The pioneering past and cutting-edge future of interventional neuroradiology

Abstract

This review provides a thorough understanding of the developments in the field of interventional neuroradiology (INR). A concise overview of the pioneering past and current state of this field is presented first, followed by a greater emphasis on its future. Five main aspects predicted to undergo significant developments are identified and discussed. These include changes in ‘education and training’, ‘clinical practice and logistics’, ‘devices and equipment’, ‘techniques and procedures’, and ‘relevant diagnostic imaging’. INR is at the crossroads of neuroradiology, neurosurgery, neurology, and the neurosciences. To progress we must value the uniqueness and vitality of this multidisciplinary aspect. While minimal access techniques offer very good anatomical accessibility to treat multiple pathologies of the central nervous system, it is also important to recognise its limitations. Medical, surgical, and radiosurgery modalities retain an important role in the management of some complex neuropathology. This review is certainly not exhaustive of all ongoing and predicted developments, but it is an important update for INR specialists and other interested professionals.

PAST

In 1927, neurologist António Egas Moniz collaborated with neurosurgeon Pedro Almeida Lima to perform the first successful cerebral angiogram (aka. arterial encephalogram) in a living patient. Originally, this technique was developed as an alternative to ventriculography and pneumoencephalography (developed by Walter Dandy in 1918 and 1919 respectively), to localise cerebral tumours by evaluating displacement of the intracranial blood vessels. ¹ Compared to pneumoencephalography, this technique also provided a better understanding of the tumours’ vascular supply. Naturally, this was then applied to directly evaluate other cerebral vascular pathology. It remained the only reliable diagnostic technique to investigate cerebral pathology until 1975 when computed tomography (CT) was introduced into clinical practice. ² The initial attempts at arterial encephalography were done with percutaneous puncture of the internal carotid artery at the neck, which was associated with complications including stroke and death. The approach was then changed to surgical cut-down onto the vessel with temporary occlusion of its proximal segment under local anaesthesia, which became the accepted method for the next 2 to 3 decades.^1,3

In 1939, the first endosaccular coiling of a cranial aneurysm was performed using a transorbital approach to place 30 feet of silver wire directly into a giant cavernous carotid aneurysm. ⁴ The description of femoral catheterisation for retrograde angiography by Farinas in 1941 significantly reduced the risk of ischaemic stroke from embolisation. ⁵ This approach was consolidated by the Sven Ivar Seldinger technique of percutaneous catheterisation developed in 1953. ⁶ In 1960, Luessenhop and Spence described a case of endovascular embolisation of a brain arteriovenous malformation (AVM). ⁷ In 1964, Luessenhop and Velasquez (1964) first described catheterization of the intracranial arteries and catheter-based embolisation of an intracranial aneurysm. ⁸

The 1970s are regarded as the beginnings of modern neurointervention. ⁹ This includes remarkable contributions from Pierre Lasjaunias towards understanding complex cerebral vascular anatomy, and Fedor Serbinenko's development of detachable latex balloon catheters for the treatment of carotid-cavernous fistulae and intracranial aneurysms.^9,10 In the 1980s, Zeumer and Theron developed endovascular stroke treatment by opening intracranial vascular occlusions with locally administered fibrinolytics.^11,12 In the 1990s, Guido Guglielmi invented detachable coils which were used for the first time in a patient with a carotid artery aneurysm. ¹³ Endovascular coiling became a well-established treatment option for both ruptured and unruptured aneurysms. During this same period, engineering innovations nurtured procedural innovations that targeted other pathological conditions, including arteriovenous malformations and acute stroke.

The International Subarachnoid Aneurysm Trial was published in 2002, showing that endovascular coiling provides a clear benefit over surgical clipping for ruptured aneurysms. ¹⁴ This was the point at which healthcare and pharmaceutical companies heavily invested in interventional neuroradiology (INR), driving technologies such as new coils, balloon-catheters, stents, endosaccular flow disruptors, and liquid embolic agents. In 2015, multiple randomised controlled trials demonstrated the clear benefit of mechanical thrombectomy for large vessel occlusion which revolutionised stroke treatment.^15–17

PRESENT

Historical breakthroughs and disruptive innovations have made it possible for large hospitals around the world to develop comprehensive neurovascular intervention departments. These centres tackle an extensive variety of pathologies, including a variety of the procedures listed in Table 1.

