Temporary balloon occlusion and ethanol injection for preoperative embolization of carotid-body tumor. (Original Article).

Abstract

We report on the preoperative embolization of a carotidbody paraganglioma by temporary balloon occlusion and ethanol injection. Complete devascularization was achieved without complication. Resection after a short postembolization interval required artery sacrifice. Histologic evaluation revealed that the tumor contained diffuse ethanol-induced microemboli. Compared with unembolized and polyvinyl-alcohol-embolized carotidbody paragangliomas, our technique resulted in no greater adverse effects on the tumor-vessel interface. This procedure is an effective and promising method of preoperative embolization of carotid-body tumors and warrants further experience and study. In this article, we also review the literature on carotid-body tumor embolization and ethanol embolization.

Introduction

The carotid body is a chemoreceptor located in the adventitia posteromedially to the bifurcation of the common carotid artery. Paragangliomas of the carotid body arise from embryonic neuroepithelium. These slowly growing, benign, hypervascular lesions typically arise in the fourth through sixth decades of life as painless, laterally mobile masses. Larger lesions might cause lower cranial nerve palsies, pulsatile bruits, and/or tonsillar, soft palate, pharyngeal, and uvular displacement or erosion. They are more common among women than men (ratio: 5:1).

The relationship of the tumor to the surrounding vasculature is classified according to the Shamblin system. Shamblin type I tumors do not encase the external or internal carotid arteries, and they are easily dissected free. Type II lesions partially surround the carotid arteries. Type III lesions completely encase the arteries and are the most difficult to resect without causing vascular or neural injury.

Case report

A 37-year-old woman came to the office for evaluation of a tender, 4-cm right cervical mass. Findings on neurologic examination were normal. Magnetic resonance imaging detected a carotid-body tumor. Tests for urine metanephrines, vanillylmandelic acid, and total catecholamines were negative. Arteriography demonstrated that a hypervascular lesion was splaying apart the external and internal carotid arteries (figure 1). The blood supply appeared to arise primarily from the external carotid artery. A right temporary balloon-occlusion test and an evaluation of xenon-contrast cerebral blood flow revealed that the patient was able to tolerate sacrifice of the internal carotid artery, if necessary.

One day before surgery, the patient underwent alcohol embolization of the numerous hair-like arterial feeders that had arisen from the external carotid artery. An 8F sheath was placed in the patient's right common femoral artery, and an 8F guiding catheter was advanced into the right common carotid artery. Two nondetachable silicon balloons were then placed through the guiding catheter; one was advanced into the internal carotid artery and placed just above its origin, and the other was advanced into the external carotid artery and placed just distal to the superior thyroid artery. The patient was heparinized with 10,000 U, and the balloons were inflated (figure 2). The patient was examined and 30mg of lidocaine was injected into the external carotid artery. No neurologic changes were noted. A 0.018-in microcatheter was then advanced into the external carotid artery, and 1.8 ml of 100% alcohol was infused by hand over approximately 15 minutes. Repeat angiography with the balloons deflated demonstrated comple te tumor devascularization and normal carotid parent vessels (figure 3).

The following day, the patient underwent surgical resection of the Shamblin type II lesion. The dissection was difficult to perform because of the partial tumor encasement; as a consequence, the common, internal, and external carotid arteries ultimately had to be sacrificed. A venous interposition graft was not placed because the patient's earlier balloon-occlusion test demonstrated that she was able to tolerate carotid sacrifice. The patient was maintained for 24 hours postoperatively on therapeutic heparin and aspirin. She was discharged on aspirin alone in stable condition. No neurologic deficits were noted.

To evaluate the effects of alcohol embolization histologically, the resected specimen was compared with two other recently excised carotid-body paragangliomas. One of these two specimens had been embolized with polyvinyl alcohol (PVA) prior to resection, and it was selected to compare the nature of the different embolic materials. The second specimen was taken from a patient whose artery was sacrificed but who had not undergone embolization, and it was selected to ascertain the effects that ethanol embolization had on the adherent sections of the artery in the case patient.

