Symptomatic posttreatment edema after stereotactic radiotherapy (SRS/FSRS) for intracranial meningiomas: patterns and predictive factors

Highlights

• •Symptomatic post posttreatment edema can be predicted in our study based on unmodifiable tumor and patient related factors.
• •Tumor size >30 mm and midline location were both statistically significant in univariate and multivariate analysis.
• •Other statistically significant factors suggested by univariate analysis were pretreatment edema and WHO histological grade 2.
• •Treatment technique, either SRS or fSRT was not associated with symptomatic post posttreatment edema occurrence.

Abstract

Background

Symptomatic posttreatment edema (SPTE) is a complication that may develop after radiotherapy for intracranial meningiomas. Our study aims at reviewing rates of SPTE in a large cohort of a single institution and identifying possible predictive factors.

Methods

We retrospectively analyzed data of 293 patients with 304 intracranial meningiomas irradiated at our institution between 2005 and 2018. We evaluated rates of SPTE and investigated numerous factors by univariate and multivariate analysis. Kaplan Meier analysis was used for estimation of actuarial local control and overall survival.

Results

Median age was 60 years. Meningiomas were treated with fractionated stereotactic radiation therapy (70 %), single fraction stereotactic radiosurgery (24 %) or fractionated stereotactic radiosurgery (6 %). Median imaging follow-up was 60 months, actuarial 10 year local control rate for patients with grade 1 meningiomas who received radiotherapy as definitive treatment was 99 %. Local control at 5 years was 94 % for grade 1 meningioma, 57 % and 53 % for grade 2 and 3 respectively. Sixteen patients (5.5 %) developed SPTE, median time to onset was 3 months (range 1–26 months). the higher rates of SPTE observed were in midline (13 %) and convexity (9 %), compared to skull base tumors (2 %). On univariate analysis, age > 60 years (p > 0.03), pretreatment peritumoral edema (p = 0.014), medline location (p = 0.018), tumor size > 30 mm (p = 0.015) and grade 2 histology (p = 0.03) were predictive of SPTE. On multivariate analysis, only tumor location and size remained statistically significant.

Conclusions

Based on our results, patients at high risk of SPTE can be identified based on patient and tumor characteristics. The best treatment technique in high risk patients is yet to be defined.

Introduction

Meningiomas account for about 30 % of primary brain tumors. They constitute a group of mostly benign neoplasms originating from arachnoidal cap cells which are most typically diagnosed in elderly women [1,2].

Since the first identification of meningiomas by Cushing [3], followed by the resection grades classification described by Simpson in 1957 [4], surgery has been the mainstay treatment. However, patient- and tumor-related limitations to surgery have led to the development of alternative treatment approaches [5]. Radiotherapy has been shown to improve local control in the adjuvant setting of subtotally resected or histologically aggressive tumors [6,7]. In addition, the emergence of stereotactic techniques and radiosurgery has made radiotherapy an upfront treatment of choice in selected cases [8,9]. With the increasing use of radiotherapy, reports on short- and long-term toxicities are of crucial importance to guide clinicians in treatment decision-making and complications follow-up. Most radiotherapy toxicity data relate to cranial nerve injury and posttreatment peritumoral edema [10,11]. Symptomatic posttreatment edema (SPTE) is a well-recognized complication that may lead to significant morbidity. Described rates in the literature vary from 1.4 % to 43 %, depending on the studied populations [[12], [13], [14], [15], [16], [17], [18], [19], [20]].

Few studies addressed the issue of SPTE and its possible predictive factors. They are all retrospective, including various tumor locations, sometimes excluding previous treatments or aggressive tumor grades. Furthermore, most published reports focus on radiosurgery, with a minority of patients in SPTE reports having received fractionated stereotactic radiotherapy [13,14,[21], [22], [23], [24], [25]].

Factors reported to be associated with SPTE include patient-related factors such as gender and age, tumor characteristics such as volume, tumor/brain interface area, location and presence of pretreatment edema, or treatment-related factors such as previous surgery, radiation dose and fractionation. Yet, results remain equivocal due to large interstudy heterogeneities.

Therefore, we aimed to analyze the experience at our institution and identify potential predictive factors of SPTE.

Materials and methods

Study population

We conducted a retrospective review of 300 patients who received radiotherapy for intracranial meningiomas in our department from January 2005 to December 2018. Inclusion criteria were as follows: patients aged 18 years or more; histological or imaging diagnosis of intracranial meningioma; radiotherapy indicated as adjuvant, definitive or salvage treatment; at least one imaging follow-up after radiotherapy.

