Stereotactic linear accelerater-based radiosurgery for the treatment of patients with glomus jugulare tumors




The optimal treatment for patients with glomus jugulare tumor (GJT) of the skull base remains controversial. Surgical excision is associated with a high incidence of cranial nerve injury, decreased quality of life, and high mortality. Fractionated radiotherapy is used to control the majority of these tumors, but disadvantages are a prolonged therapy interval and exposition of adjacent brain tissue to irradiation. The authors present the results of a study on 12 of 14 consecutively admitted patients who were treated using linear accelerator-based radiosurgery (LINAC-RS), an innovative method for the treatment of GJT.


From May 1991 to March 2001, 14 patients with GJT were treated with stereotactic LINAC-RS for continued growth of tumor or of remaining tumor after surgery. Twelve patients (9 women and 3 men; age range, 28–71 years; median age, 59 years) with a median follow-up of 4 years (range, 0.8–9,0 years), were selected for retrospective analysis. A median single dose of 15 grays (Gy; range, 11–20 Gy) was applied to the surface of the tumor.


After undergoing LINAC-RS, 8 of 12 patients (67%) reported partial or complete subjective improvement, whereas complaints remained unchanged in 4 patients (33%). Neurologic status improved in 3 patients (25%) and remained unchanged in 8 patients (67%). Magnetic resonance images showed tumor shrinkage in 8 patients (67%) and no further progression in 4 patients (33%).


LINAC-RS is an effective and safe therapy for patients with GJT and may be used as an alternative to surgical resection. Compared with fractionated radiotherapy, LINAC-RS has some advantages. However, to clarify the question of long-term tumor control, longer observation times are required. Cancer 2003;97:1093–98. © 2003 American Cancer Society.

DOI 10.1002/cncr.11118

Glomus jugulare tumor (GJT)—also known as paraganglioma or chemodectoma—is a rare tumor with an incidence of about 1 per 1,000,000 population.1 It arises in the chemoreceptoric, nonchromaffine paraganglia and is the most common tumor of the middle ear. These tumors usually are benign and slow-growing but destructive. In about 3% of patients, the tumor is malignant.2, 3 Typically, these tumors manifest themselves between the third and the sixth decades of life. Women are affected five times more often than men.4, 5 GJTs have their epicenter in the region of the jugular bulb. Due to their destructive growth, they may invade the tympanic cavity and the jugular foramen, may expand intracranially (leading to serious cranial nerve injury), and even may become life-threatening. Generally, patients with GJTs exhibit tinnitus; cephalalgia; hypacusis; anacusis; dizziness; or multiple injury of cranial nerves VII, VIII, IX, X, XI, and XII.

Because of the characteristic bone destruction associated with GJT, diagnosis also can be made from standard radiologic images (Schüller or Guillen transorbital projection) and computed tomography (CT) scans,6 although magnetic resonance imaging (MRI) remains the gold standard. Aside from providing information about the dimensions of the tumor, MRIs show the soft tissue borders of the tumor and its anatomic relation to the brainstem. On plain T1-weighted series, the typical MRI finding is inhomogeneous and hypointense signal with strong gadolinium (Gd-DTP) enhancement. On T2-weighted images, the tumor shows high signal intensity. Digital subtraction angiography (DSA) adds important information by displaying tumor blush with typical vascular supply in most instances and may be used to identify other potential head and neck paragangliomas. Catheter angiography allows preoperative superselective tumor embolization.

The standard treatment for patients with GJT is surgery that eventually is facilitated by endovascular embolization. Percutaneous fractionated radiotherapy is reserved for patients with inoperable tumors.

The objective of this retrospective study was to evaluate an innovative therapeutic approach using LINAC-RS for treating patients with GJT. Evaluation was focused primarily on assessment of tumor control rates and treatment-associated complications after radiosurgery. The results were compared with those of other studies on conventional surgery and fractionated radiotherapy.


From May 1991 to March 2001, 14 consecutively admitted patients with GJT were treated at our institution using stereotactic LINAC-RS for continued growth of tumor or of remaining tumor after surgery. Prior to radiosurgery, six patients underwent tumor resection, and five patients underwent embolization. Twelve of 14 patients (9 women and 3 men; age range, 28–71 years; median age, 59 years) with a median follow-up of 4 years (range, 0.8–9.0 years; 8 patients had ≥ 4 years of follow-up) were selected for retrospective analysis. This group underwent a complete series of follow-up MRI examinations. One patient was excluded from the study due to a short follow-up (3 months); and another patient was excluded because, 6 months after radiosurgery, she underwent tumor resection at a different institution, although she showed no new clinical symptoms, and her tumor had not progressed. Data on initial symptoms; physical, audiometric, and neuroradiologic examinations; radiosurgery planning data; and postoperative follow-up data were obtained from clinical records and from a questionnaire that was mailed to the patients. Evaluation of follow-up MRI was done together with the neuroradiologist (P.L.). Tumor size was estimated individually by measuring the greatest tumor dimension on transverse, coronal, and sagittal reconstructions; and tumors were graded according to the criteria of Fisch (Table 1).7

Table 1. Fisch Classification of Glomus Tumors
ALimited to middle ear cleft
BLimited to the tympanomastoid area
CInvolving the infralabyrinthine compartment and petrous apex of the temporal bone
D1Intracranial extension < 2 cm in greatest dimension
D2Intracranial extension > 2 cm in greatest dimension

Except for 1 patient who had a Type C tumor, all patients exhibited a Type D1 or D2 tumors. Ten of 14 lesions extended below the skull base and into the jugular foramen.

