Abstract
Introduction Transarterial embolization is a well-established adjunct in the management of hypervascular head, neck, and spine tumors. Few case-series have described the role of direct percutaneous embolization (DPE).
Method We report our experience with DPE (with or without transarterial embolization) in patients with head, neck, and spine tumors treated between 2012-2023. Baseline demographics, angiographic imaging, tumoral pathology, and relevant clinical variables were retrospectively reviewed and descriptively analyzed.
Result A total of 55 patients underwent direct percutaneous embolization (19 of whom also received transarterial embolization either before [n=7] or after [n=12] DPE). The most commonly embolized lesions were malignant carcinomas (n=16), followed by juvenile nasopharyngeal angiofibromas (n=11), paragangliomas (n=8), hemangiomas (n=7), and others (hemangioblastoma, schwannoma, neurofibroma, esthesioneuroblastoma, meningioma and hemangiopericytoma). The most common locations were sinonasal/nasopharyngeal (n=21), followed by scalp/subcutaneous (n=18), carotid body (n=6), spinal/paraspinal (n=4), skull base (n=5), and intracranial (n=1). A median of 4 (interquartile range 4-6) twenty-two-gauge spinal needles were used per embolization with a median fluoroscopy time of 50.5 (23.2-77.8) minutes resulting in median radiation exposure of 3055 (840.5-5053.5) mGy. Seven patients received more than one embolic agent, with n-butyl cyanoacrylate (glue, n=44, 81.5%) being the mostly commonly used embolic, followed by ethylene-vinyl alcohol copolymer (Onyx, n=7, 12.9%), 98% dehydrated ethanol (n=7, 12.9%), sodium tetradecyl sulfate (n=2, 3.7%) and poly-vinyl alcohol particles (n=1, 1.8%). The median volume of embolic agent injected was 5.5(4-7.6) mL resulting in total/near-total (90%-99%) angiographic devascularization in 74.5% cases. The median operative blood-loss was 250(75-700) mL. One patient underwent trans-calavarial DPE for a cerebellar hemangioblastoma and suffered diffuse subarachnoid hemorrhage from profuse tumoral bleeding. One patient had an asymptomatic parent-vessel occlusion from retrograde embolic extension.
Conclusion Our single-center study reinforces prior experience that DPE of Sino-nasal carcinomas, angiofibromas and paragangliomas with adhesive and non-adhesive liquid embolic agents is safe, feasible and effective. Further, it suggests that these benefits may also be extended to non-traditional head, neck and spine tumors. Caution must be exercised when applying these techniques to intracranial tumors with robust intratumoral arteriovenous shunting.
Introduction
Trans-arterial embolization (TAE) in the management of hypervascular tumors of the head, neck, and spine has been shown to be a safe and effective adjuvant to surgical resection or palliation.1 However, the tortuosity of parasitized vasculature and risk of embolic ischemic stroke may preclude its use.1,2 Since the introduction of direct percutaneous embolization (DPE) via intra-tumoral injection of embolic material in the 1990’s, several case-reports and short-series have supported the safety and efficacy of this technique.1,3 In the absence of level A/B evidence from randomized trials to support use of TAE/DPE for tumor embolization,1,3 we provide further retrospective data favoring DPE of head, neck, and spine tumors using liquid embolic agents (ethylene-vinyl alcohol copolymer, Onyx;™ or n-butyl cyanoacrylate, n-BCA; sodium tetradecyl sulfate, Sotradecol;® or polyvinyl-alcohol particles) in 55 consecutive patients at a tertiary-referral University Hospital. Further, we add to the contemporary practice of trans-calavarial embolization of infratentorial tumors reported in three patients prior, 4,5 by reporting our experience with DPE of a cerebellar hemangioblastoma. We describe the technical aspects of the procedure and its efficacy in reducing intraoperative blood loss.