Tumor location was defined as one of 4 main sites: convexity, midline (parasagittal and parafalcine), base of the skull, and other sites including tentorium and foramen magnum. Patients with optic nerve sheath meningiomas were excluded, reducing the number of evaluable patients to 293, with 304 tumors.

The size of the tumor was measured using the largest diameter on magnetic resonance imaging (MRI) or computed tomography (CT) scan. Histological grades were classified as per the 2007 World Health Organization (WHO) classification of meningiomas [26]. Meningiomas with no histological proof were presumed grade 1 for our statistical analysis if radiological signs didn’t show aggressive features, otherwise they were presumed grade 2. Institutional ethics review board approval was obtained for this study.

Radiation therapy planning

Unless contraindicated, all treatments were planned with the help of a volumetric contrast-enhanced MRI co-registered with a planning CT. Gross tumor volume (GTV) was defined as macroscopic tumor based on the combined imaging. When a fractionated stereotactic technique was used, surgical bed and/or GTV plus a margin of 5-20 mm considering anatomical barriers and histological grade was added to account for potential microscopic disease, thus generating the clinical target volume (CTV), and an additional margin of 3 mm was typically applied to generate the planning target volume (PTV). The CTV for adjuvant radiotherapy was based on both the tumor grade and the presence or absence of post-operative residual disease. In patients treated with stereotactic radiosurgery (SRS), a small margin of 0–2 mm could be applied to create the CTV at the discretion of the treating radiation oncologist but no additional PTV margin was added. The choice of the fractionation schedule depended mainly on grade, tumor size and distance to critical structures.

Intensity-modulated radiation therapy (IMRT) planning was used in all patients treated with fractionated stereotactic radiation therapy (FSRT). Radiosurgery was based on a linear accelerator (LINAC) with a Brown-Robert-Wells stereotactic frame until 2010, and Cyberknife® from 2010 to 2018.

Follow-up

Clinical and imaging follow-up utilized MRI or CT scans, scheduled at 3–4 months post-treatment, every 6 months for 2 years, and annually thereafter, with earlier imaging if clinically indicated. Patients with WHO grade 2 or 3 meningiomas underwent closer follow-up (2 months post-treatment and then every 3–4 months) based on clinical presentation. Symptomatic posttreatment edema (SPTE) was defined as radiologically confirmed new or worsening edema associated with neurological symptoms not attributable to other conditions. Imaging was interpreted by experienced neuroradiologists.

Statistical analysis

Imaging follow-up spanned from treatment completion to the last available imaging, and clinical follow-up extended to death or last contact. Local control (LC) and overall survival (OS) were measured from treatment completion to progression/last imaging or death, respectively, with Kaplan–Meier estimates. Prognostic factors for SPTE were analyzed using univariate and multivariate methods, including age, sex, tumor size, grade, location (midline vs. other), pretreatment edema, prior surgery/radiation, and fractionation (SRS vs. FSRT). Non-dichotomous variables used published cut-offs. Statistical tests included Student's t-test, Fisher’s exact test, and log-rank test for univariate Kaplan–Meier analysis (significance: p < 0.05). Significant variables underwent multivariate logistic regression (backward stepwise). Analyses used SPSS v23.0 (IBM).

Results

Characteristics of the study population

Patient characteristics are summarized in Table 1. The median age of our 293 patients was 60 years (18–88 years); the majority were female (72 %). Among 304 tumors, 60 % were located in the skull base, 17 % in the convexity, 15 % in the midline (9 % parafalcine and 6 % parasagittal), and 8 % in the posterior fossa/foramen magnum. Histology was unavailable for 124 tumors (41 %), presumed grade 1 except for one presumed to be grade 2 based on radiological brain invasion. For histologically confirmed cases (after biopsy or surgery), 35 % were grade 1 (n = 109), 21 % grade 2 (n = 62), and 3 % grade 3 (n = 9). Median tumor diameter was 26 mm. Pretreatment edema was present in 45 tumors (15 %). Previous treatments included surgery in 186 cases (61 %) and radiotherapy in 12 cases (4 %); one patient received prior chemotherapy. Radiation consisted of fractionated stereotactic radiation therapy (FSRT) in 214 cases (70 %), with a median dose of 54 Gy (30–60 Gy); dose per fraction varied from 1.8 to 3 Gy. Seventy-one patients (24 %) had single-fraction SRS with a median prescribed dose of 15 Gy (14–20 Gy), prescribed to a median 71 % isodose line (59–90 %). Nineteen patients (6 %) had fractionated SRS (FSRS) with a median total dose of 18 Gy (18–25 Gy) delivered in 3 to 5 sessions.

