients. Moderate clinical outcome was achieved in 60% of patients . Despite the high recanalization rate reported in these studies, the use of intracranial stenting in acute stroke treatment is debatable due to the risks associated with permanent stent deployment and the recent success of thrombectomy. However, stenting has a clear value in selective cases of rescue therapy, where other recanalization methods have failed. Furthermore, permanent stent deployment has its role in the acute phase of stroke treatment in case of an underlying intracranial stenosis in order to achieve recanalization and to reduce the risk of re-occlusion due to insufficient flow across a stenosis.
Stent recanalization allows fast and effective recanalization, but has the disadvantage of the need for double platelet anti-aggregation medication in order to prevent in-stent thrombosis and re-occlusion. This may increase the risk of symptomatic intracranial hemorrhage in the setting of acute stroke. Furthermore, there is a risk of in-stent thrombosis or delayed thrombosis. Implantation of a permanent stent is still a rescue technique in special situations, e.g. vessel occlusions due to an underlying dissection or arteriosclerotic stenosis and inability to remove the thrombus en bloc.
Proximal Thrombectomy – Thromboaspiration
Percutaneous aspiration thrombectomy has been well known and established for decades in acute limb ischemia as a percutaneous, minimally invasive development of surgical recanalization using large bore Fogarty embolectomy catheters. It is rapid and effective in cases of embolic thrombi in normal arteries, but the success decreases in the presence of underlying chronic atherosclerotic changes of the artery. Considering that stroke is mainly caused by an acute thromboembolic occlusion of an otherwise normal cerebral artery, the concept of thromboaspiration in acute stroke is obvious and has been primarily utilized for occlusion of the vertebrobasilar  and carotid arteries .
Although large, 9F balloon-guiding catheters can be used, the success of remote aspiration in the cervical segment of the carotid artery is limited by collapse of the internal carotid artery (ICA) (Figure 20.3).
Figure 20.3 A 74-Year-old man with acute symptomatic ICA stenosis.
A: DSA shows high-grade ICA stenosis with distal thrombus formation.
B: Proximal protected (MOMA) stent implantation.
C: Acute in-stent-thrombosis one hour later.
D: Collapse of the internal carotid artery and the stent during manual aspiration through a 9F balloon catheter.
E: Re-distension of the self-expanding stent after cessation of the aspiration.
F: Normalization after thrombaspiration and i.v. application of Rheopro.
The situation is, however, completely different, if the balloon and the tip of the high bore aspiration catheter can be navigated into the distal cervical or vertical petrous segment of the ICA: application of the vacuum at that level is transmitted far distally, because the fixation of the arterial wall prevents collapse of the petrous segment of the ICA and large thrombi of the distal ICA and beyond can be aspirated (Figure 20.4A) .
Figure 20.4 A: Thrombus of a T occlusion, which could be aspirated en bloc through a 9 French balloon catheter with its tip in the petrous part of the internal carotid artery.
B: Contact aspiration thrombectomy of an occlusion of the internal carotid artery after cardiac surgery: after about 2 minutes of forced aspiration the rigid and compact thrombus is corked and fixed and could be withdrawn en bloc together with the 5F aspiration catheter while maintaining the vacuum.
Using the Bernese animal model, contact aspiration techniques proved to be effective also for recanalization of arteries with a diameter comparable to the middle cerebral artery (MCA) . Suction thrombectomy with and without proximal balloon protection was fast in application and allowed repeated attempts with low complication rates (no dissection and perforation, 1.6% vasospasms). Evaluation of the thrombus-device interaction revealed a significant increase of successful retrieval rate, when the proximal base of the thrombus entered the lumen of the 4.2 and 5 French aspiration catheters (VASCO+35ASPI; Balt, Figure 20.4B), which has been designed and CE certified to be used clinically to treat acute stroke. Once the applied force exceeded the adhesion force of the thrombus, the clot could be mobilized, returned to its original shape, and could be pulled into the guiding catheter, normally without fragmentation or loss of thrombotic material. Initial elongation of the thrombus was always visible, when the catheter was slowly removed under forced aspiration. Proximal balloon occlusion had no significant influence on the success and embolization rates in this experimental setting.