Methods
This was a retrospective, observational, study conducted at our institution from a prospectively maintained database after receiving a waiver from the Institutional Review Board on the need to obtain informed patient consent given the retrospective nature of the study. The study was performed under the guidelines outlined by the Declaration of Helsinki. Patients with hypervascular lesions of head, neck, or spine that were referred to the neurointerventional service for consideration of embolization between 2012 and 2023 at The University of Minnesota were retrospectively reviewed. All patients meeting inclusion criteria underwent DPE with or without TAE, and had adequate clinico-radiographic data including their operative and pathology reports available for review. Patients undergoing TAE and/or DPE received general anesthesia with or without neuromonitoring (electroencephalography, visual evoke potentials, and somatosensory-evoked potentials). All patients received at least one transfemoral or transradial diagnostic angiogram to delineate the arterial and venous angioarchitecture. Depending on the anatomic location of the tumor, the following craniocervical arteries were injected: internal carotid, external carotid, vertebral, subclavian, costocervical, and thyrocervical. The following angiographic/radiographic findings were factored into favouring DPE over TAE: [1]Tumor location: endonasal or intra-oral or intra-aural; soft-tissue tumor <=5 inches from skin surface; vertebral body tumor with a preserved pedicle [2] Parasitized arteries: small caliber (<0.5mm) and/or sever tortuosity precluding superselective catheterization, and presence of >=4 pedicles [3] Dangerous collaterals to the intracranial circulation, and [4] Minimal or sub-total residual parenchymal opacification after TAE.
TAE Technique
Superselective catheterization of the parasitized artery was performed to look for dangerous intracranial collaterals and assess reflux potential. Typically, a 1.6F or 1.3F tip microcatheter (Headway™ Duo 156cm and 167cm respectively, Microvention, Aliso Viejo, CA) was used; with both the contrast injection and subsequent embolization performed using the same type of syringe (i.e., either a regular 1cc luer-lock syringe or 1cc high-pressure polycarbonate syringe) to ensure a visually comparable reflux assessment between contrast and embolic injections. Per the discretion of the treating neurinterventionalist, a 1.6F tip balloon-microcatheter (Scepter™ Mini, MicroVention™, Aliso Viejo, CA) was used to minimize reflux. The liquid embolic was injected through a 1cc syringe under constant negative roadmap until complete stasis of flow within each feeding pedicle was achieved.
DPE Technique
The percutaneous site was prepped and draped using standard sterile technique prior to obtaining arterial access. A prophylactic antibiotic (e.g., Cefazolin) was administered 30-minutes before DPE. Once baseline angiography was completed, the diagnostic catheter was parked in the parent vessel (e.g., ipsilateral external carotid artery for patients with juvenile nasoangiofibromas). Approximately 80 international units (IU) per kilogram of unfractionated Heparin was administered intravenously, with an additional 1000 IU administered every 60-minutes thereafter. At least three 22-gauge Quincke-tip spinal needles (3.5” or 5” long) were prepped by removing their stylet’s and attaching a 5” Microbore Extension Set. Under live roadmap guidance, a prepped 22-gauge spinal-needle was advanced into the tumor, with blood-return suggestive of penetration into the tumor-bed. Next, a ‘parenchymogram’ was performed by injecting contrast into the tumor using a 3-cc syringe (Figure: 1). Maintaining anionic precautions, the dead space of the needle and Microbore tubing was then primed with DMSO or 5% Dextrose using a 1-cc syringe, followed by injection of Onyx or n-BCA respectively under blank roadmap guidance. While exchanging for the embolic syringe, the Luer lock extension tubing was occluded to avoid reflux of blood, which may otherwise result in precipitation of embolic within the needle. The duration of the injections varied from a few seconds to several minutes. If the embolic appeared to enter a vein or an arterial pedicle, we stopped the injection and waited ∼30 seconds to 1 minute. This was repeated until the embolic was able to penetrate the desired segment of the tumor. The injection was stopped if there was significant retrograde filling of an arterial pedicle, or a venous outflow vessel, or dangerous intracranial anastomoses. Intermittent angiography was performed to determine the extent of devascularization and to target the next neovascular compartment. At the end of each embolization, the Microbore tubing was disconnected and the stylet reintroduced prior to removal of the needle. If the embolic material stopped advancing or if there was intratumoral interstitial extravasation, multiple needles were used to maximize embolization within the different tumoral compartments.
