Fractionated stereotactic radiotherapy in the treatment of exclusive cavernous sinus meningioma: functional outcome, local control, and tolerance

1. Introduction 

Meningiomas are tumors arising from arachnoid cells that account for approximately 15% to 20% of all intracranial neoplasms in largest series [1]. It is well known that surgical removal remains its main treatment and that radical resection is the only curative therapy for them [35]. Recurrence rate after surgery varies from 9% at 10 years when the tumor is removed with involved dura and bone to 40% if only partial debulking is performed [34].

However, gross total removal is not always feasible in cranial base tumors, particularly in the cavernous sinus region, and it is associated with a high rate of postoperative neurologic deficits, ranging from 19% to 86% in recently reported series. Even in centers with the most experienced hands, surgery of cavernous sinus meningiomas remains a considerable challenge for surgeons [5], [10], [33], [34], with total resectability rates ranging from 20% to 80% and bad outcomes [11], [14], [34].

In the last 2 decades, it has been supported the use of radiotherapy in these tumors [22]. The role of conventional fractionated radiation therapy is usually reserved for residual or unresectable disease, for recurrent lesions, for tumors of malignant histology, and for patients with bad medical condition in which surgery is contraindicated. Recent series have demonstrated that this technique enhances both recurrence-free survival and overall survival rates after subtotal resection, as compared with surgery alone [13], [14], [21], [34].

Because FSRT combines the precision of stereotactic positioning with the radiobiologic advantage of fractionation, it might reduce the volume of normal tissue irradiated, leading to a reduction in long-term toxicity [2], [3], [8], [12], [13], [29]. Fractionation is less damaging to normal brain when compared with single-fraction treatment. This may be of particular relevance in cavernous sinus meningiomas as they enclose cranial nerves and lie near other critical structures as the brain stem or the optic chiasm [3].

We present our experience with this technique for meningiomas of the cavernous sinus region with primary end points of local control, clinical response, and radiation-induced toxicity. Because of our relatively short follow-up, results should be considered preliminary.

2. Patients and methods 

From June 1997 to June 2001, 30 consecutive patients with cavernous region meningiomas were treated in the Catalan Institut of Oncology and Bellvitge Hospital in Barcelona. All patients were evaluated before treatment by CT and MRI. Maximal tumor diameter, tumor volume, and distance to critical neural structures (optic pathway or brainstem) were assessed in all cases. Lesions were stratified according to an anatomic scoring of tumor extension adapted from Sekhar and Altschuler [32] (Table 1). The KPS scale was recorded at the beginning of FSRT and during follow-up.

Table 1. Classification of intracavernous neoplasms (from Sekhar and Altschuler [32])

	Grade	Cavernous sinus involvement	Intracavernous ICA
	I	One area only (A, P, L, or M)	Not involved
	II	More than one area	Displaced, not totally encased
	III	Entire CS	Totally encased, at least a short length
	IV	Entire CS	Encased, with narrowing
	V	Bilateral CS	Encased

A indicates anterior; P, posterior; L, lateral; M, medial. ICA, internal carotid artery; CS, cavernous sinus.

We defined late toxicity as previously reported, that is, onset or worsening of neurologic symptoms beyond 3 months after the completion of treatment, not attributable to disease progression by radiographic or clinical criteria [20].

All patients were treated under the same protocol and as outpatient procedures. Patient immobilization was achieved using the BrainLAB (Munich, Germany) after 1999 (20 patients) or Beverly frame before 1999 (10 patients) systems. The Cosman-Roberts-Wells stereotactic frame was used for isocenter definition. A 3-dimensional CT data cube generated from continuous 2-mm CT scans was obtained. In selected cases before 1999 and in all patients afterward, gadolinium-enhanced MRI for image fusion was performed. Treatment was delivered with a 6 MV linear accelerator (Varian 2100C/600C, Varian Medical Systems, CA) to a median dose of 52 Gy at isocenter (range, 50-56 Gy), 2 Gy per fraction, and a median of 26 fractions (Table 2). Contrast-enhanced area was defined as GTV, and a margin of 2 mm was defined as PTV.