The resected lesion taken from this case measured 2.6 x 2.3 x 2.2 cm and was adherent to the artery. Neither the adherent segment nor the separately submitted arterial segments showed any gross abnormality. The entire course of the artery's attachment to the tumor was submitted for microscopic examination.

Sections of the tumor showed that the tumor cells were arranged in the typical Zellballen architectural pattern (figure 4). The neoplastic cells contained abundant, finely granular eosinophilic cytoplasm with smooth-contoured round to oval nuclei. The cells were synaptophysin-positive, and the sustentacular cells were highlighted on S-l00 and glial fibrillary acidic protein staining. In many areas, dense stromal fibrosis compressed the architecture into cords.

Many small vessels in the tumor contained foreign material that was consistent with the effects of alcohol embolization (figure 4). This material was pale blue-gray, slightly refractile, and most easily seen with increased contrast. Admixed with this material were inflammatory cells and fibrin, predominantly in strands. These foci were irregular in shape and seemed to conform to the vessel contours. The diameter of the affected vessels was in the range of 50 to 100 [micro]m. Of note, these foci did not appear on Verhoeff-van Gieson staining as PVA particles have been reported to do. (1) There were many small foci of tumor necrosis, often adjacent to embolized vessels (figure 4).

The foci of embolic material found in the PVA-embolized specimen were similar to previously published descriptions of this material (figure 4). (2) These particles demonstrated the same staining quality as did the ethanol emboli, and they contained the characteristic fibrin and inflammatory cells adherent in the pores. These foci retained the spherical shape of the particles and were much larger (1 to 1.5 mm). Areas of tumor necrosis were present, and they tended to be larger than those in the ethanol-embolized specimen. No inflammatory response to the thrombi in the tumor was seen in either case, both of which had been resected after a short postembolization interval.

The unembolized specimen showed the usual reported relationship of the tumor to the vessel (figure 5). (3) The tumor was adherent to the adventitia of the vessel, and there was no invasion of the media. No inflammatory infiltrates or stromal activity were seen at this interface. The tumor-vessel interface in our ethanol-embolized specimen was similar (figure 6). The tumor apposed the loose adventitial connective tissue, with no inflammatory infiltrates or stromal reaction present. However, changes were seen on the luminal surface. The endothelium was absent, and the intimal layer of the vessel was compacted and densely eosinophilic, a state consistent with either a chemical, thermal, or physical effect. This finding was in contrast to the intact endothelium and loose intimal layer seen in figure 5. An elastic stain emphasized the persistence of intact elastic lamina immediately below the luminal surface (figure 7). This stain also emphasized the absence of this effect in the deeper layers adjacent to the tumo r interface. This effect was absent in the separately submitted segment of artery that was sacrificed.

Discussion

Carotid-body tumors are rare. In their 1988 Mayo Clinic review of 59 tumors in 55 patients who underwent surgical resection, Nora et al reported that complete surgical excision was achieved in 52 of these patients (94.5%). (4) Only one patient died during the 59 procedures (mortality rate: 1.7%), secondary to a cerebrovascular accident. Perioperative neurologic deficits occurred in 5 of the 52 patients (9.6%) who underwent a complete resection. Eleven of the 52 (21.2%) experienced permanent cranial nerve dysfunction, mostly of the vagus nerve; 11 other patients experienced a temporary cranial nerve dysfunction. Three patients (5.8%) experienced recurrences after what was erroneously believed to have been a complete resection.

