Clinical and Imaging Response to Trigeminal Schwannoma Radiosurgery: A Retrospective Analysis of a 28-Year Experience

Abstract

Objective  The purpose of this study was to evaluate long-term clinical outcomes and tumor control after stereotactic radiosurgery (SRS) for trigeminal schwannoma (TS).

Methods  During a 28-year period (1989–2017), 50 patients underwent SRS for TS. The median patient age was 51 years (range: 15–87 years). A total of 17 patients had a previous tumor resection: 10 had a single procedure, 5 had two procedures, and 2 had three procedures. The median and mean times between tumor resection and SRS were 12 and 24 months (range: 1–90 months), respectively. Four patients had neurofibromatosis II (NF2). Based on location, tumors were classified as root type (7), ganglion type (22), or dumbbell type (21). The median radiosurgery target volume was 3.4 cm ³ (range: 0.10–18 cm ³ ), median target dose was 14 Gy (range: 12–20 Gy), and the median number of isocenters was 6 (range: 1–15). The median and mean times to last follow-up was 36.9 and 55.2 months (range: 4–205 months), respectively. Eighteen patients (36%) had longer than 5-year follow-up, and seven patients (14%) had longer than 10-year follow-up.

Results  The tumor control rate was 92% and the clinical improvement or stabilization rate was 94%. After SRS, the rates of progression free survival (PFS) at 1, 5, and 10 years were 98, 84, and 84%, respectively. Factors associated with improved PFS were female sex ( p  = 0.014) and smaller tumor volume ( p  = 0.022). In this series, we did not find that tumor type (root, ganglion, and dumbbell) had a statistically significant correlation to PFS. Forty-seven patients had neurological signs or symptoms at presentation. At last follow-up, neurological signs or symptoms improved in 22/47 (47%), remained unchanged in 24/50 (48%), and worsened due to tumor progression in 3/50 (6%). One patient (2%) developed temporary symptomatic adverse radiation effect (ARE) and three additional patients (6%) had transient imaging evidence of peritumoral reactive edema but no new symptoms.

Conclusion  As a single outpatient procedure, SRS was associated with long-term freedom from additional management in 84% of patients. Nearly half the treated patients experienced improvement in neurological symptoms or signs.

Introduction

A schwannoma, also known as a neurilemoma, is a rare nerve sheath tumor. Trigeminal schwannomas (TS), which constitutes less than 10% of all schwannomas, are characterized by tumor growth within any region of the fifth cranial nerve (CN). Depending on tumor size and location, TS can be discovered due to patient complaints, such as headache (HA), facial numbness, facial pain, or diplopia, related to compression of the third, fourth, or sixth cranial nerves. ^{1 2} Multiple cranial nerve schwannomas can be seen in patients with neurofibromatosis II (NF2).

Management strategies for TS include observation (in incidentally discovered tumors), partial or complete microsurgical resection, ^{3 4 5 6 7 8 9 10 11 12 13} or stereotactic radiosurgery (SRS). ^{1 14 15 16 17 18 19 20 21 22 23 24} Initial microsurgical resection is necessary for larger volume tumors causing symptomatic brainstem compression. SRS is a primary management strategy for smaller volume tumors and as an adjuvant salvage option after attempted resection of larger tumors associated with symptomatic mass effect. SRS is a minimally invasive treatment option with high tumor control rates, high clinical stabilization rates, and low rates of treatment related complication (including adverse radiation effects [ARE]). ^{1 14 18 19 20 21 23} Due to the limited incidence of TS, SRS studies with long-term follow-up and large patient populations are important to document the long-term benefit, as well as risks of SRS. In this report, we describe our expanded series and long-term outcomes after SRS performed as a primary or salvage option for TS. ¹

Methods

Patient Selection

We retrospectively reviewed the medical and imaging records of 50 patients (25 males and 25 females) who underwent SRS for TS during a 28-year time period (March 1989–June 2017; Table 1 ). The median patient age was 51 years (range: 15–87 years). Forty-seven patients had clinical symptoms at the time of SRS. The median symptom duration before SRS was 11 months (range: 0.7–164 months). Four patients had NF2. Seventeen patients had at least one previous surgical resection, while 33 patients had SRS as the primary management option. In patients with a previous tumor resection, 10 had a single procedure, 5 had two procedures, and 2 had three procedures. The median and mean times between tumor resection and SRS were 12 and 24 months (range: 1–90 months), respectively.

