Clinical Outcomes of Salvage Microsurgical Resection after Radiation Therapy for Sporadic Vestibular Schwannomas

Abstract

Background  The purpose of this study was to evaluate clinical outcomes after salvage microsurgery for vestibular schwannoma (VS) treated initially with modern radiotherapy techniques as compared to those treated with primary microsurgical resection.

Methods  Patients who underwent microsurgical resection of sporadic VS undergoing a translabyrinthine approach at a single academic skull base tertiary referral center were included. Baseline and postoperative dynamic gait index, functional gait assessment, House–Brackmann facial nerve grading, and completeness of resection were prospectively recorded.

Results  Of the 265 patients reviewed, 21 (7.9%) patients underwent prior radiation. Median age of the cohort was 55 years (interquartile range: 51–63). The likelihood of achieving a gross total resection was significantly lower for radiated as compared to nonradiated patients (odds ratio: 0.18, 95% confidence interval: 0.05–0.53, p  = 0.004) when controlling for tumor size. Radiated patients had better postoperative facial nerve function on the first postoperative day, but this difference was not significant at long-term follow-up. Radiated patients had lower preoperative postural stability scores than nonradiated patients on FGA (26 vs. 23, p  = 0.035). Postoperatively, radiated patients had comparable outcomes compared to nonradiated patients when controlling for age and tumor size.

Conclusion  Compared to patients with VS treated with surgery alone, previously radiated patients are less likely to achieve gross total resection in the salvage setting. Radiated patients scored better on facial nerve outcomes compared to nonradiated patients in the initial postoperative period but demonstrated similar long-term outcomes.

Introduction

While the incidence of vestibular schwannomas (VSs) has been stable (1.09–1.4 per 100,000) in recent years, rates of microsurgery as a primary treatment modality have decreased, while rates of radiation therapy (RT) as a primary treatment remained unchanged. ^{1 2 3} With increased knowledge of the natural history of VS, this trend toward conservative primary treatment has persisted across multiple cohorts. ^{4 5} RT has been shown to have good success rates, but failure and tumor growth occur in 3 to 10% of cases. ^{6 7}

In the subset of patients who have failed primary radiation treatment, less is known about optimal approach to treatment. While reported rates of primary failure after RT are low, recent meta-analyses have shown higher rates of secondary failure after repeat RT, and the next step is often salvage microsurgical resection. ^{8 9 10} These salvage cases are known to be more technically challenging and have been associated with higher incidence of adverse clinical outcomes compared to primary surgical treatment alone. ^{8 11 12} Whitmeyer et al note that patients who underwent salvage microsurgery had gross total resection in only 55.7% of cases, and those who had subtotal resection were associated with lower rates of facial nerve injury. ¹³ In this context, the purpose of this study was to evaluate clinical outcomes after salvage microsurgery for VS treated initially with modern radiotherapy techniques as compared to those treated with primary microsurgical resection.

Methods

Following IRB (institutional review board) approval (IRB# 180978), a retrospective review analyzed patients who underwent microsurgical resection of sporadic VS between November 1, 2017 and November 1, 2022 at a single academic skull base tertiary referral center. These individuals were subdivided into those who had primary radiation failure with subsequent salvage microsurgery, and those undergoing primary surgery with no history of radiation treatment. Indications for salvage microsurgery were persistent tumor growth following radiation. Only patients who underwent a translabyrinthine approach for previously untreated VS were included. Patients with a history of neurofibromatosis type 2 or history of prior craniotomy were excluded. Primary outcomes were assessed by completeness of resection abstracted from operative reports. Secondary outcomes of baseline and postoperative dynamic gait index (DGI), functional gait assessment (FGA), and House–Brackmann (HB) facial nerve grading were prospectively recorded. These gait-related outcomes are described in detail by Moshtaghi et al, but involve tasks like walking with a narrow base of support, ambulating backwards and with eyes closed. ¹⁴ Facial nerve assessments were performed within 7 days postoperatively. Complications assessed were postoperative hematoma, cerebrospinal fluid leak, new facial weakness, or new cranial nerve deficits.

Statistical analysis was performed with Fischer's exact tests and Pearson's Chi square tests. A multivariate linear regression model was used with covariates of patient's age, tumor size, preoperative radiation, and facial nerve function both pre- and postoperatively. All analyses were performed using R, version 4.2.1 (R Foundation for Statistical Computing) statistical software.