Preoperative embolization of carotid-body tumors and other paragangliomas has been reported in the literature. The advantage of this therapy is that the devascularized lesion can be removed relatively bloodlessly, which facilitates visualization of surrounding structures and lessens the need for transfusion of biologic products. Most authors have concentrated on embolization with particulate matter (e.g., PVA and Gelfoam) or glues (e.g., cyanoacrylates). (5-13) A few reports of alcohol embolization have concentrated on nonglomus tumors, with most of the alcohol introduced via a percutaneous or direct intraoperative tumor puncture route). (14-23)

Embolization of carotid-body tumors is not without its complications. Injury to the lower cranial nerves can occur when embolic material enters the vessels that supply these neural structures. Deficits can also occur through anastomotic connections, such as those between the internal maxillary and ophthalmic arteries and between the occipital and vertebral arteries. Horton and Kerber recommended testing external carotid branches with lidocaine prior to tumor embolization. (16) They studied 26 patients who underwent external carotid artery embolization of head and neck tumors and found that three (11.5%) developed transient palsies, which required therapeutic modifications. None of these three patients experienced permanent complications.

Our experience with transarterial alcohol embolization of meningiomas and dural fistulas has been positive. Alcohol is a useful sclerotic agent that allows the intervention-alist to devascularize a lesion, the feeders to which are too numerous or small to directly catheterize or to admitparticulate matter. Alcohol also has the unique ability to permeate a lesion and sclerose its vessels from within. In our case, the use of two balloons to occlude anterograde flow in the internal and external carotid arteries allowed the alcohol to permeate the tumor. Once the balloons were deflated, any remaining alcohol was immediately diluted and rendered harmless. Particulate material lacks this type of dilutional property. In addition, by minimizing the dead space in the internal carotid artery (by placing one balloon at its origin), we were able to both minimize endothelial exposure to the alcohol and to keep our total alcohol dose low (1.8 ml).

Our case illustrates a unique requirement among embolization targets: the dissection of the target tumor from the parent artery used to embolize it. There have been two major findings reported in ethanol embolization that are of concern in this clinical setting. The first is a reported inflammatory response with obliteration of tissue planes. The second is ethanol's direct tissue toxicity, both to the parent arteries and the target organs. These issues are dramatically affected by two variables: (1) the length of the interval between embolization and resection with regard to the former finding and (2) the dose and rate with regard to the latter finding. Our histologic findings confirm that we have effectively addressed both issues with our approach.

We demonstrated that the short amount of time between embolization and resection precluded the development of any inflammatory or stromal response to either the embolic material or the necrotic foci in either the ethanol- orPVA-embolized cases. This is consistent with previous experiments with PVA, which found little response to the embolic material at early time points.' This validates the approach taken at our institution, where surgery is performed within 24 hours of embolization to minimize the potential for a significant inflammatory reaction.

Injection of ethanol appears to cause two distinct phenomena: (1) direct tissue necrosis and (2) the creation of microemboli by interaction with blood components. (21,24-28) The tissue necrosis is contact-mediated and seen very soon after infusion. (21,27) At faster dosing rates, rapid necrosis can be seen in large parent arteries, visible histologically as transmural necrosis. This tissue toxicity can also be seen in the target organ, as parenchymal damage occurs too quickly to permit ischemic damage. Smaller doses can reduce parenchymal damage while maintaining complete devascularization. (26,27) Damage can also be minimized by either slower dosing rates or ethanol dilution. (25)

In the experimental and clinical studies cited earlier, most of which involved kidney and liver targets, dosing rates were generally between 0.01 and 0.2 ml/kg/sec. In contrast, the dosing rate in our study was 4 x [10.sup.-5] ml/kg/sec. Our rate appears to have prevented histologically visible damage to the media or adventitia-tumor interface of the parent artery. However, we did see endothelialintimal damage and condensation with diluted alcohol, as described by Sampei et al. (25) It is unclear whether this represents a direct alcohol effect, a thermal effect during the course of dissection, or a pressure effect from the balloon. In any case, intact elastic lamina and abroad zone of unaffected vessel wall could be seen subjacent to this area. We conclude that there are no histologic changes present that would likely predict sacrifice of the artery. We also note that in the study by Sampei et al, mild changes were seen in the smooth muscle of the media on electron microscopy. Wether these changes were prese nt in our patient--and if so, whether they would be of clinical significance--is unknown.