Table 1. Patient demographics, tumor characteristics, and radiosurgical parameters.

Parameter	No. of patients n (%)	Median	Range
Patient demographics
Age at SRS (y)		51	15–87
Symptom duration (mo)		11.4	0.7–164
Sex
Male	25 (50)
Female	25 (50)
Treatment side
Right	21 (42)
Left	29 (58)
Previous surgical resection	17 (34)
Interval between first resection and SRS (mo)		18	1–134
Interval between last resection and SRS (mo)		12	1–90
Tumor characteristics
Tumor volume (cm 3 )		3.4	0.10–18
Type of tumor
Solid	45 (90)
Cystic	1 (2)
Mixed	4 (8)
Tumor classification based on location
Root	7 (14)
Ganglion	22 (44)
Dumbbell	21 (42)
Brainstem compression	16 (32)
Radiosurgical parameter
Target volume (cm 3 )		3.4	0.10–18
Target/marginal dose (Gy)		14	12—20
Isocenters		6	1—14

Abbreviation: SRS, stereotactic radiosurgery.

Tumor Classification

Magnetic resonance imaging (MRI) defined 45 diffusely contrast enhancing tumors and 5 mixed (solid and cystic) tumors. Sixteen patients had imaging evidence of brainstem compression but no clinical symptoms of imbalance, headache, or motor weakness. Based on primary anatomic location, tumors were classified as root (7), ganglion (22), or dumbbell type (21). ² At the time of SRS, 21 patients had right-sided treated tumors, while 29 patients had left-sided treated tumors. The median tumor volume was 3.4 cm ³ (range: 0.10–18 cm ³ ).

Stereotactic Radiosurgery Technique

SRS was performed using one of six versions of the Leksell Gamma knife installed at University of Pittsburgh Medical Center (UPMC). ^{1 17} The median radiosurgery target volume (which matched the tumor volume) was 3.4 cm ³ (range: 0.10–18 cm ³ ; Table 1 ). The median marginal dose was 14 Gy (range: 12–20 Gy) prescribed to a median of 50% isodose line. The median number of isocenters was 6 (range: 1–15). Following each radiosurgery procedure, each patient received a single intravenous dose of 20 to 40 mg methylprednisolone.

Follow-up Protocol

Each patient underwent serial clinical and imaging evaluations at 6 months, 1 year, and thereafter at 2-year intervals. Neurological findings at follow-up were recorded as improved, unchanged, or worse. The tumor volume response was recorded by measuring the X, Y, and Z dimensions in cm and then multiplying by 0.5. Specifically, tumor volume response was noted as complete regression, partial regression if the subsequent tumor volume decreased by >25%, stable tumor if there was < 25% change in tumor volume, or tumor progression if there was an increase >25% in tumor volume.

If new clinical symptoms developed at any point in the follow-up, MRI was requested to assess whether tumor progression or suspected ARE was present. Patients with suspected ARE were prescribed a short dose of corticosteroids and monitored for symptoms resolution. Patients with tumor progression post SRS were considered for an additional SRS or tumor resection.

Statistical Analysis

IBM SPSS Statistics 25 was used for all statistical analysis. Logistic regression determined independent variables (i.e., patient demographics, tumor characteristics, or treatment parameters) with statistically significant predictions of outcomes post-SRS (i.e., clinical benefit, tumor change, or ARE). A Kaplan–Meier curve was created for progression-free survival (PFS). Patients without tumor progression were censored at last follow-up. A reverse Kaplan–Meier curves was created to determine time to clinical benefit. Patients without clinical benefit were censored at last follow-up. Another reverse Kaplan–Meier curve was created to determine time to tumor regression. Patients with no change in tumor volume were censored at last follow-up, while patients with tumor expansion were censored at the time of tumor progression. Cox's regression determine covariates with statistically significant correlations to PFS, time to clinical benefit, or time to tumor regression. The log-rank test was used to determine factors with statistically significant differences within PFS, time to clinical benefit, or time to tumor regression. A p -value of 0.05 was set for statistical significance.