Table 1.

Contemporary procedures in interventional neuroradiology.

	CRANIAL PROCEDURES	NECK & SPINAL PROCEDURES
DIAGNOSTIC PROCEDURES	- Angiogram - Dural venous pressure measurements - Inferior petrosal sinus sampling - Wada test (intracarotid sodium amobarbital procedure)	- Angiogram - Myelography and cisternography - Image guided lumbar punctures - Vertebral biopsy
THERAPEUTIC PROCEDURES	- Aneurysm embolisation - Arterial mechanical thrombectomy - Arteriovenous malformation embolisation - Arteriovenous fistula embolisation - Vein of Galen Malformation embolisation - Middle meningeal artery embolisation for subdural haemorrhage - Embolisation for epistaxis - Tumour embolisation (meningiomas, lingual carcinomas) - Venous mechanical thrombectomy - Venoplasty and venous stenting - Arterial angioplasty for vasospasm - Intra-ophthalmic artery chemotherapy for retinoblastoma	- Arteriovenous malformation embolisation - Arteriovenous fistula embolisation - Tumour embolisation (paragangliomas, vertebral tumours, etc.) - Carotid artery angioplasty and stenting - Vertebroplasty - Spinal nerve root blocks - Image guided lumbar punctures - Image guided epidural injections (blood patch, analgesia) - Cerebrospinal fluid-to-venous fistula embolisation
OTHER		- Pre-operative site-marking with coils

Some therapeutic procedures, such as aneurysm coiling and thrombectomy, have become well established with ongoing improvements in their implementation and technical aspects. This includes the use of relatively novel devices for aneurysm treatment such as the woven endobridge device and flow-diversion stents. Procedures which are particularly challenging may require less-frequently-used methods, such as the transorbital approach to treat indirect carotico-cavernous fistulas. ¹⁸ Other procedures are slowly becoming more prevalent. For example, endovascular embolisation of arteriovenous malformations has historically been an adjunct to microsurgery or radiosurgery. However, this has been changing in recent years with some centres currently using this technique as a first-line treatment option to cure AVMs.

Some procedures are performed although evidence is still somewhat limited. This includes the new promising treatment of middle meningeal artery embolisation for chronic subdural haemorrhage, and older techniques such as venous mechanical thrombectomy and arterial angioplasty.^19–21 This subset of treatments will require further evaluation with more scientific rigour and clinical trials.

Many other concepts, tackled in the next section, are still being developed and will generate future innovative disruptions in the way we treat patients.

FUTURE

On reflection of this prospective evaluation into the future of INR, five main aspects that are predicted to undergo significant developments have been identified:

1. Education and training

2. Clinical practice and logistics

3. Devices and equipment

4. Techniques and procedures

5. Relevant diagnostic imaging

Developments in education and training

Interventional neuroradiology, endovascular neurosurgery, and neurointerventional surgery all refer to the same speciality of minimal access operations which is at the crossroads of neuroradiology, neurosurgery, neurology, and the neurosciences. In mainland Europe and the United Kingdom endovascular intervention is mostly performed by neuroradiologists. In North and South America, it is performed by neuroradiologists, neurosurgeons and neurologists. ²² In Japan and China, it is performed mostly by neurosurgeons. ²³ Although this has the risk of creating a sense of turf war, we must value the uniqueness and vitality of this multidisciplinary aspect if we do not wish to stifle progress.

Consequently, INR is not bound by the classical limits of a typical speciality and is not restricted by standard formats of teaching and education. A challenge faced by each trainee from all pathways leading to INR is the balance between that which is considered conventional day-to-day work for their specific speciality (i.e., diagnostic radiology for radiology trainees, medical management for neurology trainees, and general neurosurgery for neurosurgery trainees) and the time devoted to neurointerventional work. INR is steadily becoming a busier and more demanding speciality due to the rising need for mechanical thrombectomy and other new treatment opportunities. This requires highly dedicated specialists with vast exposure to intervention. Therefore, there is the possibility that the dichotomy between INR and the other general work done by these other specialities will become more evident.

