A 10-year experience of linear accelerator-based stereotactic radiosurgery/radiotherapy (SRS/SRT) for paraganglioma: A single institution experience and review of the literature

Abstract

Objective

Patients with paragangliomas were treated with LINAC-SRS/SRT in this retrospective review to evaluate tumor control, clinical control, and toxicity.

Methods

16 patients (median age = 65) with paragangliomas were treated with LINAC-SRS/SRT. Patients were treated to a median dose of 25 Gy in 5 Fx and were evaluated for long-term tumor control, symptom control, and toxicity. Median follow-up was 44 months.

Results

16 paragangliomas with a median PTV of 11.7 cc were treated as above. All but 2 lesions were controlled at last follow-up, with a 5-year control rate of 88%. Eighty-one percent of patients reported improved or resolved symptoms after treatment. Toxicities included grade 2 vertigo in 1 patient and grade 3 headache from hydrocephalus requiring ventriculoperitoneal shunt.

Discussion

Linear accelerator based SRS/SRT appears to be an effective treatment option for paragangliomas. Recurrences in this cohort occurred 4-5 years after treatment, highlighting the importance of long term follow up.

Background

Paragangliomas, specifically glomus jugulare tumors (GJT), are rare, highly vascularized benign lesions that arise from the chief cells of the paraganglia in the chemoreceptor system and are closely associated with cranial nerves derived from brachial arches.^1-4,8 Typically afflicting women between the 6^th and 7^th decades, symptoms arise in 1-3% of individuals due to a four- to fivefold increase in catecholamines in part to their neuroendocrine origin.^1-3,5-6 Because of their slow growth, the average interval between appearance of the first symptom to diagnosis of GJT is four to six years.⁷ More common symptoms, which are usually a byproduct of location and size of the lesion, include otalgia, pulsatile tinnitus, hypoacusis, conductive deafness, dysphagia, dysphonia, vertigo, headache, and cranial nerve deficits.^1,5-6,8

Treatment options include surgical resection, embolization followed by surgical resection, conventional fractionated external beam radiation, radiosurgery, or a combination thereof.² Surgical resection is a definitive treatment option that offers complete resection, occurring approximately 90% of the time.^9-10 However, resection is often technically difficult due to the highly vascular nature of GJTs and their association with critical neural anatomy. Therefore, radiotherapy is an attractive, non-invasive treatment option for large, complex lesions that were previously considered inoperable.¹⁰

Historically, external-beam radiation therapy (EBRT) has been used as an adjunct for subtotal resections of GJT with the goal of preventing further growth.^11-12 The mechanism of treatment is due to the induction of obliterative endaritis, which fibroses surrounding blood vessels.⁸ While radiotherapy is effective at arresting growth, approximately 23% of GJTs reduce in size after treatment.¹³ An alternative to EBRT with standard fractionation, single-fraction stereotactic radiosurgery (SRS) such as Gamma Knife and linear accelerator-based (LINAC) SRS have been reported as primary and salvage treatment modalities for GJT.¹⁰ Similarly, hypofractionated stereotactic radiotherapy (SRT) adopts the same technique but divides the total dose into 3-5 fractions. These methods involve a shorter treatment time while delivering precise high-dose radiation to a small localized field, increasing the chances of obliterative endarteritis and reducing the risk of adverse effects with a high degree of accuracy.^5,8,11

Tumor control rates of 90-100% with lesser complication rates and improved quality of life have been historically reported utilizing SRS.^15-16 While radiosurgery halts tumor progression, it can also reduce preoperative symptoms by reducing pulsatile compression of the surrounding cranial nerves. This leads to reduced post-treatment cranial nerve dysfunction compared to surgery and SRS.^15,17 Uncommon toxicities due to excessive doses of radiotherapy include xerostomia, dermatitis, altered taste sensation, dermatitis, mucositis, mastoiditis, radiation-induced otitis, and induction of secondary malignancies.^1,8,10-11,14

While retrospective studies demonstrate safe and effective treatment with single-fraction radiosurgery,^1,18-24 the data regarding LINAC-SRS/SRT is limited.^9,25-26 The purpose of this study is to retrospectively review patients at a single institution and review the pertinent literature to date to determine efficacy and safety of LINAC-based single fraction and fractionated SRS for treatment of paragangliomas.