Results

Clinical Outcomes after Primary versus Salvage Stereotactic Radiosurgery

Clinical signs and symptoms are illustrated in Table 2 . Overall, the most typical symptoms at the time of SRS were trigeminal sensory loss, trigeminal pain, and diplopia related to sixth CN dysfunction. At last follow-up, 22 had clinical improvement, while 25 patients were unchanged. Two patients had increased trigeminal neuropathy and one patient had increased trigeminal pain.

Table 2. SRS clinical changes.

Treatment	Cranial nerve	Before SRS	Improved	Unchanged	Worse	New symptom
Primary ( n  = 33)
	III	1		1
	IV	2		2
	V neuropathy	20	12	6	2	1
	V pain	9	4	4	1
	Abducent	9	2	6
	Facial	3	2	1
	Acoustic	2	1	1
Adjuvant ( n  = 17)
	III	1	1
	IV
	V neuropathy	13	2	11
	V pain	4	2	2
	Abducent	3	1	2
	Facial	1	1
	Acoustic	1	1

Abbreviation: SRS, stereotactic radiosurgery.

Seventeen of the 22 patients with clinical improvement also had tumor volume reduction and five with improvement had no change in tumor volume. Based on logistic regression, clinical improvement could be predicted by tumor regression post-SRS ( p  = 0.039). Patients with primary SRS were more likely to improve than patients who had SRS after prior surgical procedures ( p  = 0.042; Table 2 ; Fig. 1 ). After SRS, 53.3% had clinical improvement at a median interval of 8 months. In contrast SRS after prior surgery was associated with clinical improvement in 35% which was reported at a significantly longer median interval of 63.1 months.

Fig. 1.

Kaplan–Meier survival graph showing the time to clinical improvement in trigeminal schwannoma patients treated with primary and salvage radiosurgery.

Imaging Response

MRI follow-up was available for all 50 patients. Post-SRS, 64% had tumor regression at a median time of 1 year (range: 0.25–8.6 years). Overall, the PFS at 1 year, 5, and 10 years was 97.9, 83.5, and 83.5% ( Fig. 2 ), respectively. Based on Cox's regression, females ( p  = 0.014) and patients with smaller tumor volume ( p  = 0.022) were more likely to have better PFS. No other patient factors (age, NF2 status, and symptom duration), tumor features (type, location, and brainstem compression), or treatment-related factors (marginal dose or isocenter) had statistically significant correlations to PFS.

Fig. 2.

Kaplan–Meier survival graph showing progression free survival in trigeminal schwannoma patients treated with radiosurgery.

Four patients had a complete response and twenty eight had partial tumor regression ( Fig. 3 ). In this experience, we were unable to find any specific imaging feature (e.g., homogeneous contrast, cystic change, or volume) that confidently predicted tumor response. Fourteen patients had stable tumor volumes without further growth, including 11 patients who underwent primary SRS and 3 who had SRS after prior surgery. Four patients (two had primary SRS and two had SRS after surgery) had delayed tumor progression detected at a median interval of 18 months (range: 6–51 months). The average tumor volume for these four patients was 10.4 cm ³ . Based on logistic regression, patients with larger tumor volumes were more likely to have delayed tumor progression ( p  = 0.023).

Fig. 3.

A 53-year-old woman presented with left facial numbness and underwent radiosurgery for left sided trigeminal schwannoma using a margin dose of 13 Gy (top). Her serial follow-up brain MRI show gradual tumor regression (bottom). MRI, magnetic resonance imaging.

Case Example of Delayed Additional Resection

A 15-year-old patient with a left-sided, dumbbell-type trigeminal schwannoma underwent initial subtotal resection of posterior fossa component of the tumor using a left retromastoid craniectomy. The residual tumor was treated using gamma Knife SRS. A margin dose of 13 Gy was delivered to 50% isodose volume measuring 10.5 cc. The maximum tumor does was 26 Gy. He was noted to have tumor progression 18 months later and underwent a second SRS using a margin dose of 12 Gy to 50% isodose volume measuring 24.7 cc. The maximum tumor dose was 24 Gy. Subsequent imaging showed initial tumor regression but delayed new growth which prompted a second surgical resection 9.5 years after the first SRS procedure ( Fig. 4 ). The patient underwent tumor resection using expanded endonasal transpterygoid approach. The tumor has remained stable in a 10-year follow-up after resection.

