|Home | About | Journals | Submit | Contact Us | Français|
Cavernous sinus metastases represent difficult clinical scenarios because of the lack of surgical options. We investigate the use of Gamma Knife stereotactic radiosurgery (GKRS) as a treatment option of these metastases.
To determine the patterns of failure for cavernous sinus metastases and to identify factors that predict for improved outcomes.
This is a retrospective review of 19 patients treated with GKRS for cavernous sinus metastases over a 9-year period between May 2002 and October 2011. The median marginal tumor dose was 18 Gy. Patients were followed with serial imaging. Kaplan Meier analysis was used to estimate local control and overall survival. Fischer exact test was used to determine any predictive factors for local control or survival.
Median follow-up time was 22.4 months. Kaplan Meier estimate of overall survival at 1, 2, and 4 years was 76%, 44%, and 44% survival, respectively. 11 patients experienced intracranial failure. Of these, 7 (64%) were local and 4 (36%) were distant intracranial failures. Local control was 76%, 44%, and 44% at 1, 2 and 4 years, respectively. Six of seven local failures in the series were considered to be marginal failures because they were abutting the 50% isodose volume. Head and neck primary tumors were associated with 86% of local failures (P = 0.017) and was the only factor that predicted for local failure.
GKRS appears to be a feasible and safe modality for treatment of cavernous sinus metastases. Local failures appear to be due to a marginal miss of microscopically occult disease.
Cavernous sinus (CS) metastases occur in approximately 1% of patients with metastatic cancer. While rare, CS metastases represent a unique clinical dilemma because of the neurovascular structures that run through the CS and the hazards associated with surgical resection within this region. Moreover, a significant proportion of patients with CS metastases are derived from primary cancers of the head and neck (H&N) region which have already been previously irradiated[1,3].
Stereotactic radiosurgery (SRS) represents an attractive treatment option in patients with CS metastases because of its greater biologic effectiveness as compared to fractionated radiotherapy and the ability to spare the normal brain from cognitive toxicity[4-8]. SRS affords greater control of falloff dose and thus the ability to spare the surrounding normal structures such as the hippocampus, which in previous studies has been identified as eloquent cortex essential for cognitive function.  Previous studies have described the use of radiosurgical boost for intracranial invasion from nasopharynx and other head and neck cancers for the benefit of reducing dose to normal surrounding critical structures. 
However, there is a paucity of data for patients with CS metastases treated with radiosurgery[9,10]. A particular concern is the potential for marginal failure given the natural history of direct extension from primary head and neck cancers and the sharp dose falloff afforded by radiosurgery.
This series represents a single institution retrospective review assessing the effectiveness and patterns of failure associated with Gamma Knife stereotactic radiosurgery (GKRS) for CS metastases. We examine whether factors such as tumor volume, tumor histology, and GKRS dose affect the treatment outcome.
This study was approved by the Wake Forest University Institutional Review Board. The Wake Forest University Department of Radiation Oncology Gamma Knife Tumor Registry was searched for all patients who underwent GKRS and had a noted CS metastatic lesion on MRI. None of the patients underwent surgical resection or biopsy. Radiologic criteria as described by Korchi et al was used to properly identify cavernous sinus metastases. 
Nineteen patients with CS metastases were identified who were treated with GKRS between May 2002 and January 2012 at Wake Forest University Baptist Medical Center in Winston-Salem, North Carolina. Patient outcomes were then determined using the patients’ electronic medical records and the Social Security death index.
Patient characteristics are summarized in Table 1. Patient factors such as age, race, primary cancer, tumor volume, and radiosurgical prescription dose were determined from the electronic medical record.
After evaluation by a radiation oncologist and neurosurgeon, informed consent for GKS was obtained. Patients were treated using a Leksell Model B, C, or Perfexion Unit (Elekta AB). Prior to radiosurgery, the patients underwent a high-resolution contrast-enhanced stereotactic MRI study of the brain. Treatment planning was performed using the Leksell GammaPlan Treatment Planning System (Elekta AB). A median dose of 18 Gy (range 12-20 Gy) was prescribed to the margin of each metastasis. The dose prescription was determined based on the size and volume of the metastasis following guidelines published by Shaw et al for single-fraction radiosurgical treatment of brain metastases, as long as the optic structures could be constrained to 8 Gy maximum dose to any point.
Kaplan-Meier analysis was performed to determine overall survival of our patient population. The log-rank test was used to determine statistical differences between time to event outcomes. The Fisher Exact or Chi-Square test was used to determine differences in likelihood of neurological death between cohorts.
Median survival time for the entire cohort was 22.4 months. Kaplan Meier estimate of overall survival at 1, 2, and 4 years was 76%, 44%, and 44% survival, respectively. Median survival for H&N, breast, and all other primaries were 22.4, 10.4, and 25.0 months, respectively.
At last follow-up, 15 of 19 patients in the series have died. Cause of death was local failure in 4 patients, other intracranial metastatic disease in 3 patients, extracranial metastatic disease in 6 patients, and primary tumor in 1 patient. Cause of death was unknown in a single patient.
Kaplan-Meier analysis was used to determine local control and disease free survival. Local control at 1, 2, and 4 years was 76%, 44%, and 44%, respectively. Disease free survival at 1, 2, and 4 years was 72%, 46%, and 0%, respectively. Median time to disease recurrence was 7.6 months. Of the 11 total cancer recurrences, 7 were local failures while 4 patients experienced failures at distant metastatic sites. Six of seven local failures were considered to be marginal failures because they were abutting the 50% isodose target volume.