Methods

Patient Population

From 2008 to 2017, 16 patients (all of whom were female) between the ages of 55 to 87 (median 65) were treated with either linear accelerator based SRS or hypofractionated SRT in this IRB approved study. None of the 16 patients had prior external-beam radiotherapy and 6 patients had prior failed surgery for GJT resection. Two out of the sixteen patients underwent radiotherapy for recurrence of prior GJT, one of which was treated in a salvage setting (15 Gy in 1 Fx) and the other who underwent previous surgery and was treated with adjuvant SRS.

Treatment Planning

All patients were immobilized with the Brainlab (Feldkirchen, Germany) relocatable mask system during simulation and treatment. A Gadolinium contrast-enhanced T1-weighted Magnetic Resonance Image (MRI) was acquired, with a resolution of 0.5mm × 0.5mm and a slice thickness of 2mm. The patient was then fitted with the immobilization system in the CT simulation room. A mouth bite attached to the ring was placed against the patient’s upper dentition to prevent head tilt movement while the customized thermoplastic mask was molded. In some case, mouth bite was not used due to patient’s intolerance. A CT was acquired with a resolution of 1 mm × 1 mm and a slice thickness of 2 mm. It was then rigidly registered to the MRI dataset in the Brainlab iPlan Image software. The Physician contoured the Gross Target Volume (GTV), which was expanded with 2-5 mm margin to generate the Planning Target Volume (PTV). A treatment plan with 4 – 10 non-coplanar conformal arcs was generated using a pencil beam algorithm in Brainlab iPlan Dose software. Patients were treated to 30 Gy in 5 Fx (n=1), 27.5 Gy in 5 Fx (n=4), 25 Gy in 5 Fx (n=4), 21 Gy in 3 Fx (n=1), or 15 Gy in a single Fx (n=6) prescribed to the 80-90% isodose line. Alignment was confirmed with megavoltage cone beam prior to every treatment.

Statistical Analysis

In this retrospective study, we evaluated long-term tumor control, symptom control, and toxicity. Time to failure after treatment was defined as the length of time (in months) between radiotherapy and radiographic evidence of progression. Statistical analysis was carried out using IBMÒ SPSS statistical software v20. Age, sex, prior surgeries before RT was provided, and SRT/SRS parameters were reported along with radiation dose, planning target volume (PTV), BED₃, and PTV coverage. Symptoms prior to SRT or SRS were noted along with any adverse effects or radiation-related toxicities that may have arisen after treatment. Toxicity was graded with the CTCAE standard toxicity scale.

Results

The median follow-up was 44 months. All patients receiving hypofractionated SRT and SRS were females and were symptomatic prior to receiving treatment. Pre-treatment symptoms included (in order of decreasing incidence): tinnitus (56.25%), vertigo (12.5%), cranial nerve palsy (12.5%), hearing loss (12.5%), tongue deviation (6.25%), headache (6.25%), slurred speech (6.25%), dysphagia (6.25%), pulsing sensation (6.25%), and heartbeat sounds (6.25%). After treatment with SRS/SRT, 13/16 patients experienced improvement or resolution from symptoms ranging from one month to two years. The other three patients had symptomatic control, but not improvement or resolution, at last follow-up. For example, 7 out of 9 patients with tinnitus experienced either improvement or resolution, while 1 out of 2 patients with cranial neuropathies experienced improvement, but not resolution. The other patient that did not experience either improvement or resolution was an individual exposed to SRS in a setting of 15 Gy that previously had headache, which was symptomatically stable at last follow-up. Patient characteristics are displayed in Table 1.

Table 1.

Characteristics of the 16 patients undergoing SRT/SRS for paragangliomas (glomus jugulare tumors) at Allegheny General Hospital (2009-2017).