Table 2. Treatment parameters

	Parameter	Median	Min	Max	Mean	SD
	Prescription dose (Gy)	52	50	56	52.40	2.06
	Isodose line (ICRU 50) (%)	100	100	100	100	0
	Number of fractions	26	25	28	26.17	1.05
	Dose per fraction (cGy)	200	200	200	200	0
	Treatment volume (cm3)	11.31	0.90	59.68	15.31	13.69

Optimal dose distribution with dose homogeneity within the target volume and minimal dose to normal brain and critical structures were achieved with fixed noncoplanar beams. All patients had conformal blocking with individual lead alloy. International Commission on Radiation Units and Measurements criteria were applied, which means that doses were calculated on the 100% isodose at the cross of beams at isocenter (minimum dose at PTV, 95%; range, 95%-107%).

Dose planning was executed by using ISSIS-2 (Technology Diffusion, Paris, France) treatment planning software before March 1999 and BrainSCAN treatment planning system software afterward. Both systems have been previously compared by the authors with minimal differences in total dose, dose per fraction, precision, and sparing of critical structures [40].

The main goals of treatment planning were coverage of all tumor volume with prescription dose and maximal limitation of dose to critical structures nearby.

Follow-up data were derived from chart data and interviews with patients. Imaging follow-up consisted of half-yearly contrast-enhanced MRI during the first 5 years after treatment, followed by yearly studies thereafter. Progression was defined as: 25% increasing of maximal diameter of the tumor, worsening of clinical symptoms, or both. Mini-mental examination (MME) was performed at the last follow-up in 24 patients.

3. Results 

There were 7 men and 23 women (ratio 1:3.2), with ages ranging from 34 to 77 years (median, 59). Clinical symptoms at the time of treatment are listed in Table 3. Oculomotor cranial deficit was the most frequent initial symptom in 17 patients (56.6%), followed by visual impairment in 14 patients (46.6%) and trigeminal neuralgia in 12 patients (40%).

Table 3. Clinical symptoms before treatment

	Symptoms	Patients (n [%])
	Headache	6 (20%)
	Ophthalmoparesis	17 (56.6%)
	Trigeminal neuralgia	12 (40%)
	Visual acuity impairment	14 (46.6%)
	Dizziness	1 (3.3%)
	Exophthalmos	4 (13.3%)
	VII cranial nerve paresisa	3 (10%)
	Instability	3 (10%)
	Hypopituitarism	1 (3.3%)
	Epileptic seizures	1 (3.3%)

a Patients submitted to previous surgery.

Seventeen patients (56.6%) had undergone cytoreductive surgery before radiotherapy. All these patients had WHO grade I, meningothelial subtype meningiomas. The remaining 13 patients had been discarded as surgical candidates by their neurosurgeons based on clinical criteria (past medical condition, age, etc), and for that reason, they were treated de novo. In these patients, diagnosis was made radiologically (based on CT and MRI findings) and with respect to the behavior of the lesion. These patients received FSRT as primary treatment. One patient had received previous conventional radiotherapy.

Median follow-up period was 50 months (range, 28.2-74.5 months). No patient was lost to follow-up. Tumor and patient characteristics are shown in Table 4. Of the 30 treated lesions, 23 (76.6%) had a maximum diameter superior to 30 mm (median, 42 mm; range, 15-60) before FSRT, and 73.3% of them belonged to categories III, IV, and V of Sekhar's classification.

Table 4. Patient population and tumor characteristics

	Classification	n
	Age (y)
	Median	59
	Range	34-77
	Sex
	Female	23
	Male	7
	Follow-up (mo)
	Median	50
	Range	28.2-74.5
	Sekhar classification gradea
	I-II	8 (26.6%)
	III-IV	21 (70%)
	V	1 (3.3%)
	Side
	Right	17 (56.6%)
	Left	13 (43.3%)
	Tumor size (maximal diameter)b
	Small	7 (23.3%)
	Large	19 (63.3%)
	Very large	4 (13.3%)
		Max	Min	Median
	Volume (cm3)	59.68	0.90	11.31
	Closest distance to chiasm (mm)c	30	0	0
	Closest distance to brainstem (mm)c	20	0	0

a See Table 1. b Small if maximal diameter < 3cm. Large if maximal diameter ≥3 cm and ≤ 5cm. Very large if maximal diameter >5cm. c Values obtained from 93.3% of patients. n: number of patients.

