Postoperative stereotactic radiosurgery for limited brain metastases: are we ready for prime time?

KEYWORDS:

• Brain metastases
• postoperative
• cavity
• stereotactic radiosurgery
• whole brain radiotherapy

Whole brain radiotherapy (WBRT) has long been the standard of care for patients with brain metastases. Two randomized phase III trials comparing surgical resection + WBRT and WBRT for single brain metastasis showed that the addition of surgical resection resulted in an improvement of overall survival [1,2]. In an attempt to omit WBRT totally after surgical resection, a randomized phase III trial by Patchell et al. showed that omission of WBRT resulted in a 46% local recurrence and a 70% intracranial failure rate [3]. As a result, WBRT is often recommended after surgical resection of limited brain metastases. However, caveats exist when interpreting the results of that trial. Although MRI was used to evaluate the extent of brain metastasis, thin slice volumetric MRIs with double gadolinium or gadobenate dimeglumine were not used and some punctate lesions might be missed. Furthermore, a local recurrence rate of 46% was much higher than that for stereotactic radiosurgery (SRS) alone [3]. WBRT reduced the local recurrence rate to 14%, indicating that for many patients surgical resection alone failed to address microscopic disease and radiation was needed to ‘mop up’ the surgical bed.

For patients with one to three or four intact brain metastases, four randomized phase III comparing SRS and SRS plus WBRT showed that with the omission of WBRT, despite the higher risk of intracranial failure, overall survival was not decreased [4–7]. In the M.D. Anderson Cancer Center (MDACC) and North Central Cancer Treatment Group trials, neurocognitive function was found to be significantly worse for patients receiving WBRT, indicative of the deleterious effect of this approach [7]. With the advent of more effective systemic therapies, patients live much longer than before and, therefore, are more at risk to suffer from the long-term effects of WBRT such as neurocognitive decline. Individual patient meta-analysis by Sahgal et al. have showed that for patients ≤50 years of age, SRS alone favored survival and the initial omission of WBRT did not impact distant brain relapse rates [8]. With this level I evidence, SRS alone has been established as the modern standard of care in patients in one to three or one to four metastases. In fact, the American Society for Radiation Oncology Choosing Wisely Campaign has issued a statement recommending against the routine use of adjuvant WBRT in addition to SRS for patients with limited brain metastases [9].

However, it was less clear as to what the best adjuvant treatment would be after surgical resection of a brain metastasis in a patient with limited brain metastases. Recently, Mahajan et al. and Brown et al. reported the results of their randomized phase III trials comparing SRS and observation and comparing WBRT plus SRS only to unresected brain metastases and SRS to surgical cavity and unresected brain metastases, respectively, in abstract form [10,11]. In the phase III trial presented by Mahajan et al., a total of 132 patients (128 analyzed) with one to three brain metastases wished to avoid or delay WBRT following complete surgical resection of at least one brain metastasis. Patients were randomly assigned to undergo either SRS to the surgical cavity/cavities or observation alone [10]. The radiation doses delivered 12–16 Gy according to the cavity volume at the time of SRS. The local control rates at 6 months for SRS and observation groups were 83% and 57%, respectively. The corresponding 1-year local control rates were 72% and 45%, respectively. There was no difference in rates of regional brain recurrence (58% vs. 67%), overall survival (17 months vs. 17 months), or time to WBRT between the two arms (16.1 months vs. 15.2 months). Local control was determined by tumor size with tumors larger than 3 cm associated with worse local control. The patterns of local failure relative to the CTV were not reported. For the randomized phase III trial reported by Brown et al. (N107C), 194 patients with one to four brain metastases were randomized to receive either SRS or WBRT after resection of one of the brain metastases [11]. All unresected brain metastases were treated with SRS in both arms. For target delineation of the cavity, a 2-mm margin was expanded around the CTV which was defined as the surgical cavity. The surgical access tracts for deep lesions were not specifically covered for deep lesions. The prescribed dose for cavity SRS was according to cavity volume and ranged from 12 Gy to 20 Gy. With a median follow-up of 15.6 months, there was no difference in overall survival rates between the two arms (11.5 months for SRS and 11.8 months for WBRT). Of note, patients who received SRS had significantly longer survival without cognitive decline compared to those patients who received WBRT. The rate of cognitive decline at 6 months was 85.7% after WBRT compared to 53.8% after SRS. WBRT provided better overall intracranial tumor control rates with 90% and 78.6% at 6 and 12 months compared to 74% and 54.7% for SRS alone. It also resulted in better surgical bed relapse-free survival rates as compared to SRS alone, although there was no clinically meaningful difference between treatment arms (7.7 months vs. 7.5 months). Physical well-being and quality of life were better among patients who received SRS.