In clinical practice, suction thrombectomy of thrombi in M1/2 segments of the MCA is performed by placing a large bore (normally 8 or 9 French) guiding catheter – with or without balloon – as far distally as possible into the cervical segment of the ICA. Due to technical advances 5/6 French larger bore aspiration catheters with an inner diameter up to 0.07 inch (1.5 mm) can safely be navigated directly or coaxially over a microcatheter to the proximal surface of the thrombus, guided by a biplane roadmap. Inability to aspirate blood during the slow forward approach to the site of the occlusion indicates contact between the tip of the catheter and the thrombus. To increase the adherence between the clot and the catheter, suction force has to be increased. Optimal suction power can be generated by use of two 60 ml vacuum syringes, which can simultaneously be locked to the proximal Luer-Lock port of the aspiration catheter via a three-way stop-cock. Alternatively, aspiration force with lower pressure can be applied by one 60 or 20 ml syringe, special, dedicated pumps, or the hospital vacuum, which is present in most angio suites.
Even using larger bore aspiration catheters of the newest generation, the diameter of the clot is normally too big to be sucked entirely en bloc into the 5/6 French catheter, which could be shown directly in the experimental setting using radiopaque thrombi . To increase the adherence between the thrombus and the catheter, the vacuum is normally maintained for about 1–3 minutes while the clot can be further and deeper corked into the proximal lumen of the aspiration catheter (Figure 20.4B), dependent on its visco-elastic properties as described by the Kelvin-Voigt model . While maintaining the vacuum, the aspiration catheter is then slowly withdrawn, hopefully together with the clot remaining corked at the tip. The concept of proximal flow arrest using a balloon guide catheter seems to facilitate the mobilization of the thrombus and to diminish the risk of thrombus fragmentation and distal downstream embolization or embolization into new, previously unaffected vascular territories.
Considering small modifications, these suction thrombectomy techniques have been referred to as CA (Contact Aspiration), ADAPT (A Direct Aspiration, first Pass Technique), FAST (Forced Arterial Suction Thrombectomy), or MAT (Manual Aspiration Thrombectomy). Systemic reviews and meta-analysis revealed recanalization rates between 60% and 80% (Thrombolysis in Cerebral Ischemia – TICI 2b-3) with aspiration only. This number could be increased up to 89–100%, if in the 20–40% of cases with failure of aspiration only, a stent retriever was used as an adjunctive device. Procedure-related complications were low, with embolization of thrombus fragments in new territories in about 2% of cases only.
The Penumbra System (Penumbra, Almeda, USA) is a modification of the manual proximal aspiration technique and consists of a dedicated reperfusion catheter connected to a pumping system applying continuous aspiration. A microwire with an olive-shaped tip, the separator, is used to clean the tip of the reperfusion catheter of clot fragments in order to avoid obstruction. The system was FDA-approved for acute stroke treatment in 2007 and has been investigated in several trials. The Penumbra Pivotal Stroke Trial  prospectively recruited 125 stroke patients (mean NIHSS 18) within 8 hours of symptom onset. Recanalization of the target vessel was successful in 81.6% of patients. Nevertheless, good clinical outcome was achieved in only 25% of all patients and in 29% of patients with recanalization of the target vessel. Mortality was comparatively high (32.8%) and symptomatic intracranial hemorrhage occurred in 11.2% of cases. The poor clinical outcome despite the relatively high recanalization rate in this trial prompted discussion of the impact of recanalization using mechanical thrombectomy. Subsequent single-center studies showed better clinical results using the Penumbra System. Single-center series reported successful recanalization in up to 90% with large-vessel occlusion (mean NIHSS 14) and good clinical outcome in 48% with a mortality rate of 11%. Mean procedure time was 1.6 hours.
