Stereotactic Radiosurgery for Brain Metastases
ONCOLOGY 13(10):1397-1409,1999

Complications of Radiosurgery

Acute Complications

Very few significant acute complications are observed within the first week following radiosurgery. Loeffler and Alexander reported nausea in 22 (11.2%) of 196 patients during the first 24 hours (all of these patients received > 2.75 Gy to the area postrema). They also noted seizures in 12 patients (6.1%) within 24 hours and transient motor weakness in 4 patients (2%) within 36 hours, all 4 of whom had motor cortex lesions. An update of these data by Alexander et al suggests that in order to reduce or eliminate radiosurgery-induced nausea/vomiting, all patients who are being irradiated with more than 375 cGy to the area postrema should receive antiemetic therapy prior to treatment. Ten of the patients experiencing post-radiosurgery seizures had a history of seizure disorder and, in retrospect, had subtherapeutic levels of anticonvulsants. Since 1990, the Boston group recommends that all patients with cortical lesions receive antiseizure prophylaxis before the radiosurgery procedure, regardless of their seizure history. Breneman et al reported an 8.3% (7/84) acute complication rate. Three patients experienced transient worsening of their neurologic symptoms within the first 2 weeks after radiosurgery that resolved after a short course of steroids. Three patients with lesions in the motor cortex experienced grand mal seizures within the first 72 hours after radiosurgery. Parenthetically, all of these patients had subtherapeutic anticonvulsant levels. Joseph et al reported a low incidence of severe headaches or nausea, and a 2.3% (3/120) incidence of seizures within 12 hours of completing radiosurgery. Each of these patients was also found to have a subtherapeutic anticonvulsant level at the time of treatment. Shiau et al observed complications in 100 evaluable patients. They found that six patients (6%) experienced acute deterioration of preexisting neurologic deficits, such as aphasia or hemiparesis, within 2 weeks of radiosurgery; in five of the six, this deterioration responded to steroids. Fukuoka et al described several episodes of acute toxicity in 130 evaluable patients treated with Gamma Knife stereotactic radiosurgery. In one case, a 9-cm³ metastatic lesion in the cerebellar hemisphere was treated with 30 Gy to the tumor margin. The lesion subsequently swelled and the preexisting perifocal edema increased, requiring surgical resection the day after Gamma Knife therapy. In another case, a patient with a 1.4-cm³ lesion developed hemiplegia 3 days after Gamma Knife therapy (30 Gy to the tumor margin) due to intratumoral bleeding. However, this patient recovered with minimal hemiparesis. In a third case, a massive perifocal hematoma was surgically removed 14 days after Gamma Knife therapy for a 19-cm³ metastatic lesion from thyroid cancer (30 Gy to the tumor margin). It is possible that the high doses to the tumor margin used in these cases may have contributed to these acute toxic episodes. Therefore, it is prudent to consider premedicating patients with lesions near the posterior fossa with antiemetics and to routinely check serum anticonvulsant levels in those with a seizure history. In addition, we do not use doses in the 30-Gy range, even for small lesions. For a list of common anti-seizure meds go here

Subacute Complications

Complications occurring within the first 6 months are considered subacute. The data from the major trials suggest the following as possible subacute complications:

  1. In the Joint Centers for Radiation Therapy (JCRT) series, alopecia developed in 11 patients (5.6%), all of whom received at least 4.4 Gy of radiation to the scalp.
  2. Steroid reinstitution and/or continuation was necessary in 21% of survivors at 6 months in the JCRT series, probably reflecting vascular change rather than tumor recurrence.In the multicenter report of Flickinger et al, steroids had to be reinitiated at < 6 months in < 3% of patients.
  3. The UCSF series reported six cases of neurologic deterioration between 2 weeks and 3 months after radiosurgery, including 4 cases of increased hemiparesis and/or confusion treated with steroids, and 2 cases of necrosis documented by magnetic resonance (MR) spectroscopy or autopsy. In the Japanese experience, reported by Fukuoka et al, perifocal edema developed in 19 of 130 patients (45 single, 85 multiple lesions) treated with Gamma Knife stereotactic radiosurgery. Among the 19 cases with edema, transient hemiparesis developed in 7 cases, and disappeared within 1 week to 1 month. Permanent deficits developed in four patients (hemiparesis in three and mental status change in one).

Chronic Complications

The major chronic complications of radiosurgery are radiation necrosis, cranial nerve palsies, and chronic steroid dependence.

