Surgery vs Radiotherapy for Solitary Brain Metastases

Abstract & Commentary

Synopsis: There have been no randomized trials comparing surgery and radiotherapy for solitary brain metastases, nor are there well-defined guidelines for the management of these patients. This study from the Mayo Clinic performed a retrospective analysis of the institutional experience comparing outcomes in patients treated with both approaches and concluded that the type of therapy selected was not prognostic in terms of survival but was a statistically significant factor regarding local control.

Source: O’Neill BP, et al. Int J Radiat Oncol Biol Phys. 2003;55:1169-1176.

Approximately half of the 200,000 patients diagnosed with brain metastases annually have solitary lesions. The number of patients with solitary brain metastases (SBM) is expected to rise as systemic therapies improve, given that the brain is considered to be a sanctuary site. While surgery and stereotactic radiosurgery (SRS) are believed to offer local control benefits over whole-brain radiotherapy alone, there are few data comparing surgery and SRS and their respective impact on local control. Though the surgical option is not available to half of SBM patients because of comorbid conditions, the other half may be treated with either modality.

In order to compare local control and survival results for the 2 types of therapy, O’Neill and associates at the Mayo Clinic retrospectively reviewed their experience with 97 SBM patients treated from 1991-1999. All patients underwent an MRI with contrast to confirm that their metastases were solitary, and all were felt to have been eligible for either therapy in retrospect (ie, tumors were < 35 mm, not deep-seated or located in the brainstem, and patients were free of ventricular obstruction). There were 74 patients treated surgically (76%) and 23 treated with SRS (24%). Among the surgically treated patients, 82% received whole-brain radiotherapy, and among the SRS patients, 96% received it (P = NS). There were no significant differences in the baseline demographic characteristics in the 2 treatment groups, although the SRS patients had "a less favorable mix of prognostic factors." Median age for the surgery patients was 66 years (range, 51-71 years) and 63 years for the SRS patients (range, 55-70 years). Metastases were symptomatic in 89% of the surgery patients and 74% of the SRS patients (P = NS). All SRS patients were treated with a Gamma Knife radiosurgery system.

Median follow-up was 20 months for surviving patients (range, 0-106 months). According to O’Neill et al, survival was measured from the date of the procedure. A propensity score for therapy assignment was derived for each patient as a means of accounting for confounding and selection biases, covering 25 covariates. One-year actuarial survival was almost identical between the 2 treatment groups, though surgery patients lived longer in year 2 (P = .15). There were no significant differences in complication rates, either short-term or long-term, between treatment groups. There was a trend for symptomatic patients to do worse than asymptomatic patients (P = .07). SRS patients with left-sided lesions did significantly worse than their surgery counterparts (P = .001). Fifty-nine percent of surgery patients died of systemic disease compared to 48% of SRS patients (P = NS). Thirty percent of surgery patients (19/74) developed recurrences in the brain, as did 29% (6/23) of SRS patients (P = NS). Significantly, while there were no SRS patients who experienced local failures at the site of their SBM, 11/19 (58%) of the surgery patients with relapsed disease in the brain had the disease recur at the site of original brain lesion (P = .02). Despite this, more SRS patients died of CNS tumor (29%) than did surgery patients (11%, P = .36). Multivariate analysis indicated that 2 factors were independently associated with survival: ECOG performance status and lack of systemic disease. Most notably, type of therapy was not found to be a prognostic factor for survival, even after adjustments for propensity score (P = .60).

O’Neill et al concluded that neither SRS nor surgery was superior for patients with small-to-moderate sized SBMs among their group of highly selected patients. They are in favor of a randomized trial in order to minimize confounding variables and selection bias and alerted the reader that such a trial is being planned under the auspices of the American College of Surgeons. O’Neill et al suggested that a quality-of-life assessment tool be included in this type of trial.

Comment by Edward J. Kaplan, MD

The Mayo Clinic paper was especially interesting because the results of the study offer a unique perspective for clinicians who see SBM patients, particularly those patients who are candidates for either therapy. Although the causes of death did not differ significantly between the 2 patient groups, I was surprised that local control of SBMs was markedly better in the SRS patients. In the long run, this did not seem to matter because the overall number of brain recurrences was about the same in both groups. It is unclear why the SRS patients must have had a higher number of recurrences elsewhere in the brain compared with the surgery patients. Perhaps it was related to the use of whole-brain RT, or the doses used. The whole-brain dose and fractionation schedules were not mentioned, and we are not told whether the SRS patients received their radiosurgery before or after their whole-brain RT. Survival was measured "from the date of the procedure," which likely meant either resection or SRS. Similarly, SRS dose and prescription information was lacking.

Simonova and colleagues from Prague reported their results with Gamma Knife radiosurgery for 237 patients with SBMs. They reported a local control rate of 95% with doses > 20 Gy.1 Jyothirmayi and colleagues from the Royal Marsden published their experience with 96 SBM patients treated with 20 Gy in 2 fractions, and they found that outcomes were comparable to surgical excision.2 The findings in both of these studies seem to be consistent with those of O’Neill et al.

In my opinion, given 2 options for the same problem, I would opt for the least invasive technique if outcomes are similar. As the authors pointed out, the number of SBM patients is expected to increase as systemic therapy improves, and thus the results of a randomized trial comparing SRS and surgery should offer us important information in terms of clinical recommendations in the future. It is unclear whether the use of temozolomide may enhance the results in SBM patients treated by noninvasive means.

Dr. Kaplan is Acting Chairman, Department of Radiation Oncology, Cleveland Clinic Florida, Ft. Lauderdale, FL; Medical Director, Boca Raton Radiation Therapy Regional Center, Deerfield Beach, FL.


1. Simonova G, et al. Radiother Oncol. 2000;57:207-213.

2. Jyothirmayi R, et al. Clin Oncol (R Coll Radiol). 2001;13:228-234.