In 2009, the Union of European Medical Specialists took into consideration recommendations by the World Federation of Interventional and Therapeutic Neuroradiology and published a training charter for INR. This aims to have physicians interested in the endovascular treatment of vascular disorders involving the central nervous system to be fully trained in both the technical aspects of the procedures, and to deal with the overall management of patients (so called “interventional neurotherapist”).^24–26 This would be a 4-year full-time program involving 24 months of core INR, 12 months of diagnostic neuroradiology, and 12 months of clinical neuroscience. ²⁶

Irrespective of the training pathway and structure, simulation-based training has made its way into INR. This is grounded in methodologies developed in the aviation industry in which it is well established. Applying a similar approach in INR could yield huge benefits. In fact, this is being developed for different procedures, designed for trainees to acquire the necessary skills in advance of performing procedures on patients, and then maintaining or upskilling as they become more experienced.^27–29 Simulation may also become useful for experienced operators to perform patient specific procedure rehearsals before undertaking a complicated case on the actual patient. The three main forms of simulation in INR are animal models, virtual simulation and flow models (e.g., 3D silicone models), each with its own advantages and disadvantages. Animal models are becoming less of an option due to ethical issues, limited availability, and risk of the animal not surviving repeated procedures. Virtual simulation and flow-models are becoming the more preferred options and should be considered complementary rather than exclusive to one another. The predominant limitation with virtual technology is the difficulty to simulate realistic haptic feedback, an issue that is being researched for technical improvements. ³⁰ Even allowing for such a limitation, this method of simulation remains a useful means to adequately learn procedural steps in an efficient manner. Examples of virtual simulation technology include Mentice, Cathi, and Simbionix.

Developments in clinical practice and logistics

The importance of caring for patients by interventionalists before, during and after a procedure has been advocated since the inception of interventional radiology as a speciality.^31,32 However, over 6 decades later most IR consultants around the world still do not have admitting rights and only a minority of radiology departments have allocated beds. Such responsibilities and accessibility appear to be more prevalent in eastern countries such as China, but less so in the United States, United Kingdom, and Europe. ³³ Better accessibility to beds may be achieved by adopting a system of “pooled beds” rather than beds allocated to a single department.

Clinical knowledge is a must for interventional neuroradiologists to admit patients under their care and manage them on the wards. If interventional neuroradiology is to be an equal partner in the delivery of care for the patients, it must assume primary responsibility and be completely engaged in the decision-making process of the patient's management by having a clinical standing and clinical credibility. Certain countries, such as the United States, recruit neurologists and neurosurgeons into neurointervention (as well as radiologists), whereas other countries, such as the UK, have a requisite to complete clinical foundation training before starting specialisation in radiology. Since all neurointerventionalists attain clinical training in their formative years, it is important to have a system in place which allows them to maintain and continue developing these skills along with their imaging and procedural skills. An alternative approach for interventional neuroradiologists would be to forge collaborative relationships with other clinical teams, but still assume primary responsibility for the patient.

Outpatient clinics and patient management outside of the intra- and peri-procedural settings have also become key components of an interventional radiology practice. Pre- and post- procedural clinics allow the patient to learn about the procedure, explore alternative therapies, and ask pertinent questions. The physician is also able to assess outcomes and complications, and plan additional treatment if necessary. An emerging tool in this respect is telemedicine. This was already implemented by other specialities, such as neurology, to deliver care to patients in remote areas. It has been an uncommon practice in interventional neuroradiology, until the implementation of social-distancing during the COVID-19 pandemic. Evaluation of such clinics has demonstrated that virtual platforms for INR care are feasible and sustainable. In fact, this was the preferred method by both patients and physicians for non-urgent follow-up appointments. ³⁴