The THERAPY trial  compared different aspiration techniques in addition to i.v. thrombolysis to medical treatment with alteplase alone. The Penumbra separator was used in about 25% of cases. Large bore distal access catheters have been used in 27% of the 55 patients in the treatment arm only. Although underpowered, superior clinical outcome was seen for aspiration thrombectomy: good clinical outcome 90 days later was in favor for aspiration thrombectomy (p = 0.05, 2.2 odds ratio).
In suction thrombectomy or aspiration thrombectomy, an aspiration catheter is guided to the occlusion site, then part of the clot is sucked into the tip of the catheter and withdrawn together with the catheter. Systemic reviews and meta-analysis revealed recanalization rates between 60 and 80% (Thrombolysis in Cerebral Ischemia – TICI 2b-3) with aspiration only. This number could be increased up to 89–100%, with increase of the lumen of the aspiration catheters and if in the 20–40% of cases with failure of aspiration only, a stent retriever was used as an adjunctive device.
The most recently introduced mechanical treatment approach relates to so-called “stent retrievers” or “stentrievers.” Stent retrievers are self-expandable, re-sheathable, and re-constrainable stent-like devices. Wakhloo et al. demonstrated the technical feasibility of using a retrievable, closed cell, self-expanding stent (Enterprise, Codman) for extraction of foreign bodies and clot in in vitro and in animal testing by partially deploying and then retracting the device . Kelly et al.  first described the use of a partially unconstrained stent (Enterprise, Codman) to provide a temporary endovascular bypass to achieve recanalization of an M1 occlusion refractory to previous thrombolytic and mechanical treatment. The original intention of the temporary bypass was to deploy the stent across the thrombus to facilitate intra-arterial thrombolysis with consecutive recovery of the stent into the microcatheter after clot dissolution without primary intention for thrombectomy. Pérez and coworkers  reported their first use of a fully deployable, self-expanding stent (Solitaire AB, ev3/Covidien) as a thrombectomy device for emergency treatment of an M1 occlusion after previously failed mechanical thrombectomy with a distal clot retriever. Mechanical thrombectomy using stent retrievers is an emerging treatment approach for acute ischemic stroke. The concept of stent retrievers combines the advantages of intracranial stent deployment with immediate flow restoration (Figure 20.5) and a thrombectomy device with definitive clot removal from the occluded artery.
Figure 20.5 An 85-year-old patient presenting 3 hours after sudden onset of aphasia and right hemiparesis (NIHSS 6). Emergency MRI revealing diffusion restriction in the basal ganglia, the head of the caudate nucleus, and the anterior limb of the internal capsule on the left with corresponding ADC hypointensity (A, B). Perfusion imaging demonstrating impaired perfusion with prolonged MTT in the entire left middle cerebral artery territory (C). Contrast-enhanced MRA showing an occlusion of the left M1 segment (D). DSA confirming the left M1 occlusion in antero-posterior and lateral view (E and F). Distal angiogram after passing of the occlusion site with the microcatheter to confirm proper microcatheter localization distal to the occlusion (G). Mechanical thrombectomy was performed using a retrievable stent (Solitaire FR Revascularization Device 4 × 20 mm). The Solitaire FR was placed through the microcatheter and deployed by retracting the microcatheter through the occlusion site under fluoroscopic control (H). Once fully deployed, a control angiogram was obtained to assess immediate recanalization effect, which showed immediate partial recanalization of the middle cerebral artery (arrow in I). Control angiogram after device retrieval demonstrated complete recanalization of the M1 segment corresponding to a TICI 3 grade (J and K). Control CT after 24 hours showed infarction restricted to the basal ganglia corresponding to the pre-existing DWI lesion on the pre-interventional MRI without further ischemic demarcation in the middle cerebral artery territory (L).
The complete removal of the device avoids the major disadvantages associated with permanent stent implantation, such as the need for double anti-platelet medication, which potentially increases the risk of hemorrhagic complications and the risk of in-stent thrombosis or delayed stenosis.