The creation of microscopic emboli during slow infusion by the interaction of ethanol and blood elements in situ or at the catheter tip was proposed by Ellman et al in their early study of renal devascularization. (21) The sections of our target tumor showed evidence of these microemboli. It is interesting that these microemboli resemble much smaller versions of PVA particles. These microemboli were widely distributed, which is consistent with the complete devascularization seen on angiography. The areas of necrosis associated with these microemboli were small, and nowhere did they extend to the tumor pseudocapsule or vessel interface. The PVA-embolized control case had larger areas of necrosis despite a similar interval between embolization and surgery. In our previous experience with ethanol embolization of skull base meningiomas, we noted scattered areas of necrosis but were unable to identify the emboli themselves. (29)

In summary, a small dose and a slow infusion rate resulted in complete devascularization of the tumor by microscopic ethanol-containing thromboemboli. The tumor was marked by limited necrosis, which had no histologically discernible effect on the tumor-vessel interface. The parent artery showed only endothelial-intimal damage, which was related to either the embolization or possibly the surgery itself. Overall, this represents a favorable histopathologic outcome considering that the artery had to be sacrificed surgically. It should be emphasized that in unembolized cases in which sacrifice is necessary--such as in the control case described here--there are no histologic features that would reliably predict that outcome. It is therefore ultimately unclear whether the use of alcohol in this case led to the need for carotid sacrifice. There might have been some histologically inapparent damage that led to fragility of the internal and external carotid artery walls. More conclusive answers to the questions surrou nding the effects of preoperative embolization on glomus tumors await the results and observations of more rigorous study and more clinical experience.

References

(1.) Kepes JJ, Yarde WL. Visulization of injected embolic material (polyvinyl alcohol) in paraffin sections with Verhoeff-Van Gieson elastica stain. Am J Surg Pathol 1995;19:709-11.

(2.) Link DP, Strandberg JD, Virmani R, et al. Histophathologic appearance of arterial occlusions with hydrogel and polyvinyl alcohol embolic material in domestic swine. J Vasc Interv Radiol 1996;7:897-905.

(3.) Lack EE. Tumors of the adrenal gland and extra-adrenal paraganglia. In: Atlas of Tumor Pathology. 3rd series, Fascicle 19. Washington, D.C.: Armed Forces Institute of Pathology, 1995:321-42.

(4.) Nora JD, Hallett JW, Jr., O'Brien PC, et al. Surgical resection of carotid body tumors: Long-term survival, recurrence, and metastasis. Mayo Clin Proc 1988;63:348-52.

(5.) Muhm M, Polterauer P, Gstottner W, et al. [Glomus caroticum chemodectoma: Review on current diagnosis and therapy]. Wien Klin Wochenschr 2000;112:115-20.

(6.) Marangos N, Schumacher M. [Management of large bilateral glomus jugulare tumors]. HNO 1999;47:816-20.

(7.) Fruhwirth J, Koch G, Klein GE. [Preoperative angiographic embolization of carotid glomus tumors. A method for improving operability]. HNO 1996;44:510-3.

(8.) Lavold E, Freng AB, Nakstad P. [Combined embolization and surgery of paragangliomas of the glomus caroticum. Report of a case and review of literature]. Tidsskr Nor Laegeforen 1996;116:2449-51.

(9.) Feng J, Li B, Zhou D. [Preoperative embolization of paraganglioma in head and neck]. Zhonghua Wai Ke Za Zhi 1995;33:675-6.

(10.) George B, Casasco A, Deffrennes D, Houdart E. Intratumoral embolization of intracranial and extracranial tumors: Technical note. Neurosurgery 1994;35:771-3.

(11.) Murphy TP, Brackmann DE. Effects of preoperative embolization on glomus jugulare tumors. Laryngoscope 1989;99:1244-7.

(12.) Young NM, Wiet RJ, Russell EJ, Monscil EM. Superselective embolization of glomus jugulare tumors. Ann Otol Rhinol Laryngol 1988;97:613-20.