With either technique, the choice of the liquid embolic agent was left to the discretion of the treating neurointerventionalist. The degree of residual parenchymal staining of the tumor on biplane angiography at the end of the treatment was graded by two neurointerventionalists as being: total (99%-100%), near-total (90%-98%), sub-total (50%-89%) or minimal (<50%). Total or near-total obliteration was considered a technical success.
Results
A total of 55 patients underwent direct percutaneous embolization. There were 37 (67.3%) males, with a median age of 46.5 (27-65) years. Approximately, 14 (25.5%) patients had a medical history of hypertension and 5 (9.1%) had diabetes mellitus. The most common locations for DPE were endonasal/nasopharynx (n=21), followed by scalp/subcutaneous tissue (n=18), carotid body (n=6), spinal/paraspinal (n=4), skull base (n=5), and intracranial (n=1). The most commonly embolized lesions were malignant carcinomas (n=16), followed by juvenile nasopharyngeal angiofibromas (n=11), paragangliomas (n=8), and hemangiomas (n=7). The most common malignant carcinoma was spinal metastatic disease from renal cell carcinoma (n=5), with papillary thyroid carcinoma (n=2), adenocarcinoma (n=2), squamous cell carcinoma (n=2) and sarcoma (n=2) amongst others (Figure: 2). A median of 4 (interquartile range 4-6) needles were used per embolization with a median fluoroscopy time of 50.5 (23.2-77.8) minutes. The median radiation dose per patient was 3055 (840.5-5053.5) mGy. Seven patients received DPE with more than one embolic agent. The most commonly used agent was n-butyl cyanoacrylate (glue, n=44, 81.5%), followed by ethylene-vinyl alcohol copolymer (Onyx, n=7, 12.9%), 98% dehydrated ethanol (n=7, 12.9%), sodium tetradecyl sulfate (n=2, 3.7%) and poly-vinyl alcohol particles (n=1, 1.8%). Nineteen of these 55 patients also received transarterial embolization, with 12 (63%) of those performed after DPE. A median of 2 parasitized pedicles were embolized trans-arterially with a Headway Duo microcatheter being the most commonly chosen microcatheter (n=12), followed by Excelsior SL 10 (Target Therapeutics, Boston Scientific) microcatheter (n=5), and Scepter Mini dual-lumen balloon (Microvention, Aliso Viejo, CA) microcatheter (n=2). The median volume of embolic agent injected was 5.5(4-7.6) mL per patient. Total (99%-100%), near-total (90%-98%) and sub-total (50%-89%) reduction in angiographic tumoral-blush was obtained in 20%, 54.5%, and 25.5% cases respectively. Fifty-one patients (93%) achieved atleast 80% reduction in tumoral vascularity. The technical success rate was highest in patients with juvenile nasoangiofibroma (n=11, 100%), followed by schwannoma (n=2, 100%), paraganglioma (n=7, 88%), neurofibroma (n=4, 75%), malignant carcinoma (n=16, 62.5%), and hemangioma (n=7, 57.1%). Of the 55 patients receiving DPE, 50 were followed by surgical resection, with three patients treated palliatively for refractory blood-loss anemia, and two undergoing treatments for localized pain. The median interval between embolization and surgical resection was 1 (1-4) day. The median operative blood-loss was 250(75-700) mL, with 12 patients (22.2%) receiving a postoperative blood transfusion. Only one of the 55 patients undergoing DPE had to be taken for emergent tumor resection due to uncontrolled tumoral hemorrhage. This patient had a large recurrent cerebellar hemangioblastoma which was embolized by advancing a 22-gauge spinal-needle directly through a previously placed sub-occipital titanium mesh (Figure: 3). The other complication from DPE included retrograde extension of n-BCA into parasitized vertebral artery feeders resulting in a clinically asymptomatic vertebral artery occlusion (Figure: 4). While one patient with a Jugular paraganglioma suffered a peripheral cranial-nerve VII palsy; however, this was likely due TAE in which Onyx-18 was injected under proximal balloon-occlusion in the posterior auricular artery resulting in devascularization of the trans-mastoid segment of the seventh cranial nerve without any identifiable drop-out in electromyographic activity during the embolization.