Based on the results of these two trials, a few observations were made. First, SRS improves local control compared to observation at the surgical resection bed. Second, the omission of WBRT does not jeopardize overall survival despite the high risk for intracranial failure and local failure in the cavity. Third, the omission of WBRT resulted in less cognitive decline and better quality of life. These last two points must be considered with the understanding that the benefits of upfront omission of WBRT does require careful and regular follow-up with MRI of the brain in order to diagnose and salvage recurrence. At the time of writing this editorial, these two trials have not yet been published in full manuscript form. In the MDACC trial, the local control rate in the surgical cavity arm with observation appeared to be lower than expected (45% at one year) compared to that reported in the Patchell trial. The potential for differences in preoperative tumor size (not reported in the Patchell trial) and how local control was reported (actuarial vs. crude rates) may account for this observation. The fact that a preoperative tumor size of > 3cm was associated with worse local control in the MDACC trial implied the presence of a radiation dose response as the prescribed dose was determined by the size of the cavity which was expected to correlate with preoperative tumor size. The 1-year local control after cavity SRS was also only 72%, which was lower than the 86% local control rate in the WBRT arm in the Patchell trial. The randomized comparison of cavity SRS and WBRT in N107C also showed that WBRT not only improved overall intracranial control but also the surgical bed local control. Plausible explanations include the difference in target coverage with SRS just covering the surgical cavity with 2-mm expansion without including the surgical access tracts and meningeal resections, where recurrence can occur in the form of local leptomeningeal disease [12] and the lower biologically effective dose delivered to larger cavities with low-dose SRS compared to that delivered with WBRT. However, information on the patterns of failure was not available for both trials and therefore, any attempted explanation is speculation at this point.

So what do we do with our patients with limited metastases who undergo surgical resection for one of the metastases? The ultimate decision should be made based on the risk-benefit ratio. The risk of intracranial recurrence has to be balanced against the risk of significant decrease in memory/cognitive function and quality of life. Multiple trials have shown that WBRT causes cognitive and quality of life decline and does not improve overall survival despite improvements in intracranial control. The approach of using focal SRS as an adjuvant treatment after brain metastasis resection can be regarded as one of the standard therapies based on level I evidence. However, some fine-tuning of the techniques in terms of target delineation and dose fractionation (SRS vs. hypofractionated stereotactic radiotherapy) may help further optimize outcomes. For instance, for larger cavities, instead of lowering the dose to 12 Gy in one fraction, it may make sense to offer hypofractionated stereotactic radiotherapy using 24–27 Gy in three fractions or 30–35 Gy in five fractions. Fractionating the SRS allows for a higher biologically effective dose as compared to single fraction SRS without increasing the risk of radiation necrosis. Currently, there is no level I evidence to confirm that in a postoperative cavity, increasing the dose to a large cavity is better or worse than hypofractionation. The original Radiation Therapy Oncology Group guidelines were for unresected lesions and it is unclear whether the dose efficacy and toxicity are applicable to postoperative cavities. Hopefully, the final reports of these two trials in full manuscript form will be able to provide additional details and future reports will provide detailed analysis of patterns of failure as this will help refine our treatment techniques. In the modern era, systemic targeted agents with brain penetration are being used extensively and when future trials are developed, this should be taken into account.

Declaration of interest

A Sahgal has conducted past educational seminars with Medtronic, Elekta AB, Accuray Inc., and Varian medical systems, has a research grant with Elekta AB, received travel accommodations/expenses by Medtronic, Elekta and Varian, and is a member of the Elekta MR Linac Research Consortium. ST Chao has received honorarium from Abbie, Zeiss, and Varian Medical Systems. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

This paper was not funded.