The most common symptoms, which were seen in 64% of patients, were hearing loss and pulsatile tinnitus followed by aural pressure/fullness/pain or headache (57%). Table 2 provides a comprehensive breakdown of all signs and symptoms. The baseline Karnofsky performance index ranged from 60% to 90%.

Table 2. Glomus Jugulare Tumors: Symptoms and Signs Shown by 12 Follow-Up Patients
Symptoms and signsNo. of patients
Before treatmentClinical improvement after therapy
Hypacusis or anacusis112 of 11
Pulsatile tinnitus84 of 8
Aural pressure/fullness/pain, headache86 of 8
Weakness of cranial nerves V, VII, IX–XIII54 of 5
Vertigo/dizziness75 of 7

Regarding LINAC-RS, a median single dose of 15 grays (Gy) (range, 11–20 Gy) was applied to the surface of the tumor (median tumor volume, 12.2 mL; range 4.4–57.8 mL), and the median maximum dose was 39 Gy (range, 17.2–47.5 Gy). The median number of isocenters was 4 (range, 1–7 isocenters). The dose was chosen due to the highly vascularized nature of the tumor. Thus, as with arteriovenous malformations, doses of 15–20 Gy were delivered to the surface of the lesion. However, when including the results of our earlier risk analysis,8 the dose had to be lowered due to accommodate individual anatomic variations of the patients.

The hardware and software components used in the treatment planning and irradiation procedures have been published.9–16 Figure 1 shows an example of a treatment plan.

Figure 1.

Treatment planning for a patient with a glomus jugulare tumor (GJT) on stereotactic computed tomography scans and a magnetic resonance image that were integrated using image fusion16 (STP3; Stryker Leibinger, Freiburg, Germany). Displayed are the tumor contour (red dots) together with the therapeutic isodose (15 grays [Gy]; blue line) and the 13-Gy isodose line (green line).


Eight of 12 patients examined (67%) reported partial or complete improvement of one or more of their subjective symptoms (tinnitus, headache, impaired cranial nerve function) (Table 2). Symptoms remained unchanged in 4 of 12 patients (33%). No acute complications related to radiosurgery were observed. An objective improvement of neurologic status after LINAC-RS was seen in for 3 of 12 patients (25%), and the neurologic status was unchanged in 8 of 12 patients (67%). One patient who had a D2 tumor that was treated with a margin dose of 20 Gy (normalized to the 80% isodose) in 1991 developed moderate facial palsy (House/Brackmann Grade 2) 6 months after LINAC-RS.

The follow-up MRIs show that tumor shrinkage (25–74% volume reduction compared with the preradiosurgery status) was observed in 8 of 12 patients (67%), and there was no tumor progression (local tumor control) in 4 of 12 patients (33%). Images from a typical patient are shown in Figure 2.

Figure 2.

Magnetic resonance images (MRIs) from a patient age 57 years with a glomus jugulare tumor (GJT; type D1). Three years after the patient underwent linear accelerator-based radiosurgery (therapeutic dose, 12 grays [Gy]), MRIs before treatment (left) and after treatment (right) showed substantial tumor regression (33% reduction of volume compared with baseline).


The current results show the high efficacy and safety of LINAC-RS with moderate single doses for treating patients with GJT. Five of 12 patients examined had been followed for > 6 years. With a comparably high median follow-up of 48 months, all patients exhibited either tumor shrinkage (8 of 12 patients) or stable disease (4 of 12 patients). There was no acute morbidity nor mortality. Tumor-related complaints improved in 8 of 12 patients (67%). Only one patient developed mild facial nerve palsy 6 months after radiosurgery. In our opinion, this complication was caused by the comparably high single margin dose of 20 Gy applied to a tumor that was located close to the facial nerve.