Discussion
In this series, which is the largest to-date describing the use of percutaneous embolization for head, neck and spine lesions, we find it to be safe (1.8% major complication rate) and effective (93% achieving 80% or higher angiographic devascularization). Further, this is the largest series to report DPE experience with n-BCA (used in over 80% cases) via a myriad of approaches (trans-cutaneous, trans-mucosal, trans-aural, per-aural, per-oral, trans-pedicular and trans-calavarial). Finally, we add to the currently limited experience of trans-calavarial DPE and summarize the relevant literature.
It is now well-established that preoperative trans-arterial embolization of hypervascular head, neck, and spine tumors reduces operative times and blood-loss.1 Several agents have been utilized over the years, driven largely by operator preference, with liquid embolic agents (e.g., Onyx and n-BCA) being increasingly utilized over particles given the risk of injury to vasa-nervosum and delayed recanalization.1,6 However, parasitized arterial feeders can occasionally be very tortuous making catheterization difficult. Some tumors may also have a high degree of arterio-venous shunting, further obscuring its arterial anatomy.2,8 In addition, dangerous external-carotid to internal-carotid or vertebral artery anastomoses may develop in tumor-beds making trans-arterial embolization high-risk.1,2 To bypass these limitations, direct percutaneous embolization was initially introduced in 1994 as an alternative approach to devascularize tumors,2,7 with several case-series since providing retrospective data on the safety and efficacy of this technique.6-11 Interestingly, this early experience with DPE has also shown that it results in greater penetration into the capillary-bed of these tumors when compared to TAE, wherein occlusion of the proximal feeders may result in a high-degree of angiographic devascularization without any significant reduction in operative blood-loss as the capillary-bed may remain ‘alive’ via collateral pathways (Figure 2, D).1,8 Thus, TAE may not always be achievable, and even when it is feasible, its use alone may result in radiographic devascularization without conferring a net clinical benefit .1,8
Several embolic agents have been used for DPE including Onyx, n-BCA, SQUID-12, and Polyvinyl alcohol or trisacryl gelatin microspheres, with particles found to be associated with a higher risk of cranial neuropathies from devascularization of vasa-nervosum.1 In our series, we predominantly (81.5% cases) used n-BCA diluted to a 25% concentration by diluting it with Lipiodol (i.e., 0.9mL of n-BCA diluted in 2.7mL of Lipiodol). While we did use Onyx-18 in ∼13% cases, we found that low-concentration n-BCA allowed for a more centripetal distribution of embolic material within the tumor-bed. Further, lower concentrations of n-BCA have higher viscosity allowing for more controlled and pronged injections. In addition, it’s almost immediate polymerization could become useful in cases of extravasation.5,6 Despite its rapid polymerization, we did not find any significant increase in the rate of needle-occlusion with n-BCA when compared to Onyx-18.