Virtual communication systems have also gained traction for remote assistance amongst INR colleagues. This allows interventionalists who require another opinion, to do a video consultation at any time with a colleague from a different hospital or a different country. ³⁵

The human workforce is increasingly becoming supported by new technological advancements, artificial intelligence (AI), and networks to address challenging demands. In the context of stroke management, which is a time sensitive and complex care pathway, an Integrated Stroke System Network has been described. ³⁶ This involves multiple comprehensive stroke centres connected to multiple primary stroke centres, which provides aggregated data that guides patient management in a large geographical area. This is managed by a stand-alone patient command centre managing all stroke calls. AI detection of a large vessel occlusion on a scan would automatically alert the complex stroke centres and then assign the case to a specific hospital within the network, depending on criteria such as distance from call location and bed availability.

Another anticipated change in the INR workforce has to do with the global drive to improve the representation of women in the workplace. Women's representation in medicine has gradually increased over the past decades, and over 50% of medical students in many countries are now female. ³⁷ Therefore, the prevalence of female doctors is steadily rising, and the proportion of women in some traditionally male-dominated specialities such as surgery is also increasing. ³⁸ Like neurosurgery and general interventional radiology, INR remains one of the most male-dominated subspecialties in medicine. ³⁹ A recent survey explored the reasons that could underly female underrepresentation in INR. Compared to men: significantly fewer were married and had children; significantly fewer held supervisory roles and academic titles; significantly less had a mentor; were less satisfied in their careers; more felt they receive less recognition; less proportion of work time dedicated to neurointervention. ⁴⁰ Another deterrent is fear related to radiation and orthopaedic stress during pregnancy. In fact, pregnancy remains one of the most important issues for young women in choosing a specialisation program. ⁴¹ However, a recent study suggests that with optimal radiation safety practice, the fetal dose of a pregnant neurointerventionalist is negligible. ⁴² As more of these issues are studied and addressed, we expect to have an increasingly greater number of women in INR.

Development of new devices and equipment

New devices and equipment are constantly being developed to improve the possibilities of already existing and new endovascular treatments. This section is intended to highlight some of the most remarkable and pertinent of these developments. However, it is certainly not exhaustive of the work in progress.

Wires:

A limitation of conventional guidewires in interventional radiology is the pre-defined tip shape which cannot be changed while it is in the vessel. The Columbus steerable 0.014-inch guidewire (aka. Drivewire) (Rapid Medical) was designed to address this problem. Its tip is remotely controlled and deflected within the patient from a handle at the proximal end of the wire which is pulled or pushed. This guidewire gained Food and Drug Administration (FDA) approval in 2020. It is reported to lack torquability compared with other available guidewires but offers good support at the tip which allows manoeuvres that are impossible with other available wires. ⁴³

A magnetic navigation system has also been developed for the same purpose, with better control of guidewire manipulation. ⁴⁴ This consists of two permanent magnets positioned on either side of the fluoroscopy table. These allow for omnidirectional rotation of a 0.014-inch magnetic guidewire, altering the wire tip while in the vessel. In vitro studies suggest better accuracy with faster and easier access of neurovascular anatomy, compared to manipulation of traditional guidewires with a predefined shape. This has also been proven to be safe on a small group of 10 patients. ⁴⁵ However, a major limitation with this technique is that implementation requires a costly infrastructure.

Catheters:

A recent development was that of mini-balloons like the Scepter Mini-Balloon microcatheter (Microvention) which was FDA approved in 2019. This is being clinically explored with an increasing number of technical applications being reported in the literature, such as safe delivery of liquid embolic agents and an aid in positioning a woven endobridge device. ⁴⁶

In the realm of thrombectomy, there has also been an effort to develop larger bore aspiration catheters on the premise that matching the catheter size closely to the vessel size improves the effectiveness of clot aspiration. ⁴⁷ A recently introduced aspiration catheter is the Millepede 0.088-inch (Perfuze) with reported good outcomes in the literature. ⁴⁸ Another development for clot aspiration is the pulsed/cyclical aspiration pump, with studies suggesting that this provides superior aspiration compared to continuous uniform aspiration. ⁴⁹