The optimal treatment strategy for patients with GJT invading the skull base remains controversial. Due to the high vascularization of these lesions and the complex anatomy associated with their particular localization, excision of GJT is difficult. It has been reported in the literature that blood loss during tumor resection ranges from 1.7 liters to 2.8 liters.17, 18 In both of those studies, presurgical endovascular embolization was capable of reducing blood loss by about 60%,17, 18 although the morbidity rate remained at 11.2%.17 In general, despite dramatic improvements in surgical techniques, neuroradiology, and intraoperative facial nerve monitoring, morbidity associated with surgery is high. In 1994, Green et al.19 reported that quality of life was reduced significantly in > 15% of patients after they underwent resection. In 31–81% of their patients, postsurgical destruction of cranial nerves IX–XII was seen, with 19% of patients requiring repair of paralyzed vocal cords and 4% of patients requiring gastrostomy. Neuropathy of the facial nerve with the need for tarsorrhaphy was seen in 16–23% of those patients, the VIth nerve was affected in 6% of patients, and hearing impairment or deafness occurred in 4–26% of patients. Mortality rates for patients with GJT range from 4.0% to 6.4%,19–22 and 16–18% of patients develop recurrent GJT after surgery.19, 21–24

Another treatment option for patients with inoperable GJT is percutanous fractionated radiotherapy. In one series, the morbidity rate after radiotherapy was 7%.25 Severe complications included temporal bone osteoradionecrosis, necrosis of nerve tissue, and tumor hemorrhage.23–25 Furthermore, radiation-induced second malignancies also have been reported.26 Local tumor control was achieved in 85–94% of patients who were treated with a radiation dose of 45–55 Gy.25, 27, 28

GJT treatment with radiosurgery is a new therapeutic approach. According to the paradigm of stereotactically guided, single-session, and highly focused percutaneous irradiation, as coined in 1951 by the pioneer neurosurgeon Lars Leksell,29 the precise stereotactic localization of the target point, combined with a steep dose gradient outside the target volume, allows the application of high single doses to a lesion without concomitant damage of adjacent, normal tissue, with the objective of selectively necrotizing or devitalizing small, well-delineated intracranial lesions.30 Compared with radiosurgery, fractionated percutaneous radiotherapy requires a longer treatment time: on average, about 5 weeks. Larger fields must be applied to compensate for immobilization and set-up errors. Due to field size and the application dose, repeat treatment, as a rule, is not possible. Furthermore, in most patients, a significantly excessive dose is applied to the normal brain, soft tissues, and bone tissues adjacent to the area to be irradiated. The carcinogenic potential of normal tissue irradiation also is a concern in these patients, who often are young. Since 1997, three studies have assessed the efficacy of radiosurgery for patients with GJT. In 1997, Foote et al.31 presented nine patients who underwent γ-knife radiosurgery. After a median follow-up of 20 months, 8 of 9 tumors remained stable in size, as shown by serial MRI, and the tumor was reduced in 1 patient. Similar to the results from our own study, there was no acute or chronic toxicity. Seven of nine patients reported decreased symptom intensity. In 1999, Liščák et al.32 published the results of a multicenter study. In that study, 66 patients underwent γ-knife radiosurgery at 6 European sites. After a median follow-up of 24 months, the tumor control rate was 100% (47 patients), and the treatment-related morbidity rate was 5.8% (3 of 52 patients). The complaints reported most often were temporary tinnitus (1 patient) and permanent facial nerve palsy (two patients). Both the tumor control rate and the morbidity rate from this relatively large series are comparable to our own results after patients underwent LINAC-RS for GJT. Results from a study in 2000 by Jordan et al.33 of eight patients who also underwent γ-knife radiosurgery generally agree with both our results and those of other groups.

In view of our own, promising results and the good clinical outcome, we suggest that, for patients with GJT, LINAC-based systems may be superior to γ-knife radiosurgery, mainly because of the higher mechanical flexibility of a linear accelerator compared with a γ-knife, as evident from 1) the greater field size and 2) the ability to irradiate targets below the skull base. Maximum field dimensions of up to 30 mm facilitate the treatment of larger tumor volumes (the median volume in our series was 12.2 cm3, and the maximum volume was 57.8 cm3). This technique is important in the treatment of tumor compartments located below the skull base. A promising development in LINAC-RS is the use of the mechanical micromultileaf collimator (MLC) (MRC Systems GmbH, Heidelberg, Germany), which we have used for 7 months. This device allows optimal dose conformation to nearly any geometry of the target volume. Dose inhomogeneities resulting from field overlap are avoided. Therefore, we expect a significant reduction in radiation-induced side effects by minimizing radiation exposure of adjacent healthy brain tissue and avoiding hot spots in critical structures like cranial nerves.


Our own results and those reported by others (a total of 96 patients evaluated) clearly show that stereotactic radiosurgery for the management of patients with GJT is safe and very effective. High local tumor control rates associated with a very low incidence of side effects were achieved. Important prerequisites for a good outcome are CT-based and MR-guided, three-dimensional treatment planning; optimum dose conformation for intracranial tumor compartments; and the integration of the results of risk analyses. Independent of the size of an extracranial tumor, LINAC-RS may be considered an alternative treatment for patients with GJT if the intracranial tumor compartment is relatively small. In patients with larger intracranial tumors who have substantial brain stem compression, it can be used as an adjuvant therapy after tumor excision. The use of a mechanical MLC for LINAC-RS has been encouraging. We expect that the use of a mechanical MLC will enable us to treat patients with larger tumors and to reduce radiation-induced side effects by minimizing the radiation exposure to adjacent healthy brain tissue and critical structures.


The authors thank Dr. A. Rodón for translation and editing.