The radiographic goal of pre-operative embolization is generally considered to be atleast 80% devascularization on final angiography,1 which was achieved in 93% cases in our series with ∼75% achieving total (99%-100%)/near-total (90%-98%) angiographic devascularization. This is comparable to other single institutional studies that have shown similar rates of devascularization with DPE of head and neck tumors,1-3,8-11 vascular malformations,12 and spinal tumors.13
Intra-operative needle puncture and embolization of intracranial tumors at the time of their resection has been well described.14,15 However, only a handful of case-reports have described transcalavarial puncture for DPE of intracranial tumors (Table 1). Seven of the eight previously reported cases of transcalavrial DPE have utilized prior craniectomies from aborted resection attempts 3-5,16,17 or pathologic bony erosion;5 while one patient with an infratentorial endolymphatic-sac tumor underwent a retroauricular craniotomy to facilitate DPE with subsequent tumor resection via the same craniotomy.5 Among the three reported cases of transcalavrial DPE performed for infratentorial tumors,4,5 one patient had a cerebellar hemangioblastoma while the other two patients had petro-temporal endolympatic-sac tumors; undergoing Onyx-34, 15% n-BCA and 15% n-BCA embolization resulting in 50%, near-total (>90%) and total tumoral devascularization respectively. In our series, we treated one patient with a recurrent cerebellar hemangioblastoma with DPE by advancing a 22-gauge spinal-needle (under bi-planar fluoroscopic/roadmap guidance) via a previously placed titanium mesh (Figure 3). We used 25% n-BCA resulting in nearly 60% devascularization after four punctures. Similar to our patient (Figure 3), the only other previously reported case of DPE for a cerebellar hemangioblastoma, was also immediately resected in the operating room where the non-embolized residual was noted to bleed profusely resulting in an estimated blood-loss of 1000mL.4 However, unlike the prior case where the tumor was homogenous and solid (type IV hemangioblastoma), our patient had a microcystic tumor with multiple gadolinium-enhancing cysts (type-III hemangioblastoma).4,18 Cysts with enhancing walls have been shown to contain neoplastic cells, with dilated and necrotic vascular spaces on histologic examination, and suffering higher intraoperative blood-loss than the most common type (i.e. type II) hemangioblastomas.18,19 The authors believe that the high degree of arteriovenous shunting within the neoplastic cystic walls, may have resulted in embolization of n-BCA into the tumor’s primary draining vein (superior vermian) and the straight sinus, turning the residual non-embolized tumor into a functional arteriovenous malformation with its venous outflow obstructed, causing it to bleed immediately post-embolization. This required emergent tumor decompression and resection with an estimated blood-loss of 1000mL.
The other DPE related complication encountered in our series was immediate retrograde extension of 25%n-BCA via the parasitized arterial feeders into the left vertebral artery in a patient with co-dominant vertebro-basilar circulation (Figure 4). The patient remained clinically asymptomatic. However, post-capillary retrograde extension of embolic agents has been reported to result in significant clinical impairments, usually immediately post-embolization while the glue is still in a liquid state.11 However, delayed embolization thought to have been caused by fragmented, polymerized glue has also been reported.20 While some have used compliant balloons within the parent vessel (e.g., internal carotid or vertebral arteries) during DPE for this particular reason; the recurrent inflation of these intravascular polyurethane balloons may also confer an additive thromboembolic risk, with some reporting migration of embolic agents despite balloon-inflation.20
Conclusion
Direct percutaneous embolization of head, neck, and spine tumors is safe and effective. Our single-center study reinforces prior experience that DPE of sino-nasal carcinomas, angiofibromas and paragangliomas with adhesive and non-adhesive liquid embolic agents achieves good penetration of capillaries within the tumor-bed significantly reducing operative blood-loss. Further, it suggests that these benefits may also be extended to non-traditional head, neck and spine lesions such as schwannomas, neurofibromas, and hemangiomas. Caution must be exercised when applying these techniques to intracranial tumors with high intratumoral arteriovenous shunting.
Data Availability
All data produced in the present study are available upon reasonable request to the authors
Acknowledgements
None
Footnotes
Sources of Funding: None
Conflict of Interests: B.D.J is a consultant for Microvention, Medtronic and Stryker. On behalf of all other authors, the corresponding author declares that there are no competing conflicts of interest.
Abbreviations
- CTA
- computed tomography angiogram
- JNA
- juvenile nasopharyngeal angiofibroma
- n-BCA
- n-butyl cyanacrylate
- MCA
- middle cerebral artery
- TAE
- Trans-arterial embolization