Of the 7 patients that experienced a local failure, two patients were re-treated with GKRS with plans that matched treatment volumes with the previously treated plan. One of these patients continued to fail locally and succumbed to local disease failure. The remaining patient is alive 3 months after radiosurgical salvage.
The Fisher Exact test was used to determine whether primary tumor type was differentially related to failure type. H&N primary malignancies represented 6 of the 7 patients with a local failure and was the only factor that predicted for local treatment failure (P = 0.017). None of the other assessed factors including marginal dose, histology, tumor volume, or age were related to either local failure or overall survival.
A single patient experienced new facial numbness within 6 weeks after radiosurgery to a CS breast cancer metastasis with a dose of 20 Gy prescribed to the 45% isodose line. The numbness persisted thereafter, but was non-bothersome and thus the patient was not treated with steroids.
No patients experienced toxicity that required prolonged steroids, hospitalization, or that contributed to death.
There have been few other series documenting the use of radiosurgery for the treatment of CS metastases. Kano et al reported a series of 37 patients who experienced a 1 year overall survival of 37%. In this series, early treatment of tumors (within 4 months of symptom onset) was associated with an improvement in cranial nerve function. Iwai et al reported a series of 21 patients (nine of which had nasopharyngeal cancer) which were treated with a median dose of 14 Gy using GKRS, comparable to the findings in the current study. Mori et al (2006) reported on their series of patients treated with SRS for metastases in the pituitary and CS. Of 623 patients with brain metastases, 2.15% had pituitary or cavernous sinus metastases. Cavernous sinus metastases were treated with a margin dose of 14.4 to 20 Gy. Cmelak (1997) reported their series of 47 patients with skull base lesions which included 11 patients treated with GKRS for a boost for nasopharynx cancer after fractionated radiotherapy. All others included 37 patients with skull base lesions treated with radiosurgery for recurrent or metastatic lesions. They report a 69% local control rate, and an 8.4% complication rate (however, this is including recurrent lesions treated after external beam radiotherapy). Older series include a report by Vikram (1979) which included 46 patients with metastases to the base of skull treated with external beam radiotherapy, which relieved 78% of symptoms in treated patients. Table 2 summarizes existing institutional series describing outcomes in patients receiving external beam radiotherapy (EBRT) or radiosurgery using either gamma knife radiosurgery or linac-based radiosurgery. [8,9,10,15,16]
The selection of optimal dose and volume delivered to a CS tumor is confounded by location and dosimetric constraints of adjacent structures. Specifically, tumor size, proximity to the optic nerve and prior radiotherapy have been shown to influence the prescription dose. In the aforementioned University of Pittsburgh series, the median dose to the tumor margin was 14 Gy as compared to the 18 Gy margin dose used in the current series. The Pittsburgh series demonstrated improved progression free survival in patients treated with doses higher than 14 Gy. Delivery of this dose may not be possible due to proximity to the optic nerve and brainstem. For this reason, methods such as hypofractionated stereotactic radiotherapy is a potential avenue that requires further study. One clinical dilemma that occurs with local failure within the CS is the lack of a surgical salvage option. CS surgery has a high rate of morbidity, and in a population where life expectancy is limited, salvage surgery is generally not considered a good option.
The increased rate of marginal failure in patients with head and neck primary tumors that extended into the CS has not been previously reported. Of the 6 patients who experienced marginal failure, 2 patients ultimately received further applications of GKRS. However, 4 of the patients experiencing marginal failure ultimately died of local disease failure. One implication from this data is that fractionated re-irradiation may be a reasonable option given its ability to treat a clinical target volume (CTV) expansion to account for microscopic extension of disease. By the nature of radiosurgery, a CTV expansion is generally not used, and dose falloff is much sharper than with conventionally fractionated radiotherapy, predisposing patients with microscopic disease to a marginal miss. A previous series of re-irradiation of nasopharyngeal cancers revealed that patients experienced durable rates of local control with fractionated re-irradiation, with acceptable toxicity outcomes.
Given the concern for marginal failure for patients with CS metastases from H&N primaries, it is likely that the future application of radiosurgery will be in the setting of local failure after re-irradiation or as a boost to dose escalate in the re-treatment setting. The major limitation of re-irradiation at the CS is the tolerance of the normal tissues such as the brain stem and the optics. GKRS, with a dose fall-off profile much sharper than that of external beam radiotherapy, allows for dose distributions that may be able to spare normal tissues better than external beam irradiation alone. The paradigm of conventionally fractionated re-irradiation followed by a focused hypofractionated boost was recently demonstrated by Koutcher el al in a series of patients with recurrent nasopharyngeal cancer treated with external beam radiotherapy with a brachytherapy boost. Toxicity in this series was worsened when external beam radiotherapy was used alone as compared to the use of a more focal hypofractionated brachytherapy boost. Further series will be necessary to evaluate the efficacy of this salvage option as well as the role of GKRS in this manner.
The limitations of the current study include its retrospective nature as well as the small patient numbers. However, given the relative rarity of CS metastases, it is unlikely that a prospective trial will be feasible. Further studies will be necessary in order to validate the findings of increased marginal failure in patients with head and neck primaries and investigate whether a clinical target volume margin should be added in this region.
GKRS appears to be a feasible and safe modality for treatment of CS metastases. Local failures appear to be due to a marginal miss of microscopically occult disease.