The application of stent retrievers is comparable to that of intracranial stents. Initially, the occlusion site is passed with a microcatheter (0.0165–0.027 inches) and the stent retriever is deployed by retrieving the microcatheter and unsheathing the device covering the entire thrombus. The radial force of the stent retriever is able to immediately generate a channel by compressing the thrombus and to partially restore blood flow to the distal territory in most cases, creating a channel for a temporary bypass (Figures 20.5 and 20.6).
Figure 20.6 A 66-year-old patient presenting with aphasia and right hemiparesis (NIHSS 8). DSA demonstrating left ICA pseudo-occlusion due to a high grade ICA stenosis (A and B). Control DSA after PTA of the stenosis using proximal balloon protection showing reconstitution of flow in the ICA and concomitant left M1 occlusion (C and D). Passing of the occlusion site with the microcatheter and distal injection to confirm proper microcatheter localization (E). Deployment of a stent retriever (Solitaire FR 4 × 20 mm) using triple access technique with additional aspiration demonstrating immediate partial recanalization effect (arrow in F). Simultaneous near-infrared spectroscopy measurements revealing immediate increase in oxyhemoglobin (red curve) and total hemoglobin (green curve), as well as a decrease in deoxyhemoglobin (blue curve) in the ipsilateral MCA territory (upper row in G) corresponding to reconstitution of perfusion after recanalization. No change in oxygenation measurements are seen in the contralateral MCA territory (lower row in G). Final angiogram showing complete recanalization of the ICA and the M1 segment corresponding to a TICI 3 grade reperfusion of the brain, downstream to the tandem-occlusion (H and I).
Adjuvant intra-arterial thrombolysis can be applied and the temporary bypass effect can be used to facilitate clot dissolution by increasing the thrombus surface in contact with thrombolytic drugs. However, the device is typically left in place for an embedding time of 5–10 minutes allowing engagement of the thrombus within the stent struts [37, 38]. During retrieval of the stent retriever into the guide catheter, proximal balloon occlusion and flow reversal by additional aspiration at the guide catheter is again recommended. In vivo experimental studies have illustrated incorporation of the thrombus within the stent struts. During mobilization and retrieval of the device, the thrombus-device complex remains in a straight position without obvious compression or elongation of the clot material . This might result in an increased retrieval force required to mobilize the thrombus and much higher retrieval success rate, compared to older, distal access devices . Therefore, straight thrombus position during retrieval and firm clot engagement appear to be key features of stent retrievers compared to the mechanical principle of action of other thrombectomy devices and may explain their high success rates . However, since the optimal design of stent retrievers allowing maximal clot engagement remains unclear, variations of retriever designs have been developed. The different designs vary in terms of radial force (“lower” vs. “higher,” zones with variable radial force), stent design (open-end vs. closed-end, delivery profile), stent cell design (open-cell vs. closed-cell vs. hybrid), and material.
The first dedicated combined flow restoration and thrombectomy device for acute stroke treatment was the Solitaire FR (ev3/Covidien, Irvine, USA), receiving the CE mark in 2009 and FDA approval in 2012. The device is based on the Solitaire AB Neurovascular Remodeling Device, originally developed for stent-assisted treatment of wide-neck intracranial aneurysms. Within a short period of time, numerous studies have reported the in vivo and clinical application of the Solitaire FR for stroke treatment [37–39]. The first clinical case series published by Castaño et al.  included 20 patients with M1 and carotid terminus occlusions and showed the ability for fast and efficient clot retrieval using the Solitaire AB with successful recanalization (TICI 3 or 2b) in 90% of patients and good clinical outcome (mRS ≤2) in 45% of patients. Subsequent single- and multicenter studies have demonstrated the potential to reduce the procedure time (20–55 minutes) and to increase recanalization rates to over 80–90% in large cerebral arteries, with favorable clinical outcome in a large percentage of patients (42–57.9%) , indicating the potential of this technique to be established as a major approach to endovascular stroke treatment.