Table 1.

Characteristics in 293 patients with 304 meningiomas.

Attributes	Value
Age, yearsMedian (Range)	60 (18–88)
Sex, (F/M)	211/82
Tumor location, n (%)
Skull base	181 (60)
Midline	47 (15)
Convexity	52 (17)
Other	24 (8)
WHO histological/imaging grade, n (%)
Grade 1	232 (76)
Grade 2	63 (21)
Grade 3	9 (3)
Median Tumor diameter, mm (Range)	26 (0–96)
GTV Median,cc (Range)	10.5 (0.3–190)
CTV Median,cc (Range)	49 (3–363)
PTV Median,cc (Range)	66 (0.6–478.2)
Tumors with pretreatment edema, n (%)	45 (15)
Previous treatment, n (%)
Surgery	186 (61)
Radiotherapy	12 (4)
None	106 (35)
Fractionation scheme, n (%)
FSRT 54 Gy (30–60)	214 (70)
SRS 15 Gy (14–20)	71 (24)
FSRS 18 Gy (18–25)	19 (6)

Abbreviations: F: female, M: Male, WHO: World health organization, FSRT: fractionated stereotactic radiation therapy, SRS: stereotactic radiosurgery, FSRS: fractionated stereotactic radiosurgery.

Symptomatic posttreatment edema

SPTE occurred in 16 patients (5.5 %) with a median onset at 3 months (range: 1–26 months). There were 4 men and 12 women. Median age was 73.5 years (42–86), and median tumor size was 35 mm (7–81). Locations included skull base (4), convexity (5), midline (6: 5 parafalcine and 1 parasagittal), and posterior fossa (1). Histological grade was 1 (proven and presumed) in 9 tumors and grade 2 in 7 tumors. Eight patients had previous surgery, and 2 had previous radiotherapy. FSRT was used in 11 cases (54–60 Gy in 30–33 fractions), and 5 tumors were treated with single-fraction SRS, with a prescribed dose of 15 Gy to a median 70 % isodose line. Six patients had peritumoral edema pre-treatment, while the other 10 developed de novo edema.

Neurological symptoms included motor deficits (37 %), seizures (31 %), headache (26 %), and visual disturbances (12 %). Associated asthenia was noted in 2 patients. Four patients had more than one symptom.

Nine patients (56 %) were hospitalized. Median symptom duration was 3.6 months (range: 1–13 months). Management included corticosteroids (all), antiepileptics (5), and salvage surgery (3). Histological study of resected tumors revealed grade 2 in 2 cases and grade 1 in the other. One patient aged 76 with a grade 2 convexity meningioma died of sepsis while on corticosteroids 3 months after SPTE onset. His last imaging showed extensive edema with a stable tumor. Another patient aged 73 with a grade 2 parafalcine meningioma died 6 months after symptom onset. Imaging showed progressive edema with signs of radionecrosis. No patient received bevacizumab. Two illustrative cases are presented in Fig. 1, Fig. 2.

Fig. 1.

A: Axial T1-weighted MRI of a right convexity meningioma. B: Axial T1-weighted MRI showing de-novo edema 6 months after single fraction radiosurgery (15 Gy to the 81 % isodose line). C: Axial T2 FLAIR-weighted MRI showing extensive edema. The 83 years old patient was successfully managed with corticosteroids.

Fig. 2.

A: Axial T2 FLAIR-weighted MRI of a left parafalcine occipital meningioma with peri-tumoral edema in a 75 years old patient. B: Axial T2 FLAIR-weighted MRI showing edema progression 2 months after fractionated stereotactic radiotherapy (54 Gy in 30 fractions), patient had to undergo salvage surgery, pathology showed a WHO grade 1 meningioma.

The occurrence of SPTE was associated with worse LC (p = 0.04) and worse OS (p < 0.0001). The LC at 2 and 5 years was 78 % and 70 % for patients who developed SPTE, compared to 96 % and 87 % for the rest of the cohort. The OS at 2 and 5 years was 74 % and 59 %, compared to 99 % and 93 % for the rest of the cohort.