Results

Of the 551 patients reviewed, 265 underwent a translabyrinthine approach. Twenty-one of these patients (7.9%) underwent radiation prior to microsurgery. Median age of the cohort was 50 years with salvage microsurgery patients on average 5 years older than nonradiated patients ( p  = 0.026). The majority of the cohort was composed of female patients (60.7%; Table 1 ). Median largest tumor dimension by longest linear dimension was 25 mm in the radiated group and 26 mm in the nonradiated group ( p  = 0.7).

Table 1. Preoperative characteristics for the nonradiated and salvage microsurgery cohorts.

Preoperative characteristic	Overall, N  = 265 a	Nonradiated, N  = 244 a	Salvage microsurgery, N  = 21 a	p -Value b
Age	50 (39–61)	50 (39–60)	55 (51–63)	0.026
Sex				0.3
Female	161 (62%)	146 (61%)	15 (75%)
Male	97 (38%)	92 (39%)	5 (25%)
Dynamic gait index	22.00 (21.00–23.00)	22.00 (21.00–24.00)	21.00 (20.00–22.50)	0.049
Unknown	81	75	6
Functional gait assessment	26.0 (23.0–28.0)	26.0 (23.0–28.2)	23.0 (22.0–25.5)	0.035
Unknown	78	72	6
MRI maximum linear dimension in mm	25 (18–32)	25 (17–32)	26 (23–28)	0.7
Unknown	14	13	1

Note: Significant p -values in bold.

^an (%); median (IQR).

^bFisher's exact test; Wilcoxon rank sum test; Pearson's Chi-squared test.

Among radiated patients, four (19%) achieved gross total resection, while 144 (62%) of nonradiated patients achieved gross total resection ( Table 2 ). In a multiple regression model, the likelihood of achieving a gross total resection was significantly lower for radiated as compared to nonradiated patients (odds ratio: 0.18, 95% confidence interval; 0.05–0.53) when controlling for tumor size ( Table 3 ). Regarding postoperative facial nerve function, 19 radiated patients (86.4%) exhibited HB level I or II facial nerve outcomes compared to 83.1% of nonradiated patients ( p  = 0.033). In a multivariable model correcting for age and tumor size, this relationship was not significant. Similarly, there was no significant difference between facial nerve outcomes (percentage of patients demonstrating HB I or II) at the last follow-up when correcting for these covariates.

Table 2. Postoperative characteristics between the nonradiated and salvage microsurgery cohorts.

Postoperative characteristic	Overall, N  = 265 a	Nonradiated, N  = 244 a	Salvage microsurgery, N  = 21 a	p -Value b
Surgical resection				<0.001
Gross total resection	148 (58%)	144 (62%)	4 (19%)
Near total resection (<5% of original volume)	4 (1.6%)	3 (1.3%)	1 (4.8%)
Subtotal resection (5–50% of original volume)	93 (36%)	81 (35%)	12 (57%)
Partial resection (>50% of original volume)	10 (3.9%)	6 (2.6%)	4 (19%)
Inpatient length of stay (hours)	79 (57–84)	79 (57–84)	67 (55–81)	0.073
Unknown	6	5	1
Thirty-day complications				0.2
No return	207 (81%)	193 (82%)	14 (70%)
Return to ED without admission	15 (5.9%)	12 (5.1%)	3 (15%)
Return to ED with admission	34 (13%)	31 (13%)	3 (15%)
Unknown	9	8	1
Post-op dynamic gait index	21.0 (18.0–22.0)	21.0 (18.0–22.0)	21.0 (19.0–22.5)	0.6
Unknown	84	78	6
Post-op functional gait assessment	22.0 (19.0–26.0)	22.0 (19.0–26.0)	23.0 (20.0–25.8)	0.5
Unknown	81	74	7
Percent of patients with HB I or II on POD1	217 (84%)	198 (83%)	19 (90%)	0.5
Percent of patients with HB I or II at most recent visit	226 (85%)	208 (85%)	18 (86%)	> 0.9

Abbreviations: HB, House–Brackmann, POD, postoperative day.

Note: Significant p -values in bold.

^an (%); median (IQR).

^bFisher's exact test; Wilcoxon rank sum test; Pearson's Chi-squared test.

Table 3. Multivariable regression model showing lower odds of gross total resection in the salvage microsurgery cohort after controlling for age and tumor size.