	Incidence	Median (range)
Sex
- Female	100% (n=16)
- Male	0% (n=0)
Age		65 (44-87)
Prior surgeries	37.5% (n=6)
- Interval between surgery and RT (in months)		73 (18-204)
Prior external-beam radiotherapy	0% (n=16)	0 (0)
Tumor location (right vs left)
- Right	75% (n=12)
- Left	25% (n=4)
Pre-treatment symptoms (if any)	100% (n=16)
Post-treatment improvement (if any)	81.25% (n=13)
Follow-up images acquired		3 (0-6)
Time for follow-up (in months)		44 (1-86)

Twelve right and four left sided GJTs (Table 1) were managed with a median planning target volume (PTV) of 11.7 cc (range 0.53-28.40 cc) and a median BED₃ of 77.9 Gy (range 66.7-90 Gy) (Table 2). The median PTVs for SRS and SRT were 6.15 cc and 12.87 cc, respectively. Specifically, the 6 patients that underwent SRS were treated in a setting of 15 Gy in 1 Fx while 9 patients underwent LINAC-SRT in a setting of 21 to 30 Gy in 3 5 fractions. Six patients underwent surgical resection prior to salvage radiotherapy with a median time interval to radiotherapy of 73 months (Table 1).

Table 2.

SRS/SRT treatment parameters for paragangliomas in 16 patients.

	Median (range)
Treatment dose (Gy)	25 (15-30)
Radiation dose per fraction (Gy)	5 (1-5)
Ionizing radiation (in Curie; Ci)	1.6 (0.78-85.3)
Isodose line	90% (80%-94%)
Planning target volume (volume)	11.7 (0.53-28.8)
- SRT	12.87 (0.53-28.8)
- SRS	6.15 (2.74-27.5)
Planning target volume (cc)	98 (90-100)
BED3	77.9 (66.7-90)

All but 2 lesions were controlled at last follow-up, resulting in a 5-year local control rate of 88%. The two local failures occurred 73 and 43 months post-SRS. Both treatment failures occurred after being treated with 15 Gy in 1 Fx (BED₃ of 90 Gy). The former lesion measured 27.50 cc. After failure was recognized, and successful salvage radiotherapy with ERBT to 45 Gy in 25 Fx was delivered. The second patient with a local failure previously had a 4 cm lesion that was resected prior to developing a new 8.20 cc lesion. After radiotherapy, the patient developed new 2 mm intracranial extension 45 months after SRS. Due to the small and asymptomatic nature of the recurrence, salvage has not been attempted as of the last follow-up.

Thirteen of sixteen patients reported improved or resolved GJT-related symptoms after treatment (81.25%; Table 1). Adverse events post-radiotherapy included grade 2 vertigo in one patient after receiving 15 Gy in 1 Fx SRS and grade 3 headache after receiving 27.5 Gy in 5 Fx SRT with concomitant obstructive hydrocephalus in the other. In the formers case, the patient was treated with meclizine which resulted in resolution of vertigo 12 months post-radiotherapy. In the latter case, the patient’s lesion caused obstructive hydrocephalus prior to receiving 27.5 Gy in 5 Fx. The patient’s hydrocephalus was managed by insertion of a ventriculoperitoneal shunt. This patient’s headaches post-radiotherapy was managed with dexamethasone. The total incidence of adverse effects occurring post-radiotherapy was 12.5%.

Discussion

Surgery vs. Radiotherapy

Treatment for paragangliomas should be tailored to each patient. While surgical resection offers the only form of definitive treatment, it is not without risk. Even with the advent of microsurgery, GJT have multiple postoperative morbidities due to its high vascularity and association with multiple cranial nerves.^2,8,10,12 Radiotherapy has traditionally been offered to patients for postoperative recurrence or non-surgical candidates, but stronger evidence favors radiotherapy as a first-line treatment due to its long-term efficacy to prevent tumor progression and to retain cranial nerve function.²⁷ To compare local control rates in surgical patients to patients receiving fractionated ERBT, Hu and Peresky²⁷ determined that 93% patients undergoing radiotherapy achieved local control while 78% of patients that underwent surgery and 85% undergoing a combination of ERBT and surgical removal achieved local control in a follow-up period of 11 to 16 years. Furthermore, a comprehensive review of the literature by Suárez et al. was conducted to compare surgical removal and radiotherapy success in the treatment of head and neck paragangliomas.²⁸ Among patients with GJT, significantly better rates of local control were observed with radiotherapy compared to surgery (91.5% vs. 78.2%; P = 0.002).²⁸ In addition, major treatment complications (and cranial nerve palsies) were observed at a significantly higher rate in surgical patients compared to the radiotherapy cohort after treatment (28.2% vs. 11.4%; P = 0.003), however, no significant difference was observed between deaths due to radiotherapy and surgery. The author notes that given the design of this particular study, there could likely be significant selection bias and results should be interpreted with caution.²⁸