Predictive factors of SPTE

Univariate analysis showed that age > 60 years was statistically significant for SPTE (p = 0.03). Patients with SPTE had a median age of 73.5 years (42–86 years) compared to 59 years (18–88 years) for the rest of the cohort. Pretreatment edema was also significant (p = 0.014), with 15 % of these patients developing SPTE. Regarding tumor location, 13 % of patients with midline tumors developed SPTE, compared to 9 % in convexity and 2 % in skull base tumors (p = 0.018). Other significant factors included grade 2 tumors (p = 0.03) and tumor size > 30 mm (p = 0.015).

Gender, previous treatments, and the fractionation scheme were not statistically significant. Among the 6 patients with midline tumors who developed SPTE, 5 received FSRT and 1 had SRS; this difference was not statistically significant. Similar results were observed for convexity tumors treated with SRS or FSRT, suggesting SPTE occurrence is likely unrelated to fractionation scheme. Further analysis of prior surgery in patients with convexity and midline tumors showed no significant association with SPTE (p = 0.3).

Multivariate analysis confirmed only midline location (p = 0.009) and tumor size > 30 mm (p = 0.006) as predictive factors. Statistical analysis details are provided in Table 2.

Table 2.

Univariate and multivariate analysis of possible SPTE predictive factors.

Variable	Univariate analysis	Multivariate analysis
	p Value	p Value	OR (CI 95 %)
Age (≤/> 60 years)	0.03	na	na
Sex (F/M)	0.7	na	na
Tumor size (≤/> 30 mm)	0.015	0.006	0.086 (0.015–0.501)
WHO histological/imaging grade (grade 1/grade 2)	0.03	na	na
Tumor location, (medline/others)	0.018	0.009	6.690 (1.591–28.131)
Pre-treatment edema (yes/no)	0.014	na	na
Previous surgery (yes/no)	0.3	na	na
Radiation therapy fractionation (FSRT/SRS)	0.5	na	na

Abbreviations: OR: Odds Ratio, CI: Confidence interval, F: female, M: Male, WHO: World Health Organization, FSRT: fractionated stereotactic radiation therapy, SRS: stereotactic radiosurgery, na: not applicable.

Discussion

Peritumoral brain edema is common in meningiomas [27,28] and often due to a disrupted blood–brain barrier and increased VEGF-A production [29,30].

In our study, 15 % of tumors had surrounding edema before radiotherapy. Post-treatment, either worsening of pre-existing edema or new edema can occur. While the exact mechanisms of radiation-related injury remain unclear, radiation necrosis and complex inflammatory responses involving high levels of VEGF and hypoxia-inducible factor 1 (HIF-1) proteins have been implicated in peritumoral edema after SRS. In parasagittal meningiomas, vascular damage and venous thrombosis are potential mechanisms [[31], [32], [33]].

Although prospective randomized data on the role of radiation therapy in meningioma treatment are lacking, many retrospective studies have demonstrated favorable local control rates and low toxicity rates [11,34,35]. Our results are consistent with this literature, with a 99 % actuarial LC rate at 10 years in patients with grade 1 tumors treated exclusively with radiotherapy. OS rates at 5 and 10 years were 94 % and 91 %, respectively.

The rate of SPTE in our cohort was 5.5 %, with a median time of onset of 3 months (range: 1–26 months). Patients treated with SRS developed edema after a median of 6 months (range: 1–6). These findings align with prior reports, where median onset time ranged from 4 to 9 months, with some cases occurring as late as 48 months post-radiation [13,23,24,36]. In our cohort, SPTE rates were 7 % after SRS and 5 % after FSRT. Literature reports show similar variability, with SPTE rates after SRS ranging from 2 % to 43 % [14,15,19,24]. In one study of 173 patients treated with 1 to 5 fractions using CyberKnife or Gamma Knife, the SPTE rate was 8 % [23]. Pollock et al. reported a 2 % rate in a larger cohort of 330 patients treated with SRS [19]. In another study using multiple fractionation schemes, Fokas et al. reported a 37 % rate of SPTE in patients treated with single-fraction SRS (n = 16; 11 in falx cerebri, 5 in convexity), while no SPTE occurred in the remainder of the cohort treated with fractionated schedules (4–5 Gy per fraction) [18].

In our study, the fractionation scheme was not associated with SPTE risk regardless of tumor location. Notably, among patients treated with fractionated SRS, no SPTE cases were observed after a median follow-up of 42 months. Though our sample size was limited, this supports previous reports suggesting a potential protective role of fractionation [12,37]. However, a recent report of 245 meningiomas treated with single-fraction SRS, fractionated SRS (2–5 fractions), or hypofractionated stereotactic radiotherapy (6–15 fractions) did not find any fractionation schedule to be protective [36].