Radiation Necrosis—In the JCRT experience, symptomatic radiation necrosis developed in 17 (8%) of 217 patients from 2 to 22 months following therapy; all of these patients eventually required resection. Ten of these patients had also received methotrexate, which may have contributed to the necrosis. Breneman et al reported worsening of neurologic symptoms associated with increasing mass effect in 2 (2.3%) of 84 patients after radiosurgery; these patients required surgical decompression. Both of these patients were found to have radiation necrosis.Delayed local necrosis within the treated volume was the major complication in 16.6% (20/120) of patients in the Stanford study.[29] The risk of necrosis was a function of tumor volume and prior or concurrent whole-brain irradiation. Symptomatic and pathologically confirmed necrosis with no viable tumor developed in 1 of 116 patients 11 months after radiosurgery in the report by Flickinger et al, yielding a 2-year actuarial risk of developing necrosis of 4%. However, reoperation was required in 10 patients (8.6%) because of hemorrhage, necrosis, or recurrence. In the UCSF report, four late complications occurred > 3 months after radiosurgery. This included one case of increased hemiparesis, one case of increased confusion, and two cases of brain necrosis documented by positron emission tomography (PET). In the study by Kim et al, two patients demonstrated delayed intratumoral hemorrhage and underwent craniotomy and resection. One additional patient underwent surgical resection for necrosis. In a report of 40 patients with 41 metastatic lesions, Alleyne et al noted the development of seizure activity and MRI evidence of temporal lobe necrosis in 1 patient 12 months after radiosurgery.

Cranial Nerve Palsies—Only about 1% of patients (2/196) in the JCRT series developed permanent cranial nerve palsies.This phenomenon was observed at 7 and 8 months, respectively, following treatment. One palsy involved the fifth nerve, which received 16.5 Gy, and the other involved the eighth nerve, which received 15 Gy.

Chronic Steroid Dependence—One of the potential complications ascribed to radiosurgery is persistent edema requiring chronic steroid usage, with all of its attendant complications. In the University of Wisconsin experience, the need for steroids beyond 6 months was documented in 4 (7%) of 54 of patients. This figure is quite comparable to the 8% rate of steroid dependence at 12 months noted in the JCRT experience.[32,44] Flickinger and colleagues reported a 2-year actuarial rate of developing delayed symptomatic edema of 10.8%.

In our experience, in several cases, radiosurgery results in dramatic resolution of edema, rather than persistence of this phenomenon. Further confirmatory evidence is provided by the study of Jokura et al, who found “dramatic improvement of perifocal edema with regression of tumors.”


Based on a review of the literature, the following general conclusions can be drawn about the treatment of brain metastases:

1. In general, the natural history of brain metastases is very poor. If left untreated, patients with these metastatic brain lesions have a median survival of only 1 month.

2. This natural history can be altered somewhat by the use of steroids and external-beam radiotherapy. However, neither the fractionation scheme nor the total dose seem to have much of an effect in the older RTOG studies.

3. Predictors for favorable outcome include age < 60 years, KPS > 70, a controlled primary tumor, the brain as the only site of metastasis, and fewer than three metastatic lesions.

4. Radiosurgery has been employed in over 1,750 patients, with local control rates of nearly 70%.

5. The morbidity of radiosurgery is low. The most serious risks are radiation necrosis and cranial nerve damage, both of which occur in relatively small percentages of patients (4% and 1%, respectively). Treatment volume and prior radiation therapy may be important predictors for late damage; a very high dose appears to be related to the development of acute toxic episodes.

6. Retrospective comparisons suggest that radiosurgery produces results comparable to those of surgical resection and is probably more cost-effective.

7. The role and sequencing of whole-brain radiation are not clearly defined. Shirato et al have reported a very high failure rate if whole-brain radiation is not used, but this does not appear to compromise survival.Flickinger et al have found an improvement in local control but not in survival when whole-brain radiation is used in conjunction with radiosurgery.

8. Clinical trials of radiosurgery for brain metastases are underway and need to be actively supported. These include an RTOG phase III study testing the value of a radiosurgery boost in addition to whole-brain radiation therapy in patients with one to three metastases; an ECOG phase II study attempting to define intracranial relapse patterns following radiosurgery alone for radioresistant tumors; an M. D. Anderson study comparing resection to radiosurgery; and a recently opened RTOG trial evaluating the radiosensitizer RSR-13 in conjunction with radiosurgery.