Coefficients	OR	95% CI	p -value
Facial nerve function (HB I or II) POD 1
Preoperative radiation	1.13	0.96–1.34	0.2
Age	1.00	1.00–1.00	0.7
Tumor maximum linear dimension pre-op	0.99	0.99–1.00	0.001
Facial nerve function (HB I or II) at most recent visit
Preoperative radiation	1.06	0.90–1.24	0.5
Age	1.00	1.00–1.00	0.6
Tumor maximum linear dimension pre-op	0.99	0.99–1.00	<   0.001
Gross total resection
Preoperative radiation	0.18	0.05–0.53	0.004
Age	0.96	0.94–0.99	0.001
Tumor maximum linear dimension pre-op	0.89	0.86–0.92	<   0.001

Abbreviations: CI, confidence interval; HB, House–Brackmann; OR, odds ratio; POD, postoperative day.

Note: Significant p -values in bold.

Pre- and postoperative gait and functional mobility were also assessed. Radiated patients had worse baseline physical function than nonradiated patients with the FGA being statistically significant (23 vs. 26, p  = 0.035). Similarly, the DGI showed worse preoperative metrics in radiated patients (21 vs. 22, p  = 0.049). In postoperative analysis, there was no significant difference between radiated and nonradiated patients in gait and functional mobility indices.

There were no significant differences in length of stay or complications between the radiation and radiation-naïve groups.

Conclusion

Compared to patients with VS treated with surgery alone, previously radiated patients are less likely to achieve gross total resection in the salvage setting. Compared to established literature, the lower odds of gross total resection observed in this cohort may reflect more difficult planes, increased tumor adherence to the nerve and increased sensitivity of the facial nerve to mechanical stimulation. ¹³ These challenges in tumor dissection in previously radiated patients has been noted by multiple previous studies. ^{8 11 12}

A second important takeaway from this analysis is that previously radiated patients scored worse on preoperative postural stability indices than nonradiated patients. Although preoperative balance indices are undoubtedly influenced by patients' baseline functional status prior to radiation treatment, our finding indicates a degree of balance disturbance associated with a history of radiation treatment and likely reflects an incomplete unilateral vestibular insult. This is supported by existing literature. Murphy et al noted that in a cohort receiving gamma knife radiation, development of a new gait dysfunction occurred in 18% of patients while 15% developed worsened gait dysfunction. ¹⁵ Similarly, Tuleasca et al in a cohort of post-radiated patients report new-onset vertigo and gait disturbances in 11.4 and 8.6% of patients respectively, while worsening gait disturbance occurred in 20% of patients. ¹⁶ Interestingly, postoperatively following completion labyrinthectomy, post-radiated patients had a significant improvement in functional gait. These findings suggest an easier ability to compensate following a complete vestibular loss. ^{16 17} Notably, this disturbance after radiation treatment does not appear to be dose-dependent. ¹⁸

Lastly, this work shows that after controlling for age and tumor size, there is no difference in postoperative facial nerve outcomes between radiated and radiation-naïve patients. The worse facial nerve outcomes in the early postoperative period for radiation-naïve patients, especially in the context of lower rates of gross total resection discussed above, may reflect a more conservative approach to tumor debulking that results in spared facial nerve function at the expense of leaving residual tumor. This can be further explained with the increased rate of tumor adhesion to the facial nerve and brainstem after radiation treatment reported in the literature. ^{19 20 21 22} This cohort's low rates of facial nerve injury in the setting of subtotal resection are similar to previously published work and may indicate a role for planned conservative resection in salvage cases to preserve facial nerve function. While multiple studies have explored recurrence rates in nonradiated VS, few have reported outcomes in this subset of the patient population. ^{3 15} At this point, long-term outcomes for the more conservative approach in salvage microsurgery remains to be seen.

This study has several limitations. The first is the limited follow-up period after surgery. There is need for long-term evaluation and monitoring for delayed-onset adverse events and possible tumor recurrence. Second, this study's statistical power is limited by the small sample size. Although our center has a large catchment area and significant surgical volume, our population is likely not perfectly representative of the national VS population and may carry different genetic and environmental factors that influence outcomes. Lastly, this study is limited in its categorization and quantification of tumor residual. Postoperative surveillance computerized tomography scans are this group's standard of care but are unable to capture residuals like magnetic resonance imaging. Further studies involving large, prospective, and long-term cohorts are needed to fully evaluate outcomes in this population.