Stents:

A remarkable stent concept is that of ‘stentrodes’ (stent-electrodes) for endovascular deep brain stimulation which is an appealing alternative to conventional surgical deep brain stimulation, with indications including Parkinson's disease, tremor, and dystonia. ⁵⁰ This would involve implantation of vascular stents that contain electrodes which can be activated to stimulate or inhibit specific targets. However, this requires a much more detailed characterisation of the relationships between neurovascular structures and established targets in deep brain structures than currently understood. A study investigating this concept concluded that although at present there are several challenges, endovascular neuromodulation seems feasible. ⁵¹

Another appealing alternative to conventional deep brain stimulation (aside from stent devices), which is currently gaining traction in clinical practice and worth a mention, is Deep Brain Ultrasound Therapy (aka. Magnetic Resonance-Guided High-Intensity Focused Ultrasound). This technology uses high power ultrasound focused onto a specific brain target from multiple source-points with subsequent precise heating, cell death, and ablation at the target. ⁵² This has already become quite established for essential tremor, Parkinson's disease, and dyskinesia.

Stentrode implantation has also been applied with a different strategy to facilitate brain-computer interface and machine-learning-assisted-training. This has been reported in two severely paralysed patients with amyotrophic lateral sclerosis. Following endovascular implantation of the neuroprosthesis in the superior sagittal sinus through the venous route, the training allowed the patients to wirelessly transmit electrocorticography signal associated with attempted movements to control multiple computer mouse-click actions, that ultimately resulted in improved functional independence. ⁵³

In the context of symptomatic high-grade intracranial atherosclerotic disease, drug-eluting stents, such as the DES (NOVA intracranial sirolimus-eluting stent system) or BMS (Apollo intracranial stent system) are being explored. A randomised clinical trial assessing the efficacy of these stents against baremetal stents, suggests a favourable effect with reduced risk of in-stent stenosis and ischaemic stroke recurrence. ⁵⁴

AI and robotics:

AI and modelling software have been applied to computational fluid dynamics. This helps with the development of a predictive framework which is individualised for each patient. The software helps with sizing of endovascular devices, risk assessment, and ensuring appropriate patient selection eg. Sim&Size. The aim is to make better informed decisions with more predictable and desirable outcomes.

Robotics are being developed to perform endovascular interventional neuroradiology. At present this still involves a responder unit which requires manual loading of the wire and the catheter. The robotic stage then makes use of a controller which may involve a combination of a joystick, handheld remote, foot pedal, and/or touch screen. A recent literature review identified a total of 81 published neurointerventions that have been performed with a robot. ⁵⁵ These were all diagnostic cerebral angiograms or extracranial carotid artery stenting, except for 1 therapeutic intracranial intervention involving stent-assisted coiling of a large basilar aneurysm. Only 3 diagnostic cerebral angiograms (4%) had to be converted to manual operations for completion. Aside from these cases published in literature, there have also been live conference cases of stent-assisted coiling and woven endobridge device treatment with robotics. ⁵⁶

The main limitation with the current set-ups is that the operator completely relies on visual cues due to the lack of haptic and tactile feedback. ⁵⁷ In practice there are 3 main forces transmitted to the operator's hand, which are viscous, friction and impact forces. Ways in which these can be measured to inform the operator in the future include the use of sensors in the catheter tip and the robotic arm, or real-time computational estimation models. Sensory feedback when administering contrast is also important since this helps to optimise angiographic images and avoid complications from excessively injected high pressure (such as aneurysm rupture). This is being addressed with the development of haptic syringe devices. ⁵⁸ Other problems have to do with the responder unit not being compatible with already existing devices. At present, these systems are very expensive and not cost-effective. However, when they are further developed and become more affordable, they will allow the operator to perform procedures with less radiation exposure in an optimum ergonomic position with greater precision and dexterity, while eliminating physiological tremors and fatigue. ⁵⁹ Another massive advantage of these systems would be to operate from a geographically remote location (Telerobotic Stroke Network). This could be a solution to the shortage of trained operators in INR, potentially reduce inequality in the geographical distribution of such subspecialist care and facilitate quicker accessibility to time-sensitive procedures such as mechanical thrombectomy.^55,57