Fig. 4.

A 15-year-old boy underwent SRS for residual a left sided, 10.5 cc dumbbell-type trigeminal schwannoma ( A ). He was noted to have tumor progression 18 months later and underwent a second SRS ( B ). Subsequent imaging showed initial tumor regression but delayed new growth ( C ) which prompted a second surgical resection 9.5 years after the first SRS. A significant tumor debulking was achieved ( D ). The tumor has remained stable in a 10-year follow-up after resection ( E ). SRS, stereotactic radiosurgery.

Adverse Radiation Effects

Four patients had imaging evidence of ARE at a median interval 25 months (range: 6–56 months). For these patients, follow-up imaging (MRI) showed the development of peritumoral contrast-enhancement and transient tumor enlargement. Three patients with imaging ARE were asymptomatic. One patient developed V2 and V3 sensory loss which was persistent at last follow-up (56 months post-SRS) in the absence of tumor growth. All patients with ARE were treated with a short course of oral corticosteroids and had resolution of imaging changes over time. Based on logistic regression, we did not find any factors with statistically significant predictions of ARE.

Extended Term Results

In the entire clinical series, the median and mean times to last follow-up were 36.9 and 55.2 months (range: 4–205 months) respectively. Thirty-three patients had more than 2-year follow-up, 18 had more than 5-year follow-up, and 7 had more than 10-year follow-up.

Within this study, patients with a follow-up greater than 5 years qualified as a long-term follow-up patients. Among 18 patients with a long-term follow-up, the median time to last follow-up was 109.8 months (range: 61.6–205 months). Eleven had tumor regression at median of 63 months (range: 6–114 months), 4 had tumor stabilization at last follow-up, and 3 had delayed tumor progression at median interval of 80 months (range: 76–114 months). The 10-year PFS for these long-term patients was 88%. Within this group, there was no independent variable with statistically significant prediction of clinical improvement, tumor change, or ARE based on logistic regression. There was no independent variable with statistically significant correlation to PFS based on Cox's regression or log-rank test within the long-term follow-up group.

Notably, one patient in the long-term follow-up group had a 17-year follow-up. Specifically, this was a 24-year-old male with ganglion-type TS who underwent salvage SRS (marginal dose: 14 Gy) 1 month after surgical resection. This patient experienced clinical improvement in his third and sixth CN palsy and tumor regression beginning at 6 months post-SRS. At last follow-up (205 months), no new clinical symptoms were present, the tumor had completely disappeared, and there was no evidence of new tumor growth or ARE.

Discussion

Role of Stereotactic Radiosurgery in the Treatment of Trigeminal Schwannomas

While TS remains a relatively rare intracranial tumor, it can lead to troubling clinical symptoms such as trigeminal neuropathy, trigeminal pain, and abducens or oculomotor palsies. For patients with small, incidentally detected tumors, observation and periodic surveillance imaging may be the best initial strategy. Patients with large symptomatic tumors (> 10 mL) generally are considered candidates for surgical resection. Initial complete surgical resection is associated with significant increased morbidity. ^{7 8}

SRS is an alternative primary option for patients with smaller volume tumors, or an adjuvant salvage option for patients after incomplete tumor removal. Because of potential medical comorbidities, SRS may be an alternative minimally invasive approach

Review of Stereotactic Radiosurgery for Trigeminal Schwannomas

Published studies of SRS for TS are shown in Table 3 . Due to low incidence of TS, it is important to elucidate the significance of past findings of SRS studies. This can also lead to a better analysis of treatment recommendations for individual patients.

Table 3. Literature review of SRS for trigeminal schwannoma.