Severe acute toxicity was not seen in our series. Forty-six percent of patients complained of transient asthenia, dizziness, and headaches. All these symptoms were mild and resolved after conclusion of treatment.

Among the 28 patients with clinical symptoms at the time of treatment, 14 (50%) exhibited improvement of their clinical status. This improvement was attributable mostly to regression of diplopia and trigeminal neuralgia (less facial pain or tapering of medication) (Fig. 1). On the other hand, neurologic deterioration was seen in 2 patients (6.6%) after treatment; one of them experienced worsening of trigeminal neuralgia and the other showed visual acuity impairment verified by ophthalmologic examination. Regarding performance status, 11 patients (39.28%) improved their KPS score. Sixteen patients (53.3%) had KPS scores of at least 80 after treatment.

View Large Image Download to PowerPoint

Fig. 1. Clinical symptoms and response to treatment. *VII cranial nerve; **Visual acuity; ***V cranial nerve.

During follow-up, 2 patients (6.6%) showed progression of the treated tumor. Both cases had histopathologically proven WHO grade I meningiomas. Reduction of the tumor was seen in 6 patients (20%). Late radiation toxicity was seen in 2 patients, which consisted of moderate short-term memory loss and dysphasia in one case and neuropsychologic deficit with seizures in the other. These patients had an MME score of less than 25 points and failed in attention, memory, speech, and calculation and harbored left-sided meningiomas. The cranial MRI performed at the time of the neuropsychologic deficit did not show any change of the radiological signal in the temporal lobe. Other outcome factors such as pituitary insufficiency or other late long-term effects such as delayed oncogenesis were not seen in this series, perhaps because of the relatively short follow-up.

Age (OR, 0.200 [95% CI, 0.020-1.978]), previous cytoreductive surgery (OR, 5.000 [95% CI, 0.506-49.438]), treatment dose (OR, 2.059 [95% CI, 0.202-20.959]), tumor volume (OR, 0.423 [95% CI, 0.065-2.766]), or stratification (OR, 0.667 [95% CI, 0.097-4.605]) did not show statistically significant influence on radiological tumor control. When analyzing clinical response, none of these same variables had influence on clinical response or radiation therapy side effects: OR, 0.395 (95% CI, 0.087-1.797) for age; OR, 0.817 (95% CI, 0.190-3.505) for previous cytoreductive surgery; OR, 0.692 (95% CI, 0.136-3.518) for treatment dose; OR, 0.762 (95% CI, 0.179-3.241) for tumor volume; and OR, 0.692 (95% CI, 0.136-3.518) for stratification according to Sekhar's classification. All patients in this series were alive at the time of analysis.

4. Discussion 

Meningiomas are the most common tumors affecting the cavernous sinus. The management of these continues to be controversial because of their potential morbidity. It is certain that total resection of a meningioma in this area, such as anywhere else, is superior to any other modality of treatment [11]. However, more than two thirds of the patients in our series (76.6%) harbored lesions which had a maximum diameter superior to 30 mm, and 73.3% of them belonged to categories III, IV, and V of Sekhar's classification. Achievement of all the goals of open surgery is especially difficult in these cases in which critical structures nearby are displaced or even encased by the tumor, with a high rate of complications in even the most experienced hands [7]. On the other hand, that most of our patients had tumors with borders lying less than 3 mm from the optic pathway or brainstem would have discarded them from most radiosurgical protocols with single dose.

4.2. Radiosurgery vs FSRT 

Fractionated stereotactic radiotherapy combines the precision of stereotactic positioning with the radiobiologic advantages of fractionation, allowing higher total doses and safe target coverage; FSRT uses radiobiologic principles, which allow normal tissue to tolerate greater total doses of radiation when given in smaller, daily fractions. This might increase local control and reduce the risk of necrosis and long-term toxicity.