A recent meta-analysis found that the incidence of symptomatic edema was higher in patients treated with SRS (17.4 %) compared to FSRT (4 %). The authors hypothesized that fractionation allows better sparing of normal tissue while delivering high doses with precision, especially in irregularly shaped tumors [38].

Our findings also suggest that midline meningiomas are more likely to result in SPTE. This association was statistically significant in both univariate and multivariate analyses, echoing prior reports [36,39]. For instance, in a cohort of 102 SRS-treated patients, parasagittal meningiomas had a 4-fold increased risk of SPTE [24]. Chang et al. hypothesized that hemispheric meningiomas tend to grow deeper into the cortex, unlike skull base tumors which spread along the dura, limiting the risk of edema [40]. Similarly, Cai et al. reported a 17 % increase in SPTE risk for every 1 cm² increase in tumor–brain surface area [13].

Pretreatment edema was another significant predictor of SPTE. Sheehan et al. and others have demonstrated that patients with pre-existing edema are more likely to experience progressive edema post-radiotherapy [16,21,41,42]. Age also correlated with SPTE risk in univariate analysis, consistent with Novotny et al.’s findings using a 60-year threshold [22]. Tumor size > 30 mm was significantly associated with SPTE in our study; Morimoto et al. identified a close threshold of 25.6 mm in their analysis of three- to five-fraction SRT regimens [43]. Other studies have reported various thresholds [21,39]. WHO grade 2 tumors were also associated with SPTE in univariate analysis, similar to the findings of Conti et al. [36].

Several dosimetric parameters have been studied as SPTE predictors. One study found no significant association between SPTE and the volume of brain receiving 8–12 Gy [15]. The prescription, margin, or maximum doses were significant in some studies [14] but not others [24,35]. In a recent review of 26 studies, Milano et al. found that larger tumor size, higher prescription dose, non–skull base location (especially parasagittal), absence of prior resection, and pre-treatment edema increased SPTE risk—though inconsistently across studies [41,44].

SPTE management typically involves corticosteroids and supportive care. All our patients received corticosteroids; some also required antiepileptics or salvage surgery. One patient died of sepsis, likely related to corticosteroid use; another died of progressive edema. Both had WHO grade 2 tumors. While SPTE was associated with poorer LC and OS in our analysis, causality is uncertain due to confounding factors like corticosteroid side effects, tumor grade, and age.

Most SPTE cases in Patil et al.’s study responded to corticosteroids [24]. Conversely, Kalapurakal et al. reported five non-responsive cases, including one fatality from neurological complications [37]. Tanzler et al. also reported a corticosteroid-related death in their FSRT cohort [20]. Bevacizumab, an anti-VEGF agent, has shown promise in small series of corticosteroid-refractory cases, particularly when radiation necrosis was present [45,46].

Limitations of our study include its retrospective design, heterogeneous population, and potential selection bias. However, this diversity reflects real-world clinical practice, where patient- and tumor-related factors vary widely. Some results, especially those with borderline significance, should be considered hypothesis-generating and warrant prospective validation.

Conclusion

In our experience, the risk of symptomatic posttreatment edema (SPTE) is relatively low, but it can lead to significant morbidity and, in rare cases, mortality.

Risk factors identified by our analysis were tumor-related (midline location, tumor size > 30 mm, presence of pretreatment edema, and WHO grade 2) and patient-related (age > 60 years). The fractionation scheme did not significantly influence SPTE risk.

Larger, multi-institutional studies are needed to further validate these predictive factors and determine optimal treatment strategies in high-risk patients. We advocate for a collaborative effort to collect standardized clinical and imaging data across centers using a secure, internet-based platform.

Funding

No funding was received for this work.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Houda Bahig and David Roberge received research grant from Varian Medical Systems, and speakers Honoraria from Siemens Healthineers/Varian Medical Systems, unrelated to the current work. David Roberge also reports honoraria and research funding from Accuray, BrianLab and Elekta unrelated to the current work. Carole Lambert reports honoraria in relation to a lecture for Abbvie and advisory role for Ferring and Sanofi unrelated to the current work. Giuseppina Laura Masucci also reports Astrazeneca honoraria et cerapedics honoraria unrelated to the current work. Cynthia Ménard reported research fundings from Promaxo, Lantheus, Varian, and Tersera unrelated to the current work.

No other disclosures are declared.