Publication year	Study	No. of patients	Tumor volume (cm 3 ) Median (range)	Marginal dose (Gy) (Range)	Time to follow-up (mo)	Tumor control rate (% of patients)	Clinical improvement n (%)	Worsening or new symptoms n (%)	Symptom description
1999	Huang et al 16	16	5.3 (1–17.8)	15.2 (12–20)	44 (8–116)	100	31.3%	0%	N/A
2004	Nettel et al 17	12	4.5 (0.46–11.2)	15 (13–20)	40 (median)	91	12 (52)	2 (16.7)	Facial numbness, facial weakness
2005	Pan et al 18	56	8.7 (0.8–33)	13.3 (10–15)	68 (mean)	93	37 (66)	4 (7.1)	Facial numbness
2006	Sun et al 23	58	4.6	13.1 (10–14.4)	42.5 (mean)	93.1	48.3%	12%	Trigeminal numbness
2007	Peker et al 19	15	4 (0.9–17)	16 (14–22)	61 (mean)	100	40%	1 (6.7)	Diplopia, facial numbness
2007	Hasegawa et al 15	37	10	14.2 (11–16)	54 (mean)	84 (5 year)	40%	6 (16.2)	Facial pain, facial numbness
2007	Sheehan et al 21	26	3.96 (0.63–8.5)	15 (10.2–17)	48.5 (mean)	88	18 (72)	3 (12)	Facial pain
2007	Phi et al 20	22	4.1 (0.2–12)	13.3	37 (mean)	95	15 (67)	6 (27)	CN neuropathy
2009	Kano et al 1	33	4.2 (0.5–18)	15 (12–20)	72 (mean)	82 (5 year)	11 (33)	3 (9)	Facial sensory loss, facial pain
2012	Yianni et al 24	74	5.3 (0.4–19.9)	16.4	48.2 (mean)	92.7 (5 year)	8 (11)	7 (9)	Diplopia, facial numbness, facial paresthesia
2013	Sun et al 22	52	7.2 (0.5–38.2)	13.9 (11–17)	61 (mean)	86.9	35 (67.3)	2 (4)	Facial numbness, TGN
2013	Hasegawa et al 14	53	6 (0.2–33)	14 (11–20)	98 (median)	90 (5 year)	20 (49)	11 (22)	Facial numbness, Facial pain, III/IV palsy
Present study	Present study	50	3.4 (0.1–18)	14 (12–20)	55.2 (mean)	92	22 (47)	4 (8)	Facial neuropathy, facial numbness

Abbreviations: CN, cranial nerve; N/A, not available; SRS, stereotactic radiosurgery.

In the published literature, the highest number of patients in a single study was 72 patients. ²⁴ SRS was shown to have a clinical improvement rate up to 72%, tumor control rate up to 100%, and ARE as low as 0%. ^{16 19 21} However, on average, the number of patients per study was 39, marginal dose was 14.5 Gy, tumor volume was 5.5 cm ³ , time to follow-up was 56 months, chance of clinical improvement was 48%, chance of tumor control was 91%, and chance of clinical worsening or new symptoms was 12% (most commonly trigeminal neuropathy/pain). ^{1 14 15 16 17 18 19 20 21 22 23 24}

Kano et al found that a higher marginal dose was correlated with improved tumor control. ¹ Snyder et al found that a higher marginal dose was associated with increased rates of tumor expansion. ²⁵ Since SRS for TS is a single session of radiation to a small localized region, it is logical that increased dosages can be associated with improved tumor control and increased chance of ARE. Another study found that patients with dumbbell tumors and brainstem compression had less satisfactory tumor control rates. ^{1 14} This may indicate that while SRS is valuable as a salvage option postsurgical resection, it may be most successful as a primary treatment for small tumors, less than 3.0 cc. ^{1 24}

Our literature search yielded relatively few long-term outcome studies. Hasegawa et al reported 53 patients, with the longest follow-up (median: 98 months). ¹⁴ These authors showed a 5-year tumor control rate of 90%, 87% tumor control rate at last follow-up, and a clinical improvement rate of 49%. ¹⁴ Additional studies, with long-term follow-up will be valuable to determine extended term efficacy of SRS for TS.

Role of Resection and Review of Published Surgical Series

Previous studies of surgical resection for TS are summarized in Table 4 . In the overall surgical published series, the maximum number of patients in a single study was 105, up to 96% of patients had radical removal, up to 7% had mortality, and up to 87% had some neurological deficit. ^{3 4 5 6 7 9 10 12 13} However, on average, there were 50 patients per surgical series, 78% had radical removal, 1% with mortality, and 35% with some morbidity. ^{3 4 5 6 7 9 10 12 13}

Table 4. Literature review of surgery for trigeminal schwannoma.