Several devices being developed are also aimed at pre-hospital screening for stroke. These include the robotic helmet with transcranial doppler to detect large vessel occlusion. This technology is also being applied to monitor for delayed cerebral ischaemia in the context of subarachnoid haemorrhage (SAH). ⁶⁰ Another technology intended for early detection of ischaemic and haemorrhagic stroke is the brain bioimpedance monitor. ⁶¹ Certain applications also make use of facial recognition AI technology to aid in neurological assessment. These are all portable devices suited for use outside and inside the hospital setting. Other methods based on the concept of bringing care to the patient is that of Network Mobile Stroke Units - with a CT scanner inside the ambulance which allows en-route thrombolysis and decreases time for mechanical thrombectomy or flying interventional teams who reach the patient in a primary stroke centre. ⁶² Interventional neuroradiologists and stroke physicians are also becoming more accustomed to the use of mobile applications that allow them to review AI-processed imaging for the detection of stroke, including unenhanced CT heads, CT angiograms and CT perfusion (e.g., Brainomix, RAPID, Viz.ai)

Development of new techniques, procedures, and indications

Venous procedures:

A recently published case study demonstrates the feasibility of endovascular treatment for intractable communicating hydrocephalus. ⁶³ This utilises a new biomimetic valved micro-implant device (replicating the function of an arachnoid granulation) to passively transport cerebrospinal fluid (CSF) from the subarachnoid space to the venous system (eShunt System by CereVasc). The implant consists of an approximately 3cm long tubing with a malecot tip. The endovascular technique requires a retrograde percutaneous transvenous approach from femoral to the jugular vein. The implant is deployed over a delivery catheter, and a transdural controlled penetration is performed at the inferior petrosal sinus by unsheathing a needle at the distal tip of the delivery catheter. The miniature shunt tubing is positioned to drain CSF from the cerebellopontine angle into the internal jugular vein. This mimics the function of arachnoid granulations by restoring natural CSF reabsorption into the venous system. The implant has a differential pressure slit valve to regulate CSF flow in a pressure-driven manner, proportional to the positive pressure gradient between ICP and venous blood pressure. A prospective, single-centre, single-arm pilot study is currently underway, which is aiming to enrol 30 patients (ETCHES I study: NCT04758611).

Another case study of a patient with debilitating pulsatile tinnitus was evaluated with high-fidelity computational fluid dynamics using 3D data from digital subtraction angiography. This demonstrated turbulent flow distal to venous sinus stenosis secondary to a large arachnoid granulation. ⁶⁴ The patient subsequently went on to have stenting (Wallstent by Boston Scientific) of the stenotic segment with complete resolution of the tinnitus which was sustained on follow-up after more than 2 years. CFD modelling after stenting also revealed resolution of the turbulent flow.

Paediatric procedures:

Vein of Galen malformations are vascular anomalies seen in the paediatric age group with high-flow arterio-venous shunting which often causes heart failure, irreversible damage in pulmonary vasculature, and brain damage. Prior to the 1980s (when endovascular intervention started to be applied to this pathology), the prognosis was dismal with 90–100% mortality. ⁶⁵ Presently, this is much improved with good outcome in approximately a third of cases due to several endovascular treatment strategies performed in designated centres. ⁶⁶ Approximately 32–37% survive without significant impairment (some with minor neurological injury) and 6–14% survive but have significant neurological and cognitive compromise.^67–68

A new strategy that is being proposed is to perform fetal embolisation of this malformation in-utero. The proposed procedure involves direct coiling of the main venous pouch, the median prosencephalic varix. Venous access will require advancing a needle under ultrasound guidance, through the maternal abdominal and uterine walls, then through the fetal posterior fontanelle into the torcula Herophili (dural venous confluens). The rationale behind this is to deliver early treatment with partial occlusion of the malformation, before post-natal cardiopulmonary and brain changes occur. ⁶⁹ The fetal environment is thought to provide some advantages because at birth there is loss of the low-resistance placental circulation and a physiologic steep increase in brain perfusion. ⁷⁰ Therefore, treating it in-utero allows rerouting of the deep venous system to occur under conditions of a lower-flow state. The safety and efficacy of this procedure is being tested through a prospective, single-arm non-randomised study aiming to enrol 20 patients (NCT04434729).