(13.) Iaccarino V, Sodano A, Belfiore G, et al. Embolization of glomus tumors of the carotid: Temporary or definitive? Cardiovase Intervent Radiol 1985;8:206-10.

(14.) Casasco A, Herbreteau D, Houdart E, et al. Devascularization of craniofacial tumors by percutaneous tumor puncture. AJNR Am J Neuroradiol 1994;15:l233-9.

(15.) Lonser RR, Heiss JD, Oldfield EH. Tumor devascularization by intratumoral ethanol injection during surgery. Technical note. J Neurosurg 1998;88:923-4.

(16.) Horton JA, Kerber CW. Lidocaine injection into external carotid branches: Provocative test to preserve cranial nerve function in therapeutic embolization. AJNR Am J Neuroradiol 1986;7:105-8.

(17.) Livraghi T, Festi D, Monti F, et al. US-guided percutaneous alcohol injection of small hepatic and abdominal tumors. Radiology 1986;161:309-12.

(18.) Solbiati L, Giangrande A, De Pra L, et al. Percutaneous ethanol injection of parathyroid tumors under US guidance: Treatment for secondary hyperparathyroidism. Radiology 1985;155:607-l0.

(19.) Valavanis A. Preoperative embolization of the head and neck: Indications, patient selection, goals, and precautions. AJNR Am J Neuroradiol 1986;7:943-52.

(20.) Tanaka K, Okazaki H, Nakamura S. et al. Hepatocellular carcinoma: Treatment with a combination therapy of transcatheter arterial embolization and percutaneous ethanol injection. Radiology 1991;179:713-7.

(21.) Ellman BA, Parkhill BJ, Curry TS III, et al. Ablation of renal tumors with absolute ethanol: A new technique. Radiology 1981;141:619-26.

(22.) Heiss JD, Doppman JL, Oldfield EH. Brief report: Relief of spinal cord compression from vertebral hemangioma by intralesional injection of absolute ethanol. N Engl J Med 1994;331:508-11.

(23.) Lippi F, Ferrari C, Manetti L, et al. Treatment of solitary autonomous thyroid nodules by percutaneous ethanol injection: Results of an Italian multicenter study. The Multicenter Study Group. J Clin Endocrinol Metab 1996;81:3261-4.

(24.) Craven WM, Redmond PL, Kumpe DA, et al. Planned delayed nephrectomy after ethanol embolization of renal carcinoma. J Urol 1991;146:704-8.

(25.) Sampei K, Hashimoto N, Kazekawa K, et al. Histological changes in brain tissue and vasculature after intracarotid infusion of organic solvents in rats. Neuroradiology 1996;38:291-4.

(26.) Yamakado K, Takeda K, Nishide Y, et al. Portal vein embolization with steel coils and absolute ethanol: A comparative experimental study with canine liver. Hepatology 1995;22:1812-8.

(27.) Buchta K, Sands J, Rosenkrantz H, Roche WD. Early mechanism of action of arterially infused alcohol U.S.P. in renal devitalization. Radiology 1982;145:45-8.

(28.) Ellman BA, Green CE, Eigenbrodt E, et al. Renal infarction with absolute ethanol. Invest Radiol 1980;15:318-22.

(29.) Jungreis CA. Skull-base tumors: Ethanol embolization of the cavernous carotid artery. Radiology 1991;181:741-3.

From the Department of Neurosurgery (Dr. Horowitz, Dr. Jungreis, Dr. Levy, and Dr. Kassam), the Department of Radiology (Dr. Horowitz and Dr. Jungreis), the Department of Pathology (Dr. Whisnant), and the Department of Otolaryngolgy (Dr. Snyderman and Dr. Kassam), the University of Pittsburgh Medical Center--Presbyterian University Hospital.

Reprint requests: Michael Horowitz, MD, Department of Neurosurgery, Suite B 400, 200 Lothrop St., Pittsburgh, PA 15213-2582. Phone: [phone]; fax: [phone]; e-mail: [e-mail]