Publication year	Study	No. of patients	Radical removal (%)	Mortality (%)	Morbidity (%)	Morbidity description
1988	McCormick et al 8	14	43	7	78	CSF leakage, CN palsy
1989	Pollack et al 10	16	75	0	6
1996	Konovalov et al 7	111	77	3	87	Facial numbness
1999	Yoshida and Kawase	27	74	0	74	Facial hypesthesia, CN palsy
2003	Goel et al 5	73	70	3	7
2007	Pamir et al 9	18	94	0	28	CSF leakage, epidural hematoma, meningitis
2010	Fukaya et al 4	57	81	2	68	CN palsy
2012	Wanibuchi et al 12	105	82	0	9	Facial hypesthesia, Diplopia
2014	Chen et al 3	55	95	0	5	Facial hypesthesia, diplopia, ptosis
2014	Samii et al 11	20	75	0	4	CSF leakage
2014	Jeong et al 6	49	96	0	18	Meningitis, CSF leakage

Abbreviations: CN, cranial nerve; CSF, cerebrospinal fluid.

In comparison to radiosurgery, after surgery, patients have reduced tumor control (78 vs. 91%), and greater morbidity (35 vs. 12%), including postoperative mortality (1%). While many patients had preoperative trigeminal neuropathy, 11.6 to 33% of patients experienced worsened function in the form of facial pain, or hypesthesia. The operative risk of trigeminal neuropathy is probably related various factors such as tumor size, preoperative dysfunction, surgical approach, and the extent of resection. Samii et al reported 83% gross total resection (GTR) rate and 33% rate of trigeminal neuropathy. ²⁶ Al-Mefty et al reported 28% risk of trigeminal neuropathy. ²⁷ Goel et al achieved GTR in 70% and reported 17.8% rate of trigeminal neuropathy. ⁵ Wanibuchi et al in a series of 107 patients reported GTR in 75% and trigeminal neuropathy in 11.6% patients. ¹² Other complications after surgery include cerebrospinal fluid (CSF) leakage and meningitis. Differences in outcomes between initial surgery and SRS may in part be related to the fact that surgery is more commonly performed for larger tumors (>10 cm ³ ) with brainstem compression. Residual trigeminal neuropathy after surgery may be an acceptable clinical outcome in such patients. Similarly, adjuvant resection is an important option for patients whose tumors have failed initial SRS.

Significance of the Present Study

In this study, we expand on our previous series to document the long-term clinical and tumor outcomes of TS following SRS. A total of 50 patients were treated with a median marginal dose of 14 Gy. The mean time to follow-up was 55 months and 18 of 50 patients had a follow-up greater than 5 years. At last follow-up, 94% had clinical stabilization or better, 47% had clinical improvement, and 92% had tumor control. Only one patient had symptomatic permanent ARE.

The results of this study add to the growing literature and also demonstrate the long-term value of SRS for TS. Specifically, the overall results of this study are nearly identical to the average values in the published literature for tumor control (91%), clinical improvement (48%), and time to follow-up (56 months). No patients developed secondary malignancies.

Interestingly, in this study we found that patients with primary SRS had a higher chance of and significantly faster time to clinical improvement. In addition, tumor regression predicted clinical improvement. While in this study we did not find any factor predictive of clinical improvement, a past study illustrated that a higher marginal dose was associated with improved tumor control. ⁸

Primary SRS and smaller tumor were also significantly associated with improved PFS. Notably, in the overall study, the 10-year PFS was 85.5%. However, in long-term follow-up patients (> 5 years), 10-year PFS was 88%. All seven patients with greater than 10-year follow-up, experienced tumor regression, and clinical stabilization or better at last follow-up. The results of this study show remarkably high, overall, short-term, and long-term tumor control and clinical stabilization or better rates. We believe these results demonstrate the high-treatment efficacy of SRS for TS.

Limitations

Although our patient series was derived from an institutional review board (IRB)–approved prospectively maintained database, this study is retrospective in nature. Further long-term follow-up is needed.

Conclusion

TSs are appropriate candidates to consider primary, as well as salvage SRS, after initial surgical removal. Radiosurgery led to a 92% tumor control rate and 94% clinical improvement or stabilization rate. A large multicenter study of SRS for TS is ongoing by the International Radiosurgery Research Foundation (IRRF), the results of which could provide additional valuable evidence on long-term clinical outcomes and tumor control.