Oncological procedures:

Poor clinical outcomes of malignant brain tumours are driving a search for an alternative to the more established systemic therapies. A significant limitation to the latter has been the ineffective delivery of several therapeutic agents. The technological improvements that have revolutionised treatment of ischemic stroke and brain aneurysms have great potential even in oncological management. These allow for endovascular super-selective intra-arterial (ESSIA) infusion of chemotherapeutic and biological agents which selectively target tumour vasculature. ⁷¹ The technique has also been combined with agents or methods that disrupt vascular permeability (disrupt the blood-brain-barrier and blood-tumour-barrier), such as the infusion of Mannitol (osmotic agent) or infusion of microbubbles combined with focused ultrasound. Therapeutic agents that have been trialled such as temozolomide, cetuximab, and bevacizumab, achieved mixed results. ⁷⁰ There is also an ongoing phase I trial exploring ELISA infusion with allogenic bone marrow human mesenchymal stem cells loaded with oncolytic adenovirus (NCT03896568). Some of these procedures are being guided by advanced perfusion imaging to enhance precision. ⁷² A pre-procedural high resolution 3D MRI with contrast is used along with an intra-procedural cone beam CT, 3D rotational angiography, and super-selective angiographic injections are all used for this method. Image fusion techniques of these different modalities allow the interventionalist to confirm that a targeted vessel is appropriate for infusion in real-time.

Thrombectomy procedures:

Mechanical thrombectomy is now an established treatment in the management of acute ischaemic stroke. However, there are many patients at the fringes of current treatment guidelines, who are being investigated through several randomised controlled trials which are underway. These include trials; for low ASPECT scores such as INEXTREMIS-LASTE (NCT03811769), TESLA (NCT03805308), TENSION (NCT03094715), and SELECT2 (NCT03876457); for low NIHSS such as MOSTE (NCT03735979); and for distal vessel occlusion such as ESCAPE MeVO (NCT05151172) and DISTAL (NCT05029414). Other aspects to be explored are those patients with an mRS >2 prior to acute symptoms and elderly patients.

Adaptations to stroke treatment based on clot composition is a possibility that is being advocated as we understand more about clot constituents. A hyperdense middle cerebral artery sign on CT and the vascular susceptibility sign on MR Susceptibility -weighted imaging are known to suggest a clot rich in red blood cells. However, technology such as the Clotild smart 0.014-inch guidewire system (Sensome), which was FDA-approved in 2021, has also been developed to characterise clot composition. ⁷³ This has an integrated microchip with an impedance microsensor, and information is transmitted via Bluetooth to a medical tablet. The idea is that this then allows the operator to adapt and have a better chance of first-pass effect by selecting appropriate tools, such as specific stent retrievers designed for a particular type of clot. The wider understanding of clot composition is also allowing the development of new therapeutic avenues so that clots which are resistant to conventional tissue plasminogen activator (tPA) may be targeted with other thrombolytic agents such as ADAMTS13 and DNAse1 (as alternatives or combined with tPA).^74–75

Numerous other strategies, other than tPA and mechanical thrombectomy have been investigated for neuroprotective properties to limit secondary tissue loss and improve functional outcomes in acute ischaemic strokes. ⁷⁶ These include sphenopalatine ganglion stimulation, therapeutic hypothermia, ischaemic conditioning, 3K3A-activated protein C, NXY-059, nitric oxide, magnesium sulfate, NA-1 delivery, statins, Glyburide, Insulin, Neu2000KWL, Nimodipine, antioxidants and antibodies, GABA receptor agonists, GABA transaminase inhibitors, GABA transporter blockers, and stem cells. In many cases, randomised controlled studies have generally yielded disappointing results. However, it has since become clearer that strategies of neuroprotection need to be evaluated in the context of reperfusion therapy. Reperfusion in the form of thrombolysis or mechanical thrombectomy may be a necessary condition for positive effects of neuroprotective treatments. ⁷⁷