Single- vs fractionated SRT tumor and clinical control

An alternative to radiotherapy with standard fractionation, SRS such as GK and LINAC have been reported as primary and salvage treatment modalities for GJT.¹⁰ Patients who are candidates for SRS typically have tumors that do not invade the brain tissue and are well-demarcated on MRI.¹⁰ Suárez et al., in a retrospective review, concluded that SRS and EBRT were not significantly different in maintaining tumor control and observed no significant differences in overall complication rates (93.7% vs. 89.1% and 6.5% vs. 10.4%; P = 0.32 and 0.53).²⁸ In addition, Schuster et al. examined radiographic tumor control and symptom improvement between single- and fractionated LINAC stereotactic radiosurgery for treatment of GJT.²⁶ Although single fraction SRS in a setting of 12 to 15 Gy have been typically outlined,²⁷ Schuster et al. used a median radiation dose of 15 to 18 Gy in 1 Fx.²⁶ The authors reported a 92% stability or regression in tumor size with a disease improvement incidence of 40% in single fractionated patients and a 44.4% improvement in disease utilizing multi-fractionated radiotherapy.²⁶ Five-year tumor control in our series was 88% with 2 of 6 SRS patients experiencing treatment failure and 100% tumor control in the 10 patients that underwent hypofractionated SRT (t-test 0.22; P < 0.05). Fiengenberg et al. also reported similar findings in a series of five patients undergoing SRS (15 Gy in 1 Fx) in which 40% of their cohort experienced treatment failures.²⁴ Results must be interpreted with caution as both data sets are obviously limited due to a smaller population pool that amplifies treatment failures. In addition, one needs to take into consideration past surgery, tumor size, biological effective dose utilized, and any prior treatment when trying to analyze failures in such a study.

The rates of clinical control, defined as a stabilization or improvement in symptoms after treatment, utilizing LINAC-based radiotherapy are variable in the literature. ^1,9,18-26 For example, Wegner et al. reported 100% stable or improved symptoms in their cohort of 18 patients utilizing a median dose of 20 Gy in 3 fractions,²⁵ while a study conducted by Chun et al. utilizing a median dose of 25 Gy in 5 fractions reported only a 60% improvement of pretreatment symptoms using CyberKnife-SRT.²⁹ In addition, 161 patients across the 11 studies found to utilize LINAC-based radiosurgery and radiotherapy were analyzed showing comparable results when comparing single- and multi-fractionated radiotherapy (Table 3). ^1,9,18-26 Furthermore, at 3-year follow-up, our patient population experienced comparable rates of clinical control (100%) relative to a meta-analysis of the past data presented by Guss et al. (95%).⁵ In our series, the median radiation dose delivered was 25 Gy compared to previous studies utilizing LINAC-SRS or SRT (15-20.4 Gy).⁵ Poznanovic et al., while approaching similar levels of efficacy in reducing clinical symptoms related to GJT, utilized 15 to 16 Gy to the periphery of the PTV.¹ Their mean tumor volume was 7.24 cc¹ while our cohort had a slightly larger mean tumor volume of 13.28 cc.

Table 3.

Summary of major series presenting outcomes of glomus jugulare patients treated with SRS/SRT.

Series	Year	Patients (n)	Modality	Median prescription dose/fraction (n)	Tumor control	Symptom Improvement	Toxicities	Median follow-up (months)
Poznanovic et al. (1)	2008	8	LINAC	15/1 (8)	100%	87.5%	2	17.5
Henzel et al. (9)	2007	17	LINAC	Culm 57/3-5 (17)	100%	56.3%	10	40
Adrande et al. (18)	2013	15	LINAC	13/1 (14) 20/5 (1)	100%	20%	1	33
Lim et al. (19)	2003	9	LINAC/CK	18/1 (7) 18/3 (3)	100%	Not Reported	2 (Both occurring in 1 fx)	26
Lim et al. (20)	2004	13	LINAC/CK	18/1 (13) 24/3 (3)	100%	Not Reported	2 (Both occurring in 1 fx)	41
Lim et al. (21)	2007	18	LINAC/CK	20/1-3 (4 tumors treated in 3 fx and 1 tumor treated in 2 fx)	100%	Not Reported	3 (All 3 occurring in 1 fx)	35
Maarouf et al. (22)	2003	13	LINAC	15/1 (13)	100%	23.1%	1	48
Majdoub et al. (23)	2015	27	LINAC	15/1 (27)	100%	37%	1	132
Feingenberg et al. (24)	2002	5	LINAC	15/1 (5)	60%	40%	1	27
Wegner et al. (25)	2010	18	LINAC	20/3 (25)	100%	22.2%	0	22
Schuster et al. (26)	2015	18	LINAC	18/1 (5) 20/5 (9)	100% 88.9%	40% 44.4%	0 1	28.8
Current Series	2017	16	LINAC	15/1 (6) 26.5/5 (10)	66.7% 100%	83.3% 80%	1 1	44