Patients with WHO grade I meningioma usually have long survival. Conventional radiotherapy irradiates larger volumes of normal brain, and patients may be at risk for long-term side effects.

As Solberg et al [37] state, the most logical targets for FSRT are those composed of neoplastic cells surrounded by normal tissue especially sensible to irradiation toxicity. Fractionation in 1.2 to 2 Gy once or twice per day will avoid most of the complications attributable to single-dose or higher fractionation.

The advantage of conventional radiosurgical treatments over FSRT in sparing normal tissue is usually seen for tumors less than 3.5 cm in diameter [3], [23], [36]. Moreover, the most important factor affecting the result of radiosurgical treatments is the coverage of the whole tumor volume. Shin et al demonstrated that meningioma control rates fell drastically when any portion of the lesion did not receive the therapeutic radiosurgical dose [33]. Thus, based on that data, the author stated that when a dose of 14 Gy to the tumor margin is difficult because of a large tumor size or its proximity to the visual pathways, radiosurgery should not be selected as the initial treatment.

In that way, patients with tumors with diameters greater than 3 cm and/or borders lying less than 3 mm from the optic apparatus will be excluded from most radiosurgical protocols (Table 5), and will be referred to neurosurgical units for partial tumor resection [3], [5], [7], [15], [20], [27], [39].

Table 5. Radiosurgical and fractionated radiotherapy series for cavernous sinus meningiomas

	Series year (Ref.)	No. of patients	Median dose, range (Gy)	Median follow-up, range (mo)	Median tumor volume, range (cm3)	Improvement (%)	Worsening (%)	Radiological responsea (%)
	Radiosurgical series
	Shin et al 2001 [33]	40	36 (24-45)	42 (12-123)	4.3 (0.7-25.7)	25	2.5	37.5
	Chen et al 2001 [7]	69	30 (25-32)	–	4.7 (0.2-2.7)	9	1.5	–
	Iwai et al 1999 [17]	24	–	17.1 (6-36)	10.9 (1.2-28.6)	46	4	46
	Liscak et al 1999 [19]	67	–	–	7.8 (0.9-3.4)	35.8b	0	–
	Chang et al 1998 [6]	24	–	45.6 (19-80)	6.83 (0.5-22.5)	42	4	37
	Pendl et al 1998 [27]	43	–	39 (18-62)	15.4 (1.14-82)	57c	0	34.5
	Duma et al 1993 [15]	34	–	26(6-49)	5.2 (0.5-20.4)	24	6	56
	Fractionated radiotherapy series
	Maguire et al 1999 [20]	28	53.1 (30.6-60)	41 (3-145)	–	–	7.1	–
	Alheit et al 1999 [3]	24 (6 CS)	55	13 (3-43)	21.7 (4.4-183)	46.6	8.3	12.5
	Debus et al 2001 [13]	180 (21 CS)	56.8	35 (3 m to 12 y)	52.5 (5.2-370)	44.8	2.1	14d
	Present series 2005	30	52 (50-56)	50 (28.2-74.5)	15.74 (0.90-59.68)	50	6.6e	20

Two patients developed late neurocognitive deficits (Fig. 1).

a Nonhomogeneous criteria in different series. b Includes patient's subjective improvement. c Percentage of improved cranial nerve deficits (not stated). d Only considered if at least 50% of tumor volume reduction. e Worsening of pretreatment neurologic symptoms.

In FSRT, the use of 3D planning procedures with noncoplanar fields achieves very good conformational adaptation to irregularly shaped target volumes. Moreover, modern repeat fixation with noninvasive methods allow accuracy in a fractionated regimen comparable with that of single fraction radiosurgery, with equivalent time consumption for the total treatment in both techniques [2], [4], [7], [12], [30], [37], [40].

5. Conclusions 

It is important to consider that our series is homogeneous in tumor location and it is focused only in meningiomas of the cavernous sinus. Our experience reassesses that in symptomatic patients, neurologic improvement will occur in a significant number of cases without morbidity for a majority of them. However, because of the long natural history of these tumors, larger cohort and longer follow-up are needed to better define the efficacy and late toxicity of FSRT.