The SWIFT study (Solitaire FR with the Intention for Thrombectomy ) was a prospective, randomized trial comparing the efficacy and safety of the Solitaire FR with the MERCI device within 8 hours of symptom onset. The trial was halted sooner than anticipated due to a significantly better clinical outcome in the Solitaire FR patient group. Successful recanalization was achieved in 83.3% of cases with the Solitaire FR compared with 48.1% with the MERCI retriever, with good clinical outcome of 58.2% versus 33.3%, respectively. Overall, 40% of patients had already been treated with intravenous rtPA, but failed to improve. SICH occurred in 2% of the Solitaire FR group and in 11% of the MERCI device group, with mortality rates of 17% and 38%, respectively.
The TREVO 2 study (Thrombectomy REvascularization of large-Vessel Occlusions in acute ischemic stroke ) was a randomized trial comparing the Trevo Pro retriever (Concentric Medical/Stryker Neurovascular, USA) with the MERCI device within 8 hours of symptom onset. Successful recanalization was achieved in 89.7% in the Trevo group compared to 63.3% in the MERCI group with good clinical outcome in 55% and 40%, respectively. sICH occurred in 6.8% in the Trevo group and in 8.9% of the MERCI group with mortality rates of 33% versus 24%, respectively.
The results of these trials support the assumption that there are distinctive mechanical mechanisms of action and therefore different success and efficacy rates depending on the mechanical approaches applied with superiority of stent retrievers over distal thrombectomy devices.
Using stent retriever devices compared to intravenous thrombolysis alone in several randomized controlled trials has consistently demonstrated superior clinical outcome in acute ischemic stroke patients with large-vessel occlusion of the anterior circulation undergoing mechanical thrombectomy; these studies and subanalyses have been described in detail in the previous chapter. The MR CLEAN trial was the first trial terminated showing an absolute difference of 13.5% (95% CI 5.9–21.2) in the rate of functional independence (mRS 0–2) in favor of mechanical thrombectomy compared to intravenous thrombolysis alone (32.6% vs. 19.1%). There were no significant differences in mortality or rate of SICH. In consequence of these positive results several ongoing randomized trials at the time have been prematurely stopped for interim analysis showing similar beneficial outcome. A meta-analysis based on the individual patient data of five randomized trials (MR CLEAN, EXTEND IA, ESCAPE, SWIFT PRIME, REVASCAT) including 1 287 patients by the Hermes Collaboration group  showed a significantly reduced rate of disability at 90 days in patients undergoing stent retriever thrombectomy compared to the control group (adjusted cOR 2.49, 95% CI 1.76–3.53; p <0.0001). The number needed to treat with endovascular thrombectomy to reduce disability by at least one level on the mRS for one patient was 2.6. Subgroup analysis of the primary endpoint showed no heterogeneity of treatment effect across pre-specified subgroups for reduced disability (p = 0.43). Effect sizes favoring endovascular thrombectomy over control were present in several subgroups, including in patients aged 80 years or older (cOR 3.68, 95% CI 1.95–6.92), patients randomized more than 300 minutes after symptom onset (1.76, 1.05–2.97), and patients not eligible for intravenous alteplase (2.43, 1.30–4.55). Mortality at 90 days and risk of SICH did not differ between the treatment groups. This meta-analysis further confirmed the substantial benefit of mechanical thrombectomy using stent retrievers across a range of subgroups in stroke patients suffering from acute ischemic stroke due to large anterior circulation vessel occlusion.
Mechanical thrombectomy using stent retrievers combines the advantages of intracranial stent deployment with immediate flow restoration and a thrombectomy device with definitive clot removal from the occluded artery. The complete removal of the device avoids the major disadvantages associated with permanent stent implantation, such as the need for double anti-platelet medication. Protected stent retriever thrombectomy avoids fragmentation of the thrombus and iatrogenic distal emboli with racanalization and reperfusion rates of about 90% and a number needed to treat (NNT) of 3, as confirmed by multiple clinical trials.