Adverse Effects to SRT/SRS

In the past, radiotherapy has been criticized for high exposure of normal neural tissue and bone to radiation, resulting in considerable radiation-induced complications.^27-28 However, SRT/SRS delivery a highly conformal dose to targets while sparing adjacent normal tissues. Gottfried et al., in a comprehensive literature review, reported a neurological complication rate in 8.5% of patients undergoing SRS.¹⁰ Neurological complications are either transient or permanent with transient complications exacerbating pre-existing cranial nerve deficits, tinnitus, or vertigo while permanent complications include worsening facial nerve function, vertigo, and progressive hearing loss or deafness.²⁷ Out of the 8.5% of patients in the meta-analysis performed, 6.4% of these had transient adverse effects while 2.1% experienced permanent effects.¹⁰

Upon review of the literature for patients undergoing single- and multi-fractionated radiotherapy (Table 3), toxicities involved typical post-radiotherapy side effects including vertigo, nausea, and headache along with lower cranial neuropathies.^1,9,18-26 Furthermore, the rate of lower cranial neuropathies in SRS and SRT (8.1%)^1,18-24,26 were comparable to the rate described by Gottfried et al.¹⁰ However, of the 13 patients that experienced cranial neuropathies, only one of those underwent hypofractionated SRT (median 20 Gy in 5 Fx; incidence 0.62%) and 1.24% of patients experienced a permanent cranial neuropathy, all in a setting of single-fraction SRS (median 15 Gy in 1 Fx).^1,18-24,26 In addition, 6.8% of patients in previous studies reported post-radiotherapy neurologic toxicities including nausea, headache, and vertigo.^1,9 These typically were reported in a setting of fractionated radiotherapy while only one patient out of the 161 cases reviewed underwent single-fraction SRS.^1,9

In our experience, one out of the two patients experiencing adverse effects to treatment underwent single-fraction SRS. This patient developed grade 2 vertigo post-radiotherapy, and occurred in a setting of 15 Gy in 1 Fx. In 3 studies involving LINAC-based SRT, a wide range of patients (0-58.8%) experienced adverse effects,^9,25-26 which is comparable to our cohort of patients (1/9; 11%). The range of toxicities seen is likely attributable to the small sample size of the populations being observed, therefore caution must be taken due to the amplification of treatment failures. Rarely, severe toxicities such as osteoradionecrosis, brain necrosis, radiation-related cranial neuropathy, external auditory canal stenosis, trismus and radiation-related cranial neuropathy have been reported, none of which were observed in our series.²⁷

Limitations

The standard limitations of a retrospective study involving rare tumors including selection bias and small sample size apply here. The lack of males in this study may lead to a selection bias, as there could be differences in physiology and response to radiotherapy that could play a role across determining treatment efficacy in either gender. It has also been reported in the literature that recurrent tumors can arise up to 25 years after initial treatment with radiotherapy,³⁰ therefore, a longer follow-up time post-radiotherapy should be utilized to more accurately ascertain true efficacy in halting the progression of the lesion.

Conclusion

We add to the literature a growing but still limited database describing definitive and salvage radiation-based treatment for paragangliomas, which has predominantly been with gamma-knife to this point. This single institution study demonstrates that glomus jugulare tumors can be treated safely, effectively, and with improvement in symptoms with LINAC-based hypofractionated stereotactic radiotherapy. In our limited series, there were no local failures with LINAC-fractionated SRT and 2/6 failures with single-fraction SRS. Further investigation is warranted to unveil if any true difference in outcome exists between the fractionation schemes.