Developments in relevant diagnostic imaging

Vessel wall magnetic resonance imaging (VW-MRI) involves contrast injection and acquisition of high spatial resolution multiplanar 2D or 3D sequences with multiple signal weightings, suppression of CSF, and suppression of luminal blood signal (‘black blood’). This technique provides added detail which can play a role in the management and evaluation of aneurysms, AVMs, atherosclerosis, vasculitis, intracranial vessel dissections, and other neurovascular pathologies. ⁷⁸

Increasingly, VW-MRI is playing a role in the management of ruptured and unruptured intracranial aneurysms. In a patient with multiple aneurysms who has suffered a SAH, VW-MRI can help identify the culprit aneurysm by demonstrating mural enhancement. It can also help to predict the risk of developing vasospasm which is indicated by concentric arterial wall enhancement. ⁷⁸ In the case of unruptured aneurysms, the risk of rupture and the decision to treat is currently largely based on the size, location, morphology, and growth rate. However, VW-MRI additionally provides a means of assessing for inflammatory activity, neovascularisation, and atheromatic lesions in the aneurysm which are indicative of wall fragility, instability, and increased risk of rupture. ⁷⁸ The same approach has been proposed (although with more limited evidence) for cerebral AVMs, to identify the site of bleeding in a nidus and possible utility in identifying AVMs at high risk of rupture. VW-MRI also serves to determine whether paraclinoid aneurysms are intradural (posing a risk for SAH) or extradural, which is important for prognostication and therapeutic decision. ⁷⁸ (Note: unlike extracranial arteries, normal intracranial arteries lack vasa vasorum and do not show concentric mural enhancement.)

Another imaging modality that is being utilised for patients with acute ischaemic stroke who have undergone thrombectomy is dual energy CT. This has been shown to have very good accuracy and specificity in differentiating cerebral staining with iodine contrast agent from cerebral haemorrhage. ⁷⁹ In turn, this has implications on decision making with respect to the antithrombotic strategy.

Conclusion

INR has gone a long way since its rudimentary beginnings. There have been continuous adaptations of INR tools and concepts to the delicateness and idiosyncrasies of the cerebral vasculature, enabling remarkable advancements. This review is certainly not exhaustive of all ongoing and predicted developments, but it is an important update to INR specialists and other interested professionals. Modern endovascular and other minimal access techniques offer very good anatomical accessibility to treat multiple pathologies of the central nervous system efficiently and safely. The appeal of such an approach cannot be matched by conventional surgical procedures. However, it is important to recognise that INR also has its limitations, and other treatment modalities such as medical, surgical, and radiosurgical treatments retain an important role in the management of some complex neuropathology.

The Gartner hype cycle is a graphical depiction of expectation against time, representing the maturity, adoption, and application of innovative technologies. This suggests five predictable phases; a “technology trigger” causes a rapid increase in technology visibility, ultimately reaching a “peak of inflated expectations,” then declining in visibility to the “trough of disillusionment,” followed by an increase at a more moderate rate on the “slope of enlightenment,” before finally reaching the “plateau of productivity.” ⁸⁰ In some respects, this may be applicable to INR. In essence, concepts that have demonstrated significant potential for clinical practice and that appear promising for the future of INR should be capitalised to motivate change. However, these must be regarded with caution and with full awareness that appropriate evaluation and development takes time. Any concept which is implemented into clinical practice prematurely without undergoing the scrutiny of scientific rigour can work against the progress of this speciality.

Abbreviations

AI

Artificial intelligence

AVM

Arteriovenous malformation

CT

Computed tomography

CSF

Cerebrospinal fluid

FDA

Food and drug administration

INR

Interventional neuroradiology

MRI

Magnetic resonance imaging

SAH

Subarachnoid haemorrhage

tPA

Tissue plasminogen activator

VW-MRI

Vessel wall magnetic resonance imaging