Neuro-Oncology Practice Clinical Debate: stereotactic radiosurgery or fractionated stereotactic radiotherapy following surgical resection for brain metastasis

Abstract

The treatment of resected brain metastasis has shifted away from the historical use of whole-brain radiotherapy (WBRT) toward adjuvant radiosurgery (stereotactic radiosurgery [SRS]) based on a recent prospective clinical trial demonstrating less cognitive decline with the use of SRS alone and equivalent survival as compared with WBRT. Whereas all level 1 evidence to date concerns single-fraction SRS for postoperative brain metastasis, there is emerging evidence that fractionated stereotactic radiotherapy (FSRT) may improve local control at the resected tumor bed. The lack of direct comparative data for SRS vs FSRT results in a diversity in clinical practice. In this article, Greenspoon and Roberge defend the use of SRS as the standard of care for resected brain metastasis, whereas Palmer and Brown argue for FSRT.

Clinical Scenario

A 52-year-old woman presents with a first-time generalized seizure 4 years after initial treatment of a locally invasive estrogen receptor–/progesterone receptor–negative, HER2–positive breast cancer. She had undergone routine posttreatment clinical and radiographic follow-up without evidence of recurrence, though intracranial imaging had not been obtained. Brain MR imaging demonstrates a single right frontal ring-enhancing mass measuring 3.5 cm in maximum dimension abutting the dural meninges, with surrounding vasogenic edema and 5 mm of midline shift. Systemic restaging imaging demonstrates no additional sites of disease. She undergoes a gross-total resection of the mass without complication, and pathology confirms the diagnosis of metastatic HER2–positive breast cancer. She is referred to radiation oncology after surgical resection to discuss treatment to reduce the risk of recurrence. What treatment plan would you advise for her?

Single-Fraction Radiosurgery View (Drs Greenspoon and Roberge)

The management of large and symptomatic brain metastases has evolved over time. When possible, large metastases causing mass effect or needing tissue diagnosis should undergo surgical resection. Given the nature of neurosurgery for brain metastasis, the risk of local recurrence with no additional therapy following surgery is unacceptably high.^1,2 The first treatment evaluated to improve disease control following surgery for a brain metastasis was whole-brain radiation therapy (WBRT). In randomized trials this proved to be effective at decreasing both the risk of local recurrence in the surgical bed and distant recurrence in the brain.

With the passage of time, the predominant treatment of limited in situ brain metastases evolved from WBRT to WBRT with the addition of stereotactic radiosurgery (SRS) to SRS alone.^3–5 Although SRS alone was shown to lead to a larger risk of distant recurrence within the brain, these recurrences were salvageable with further treatment leading to no difference in overall survival with the omission of WBRT.^3,5,6 In all these randomized trials for in situ brain metastases, radiosurgery was delivered using a single-fraction dose scheme in which the dose is decreased as the treatment volume increases. Doses that would seem modest for other sites of metastases (18-25 Gy) have led to high rates of local control within the brain.^1,3–5,7

Though high-level evidence supports single-fraction SRS as the standard for in situ brain metastases less than or equal to 3 cm in size, until recently there remained uncertainty regarding the value of SRS in the setting of a resected brain metastasis. Recently the results of the first 2 RCTs in this field were published.^6,8 In a cooperative group setting, single-fraction SRS was compared with WBRT for the management of a resected brain metastasis.⁶ Radiosurgery dose was prescribed based on the target volume and ranged from 12 Gy to 20 Gy depending on the size of the surgical cavity. A 2 mm margin around the surgical cavity was added to define the clinical target volume (CTV) for treatment. The primary endpoints of the study were survival and cognitive deterioration-free survival. This study was published in 2017, showing no difference in survival between the 2 arms (WBRT vs SRS) and improved cognitive deterioration-free survival in the SRS arm. Following this publication the adoption of single-fraction SRS increased and was recognized as a standard treatment option as documented in guidelines published since 2017.⁹

Single-fraction SRS as compared to WBRT leads to improved patient-reported quality-of-life outcomes, including physical well-being, functional well-being, and fatigue. Certainly, there were limitations of the single-fraction SRS identified in the study. Local control of the surgical bed for the SRS arm was high at 6 months (80%); however, it was inferior to WBRT at 12 months (61% vs 81%). Although it is possible that fractionation would improve local control, given the very high local control in randomized trials evaluating in situ metastases and the effectiveness of adjuvant WBRT delivered at modest doses, it is plausible to conclude that the local control rates seen with the adjuvant SRS arm are due to microscopic disease outside the CTV and not to dose fractionation.⁷ Further supporting evidence that a limitation of adjuvant SRS is target volume is the pattern of failure that practitioners see with adjuvant SRS—nodular leptomeningeal recurrence well beyond the surgical cavity. This relatively new pattern of recurrence represents spread through the cerebrospinal fluid as a result of surgical intervention. These disseminated cells are theoretically outside the target volume, allowing them to seed throughout the brain.¹⁰

Following resection of a brain metastasis, there is currently high-level evidence supporting the use of either WBRT or single-fraction SRS. The comparative advantages and disadvantages are well understood and supported by the prospective literature.⁸ Certainly there are other potential treatments to improve on the limitations of these 2 options, including preoperative SRS. Among oncologists who support the use of postoperative focal treatments, results seen in prospective trials could theoretically be improved by measures such as delivering SRS to a more generous CTV, including at-risk dura and more brain parenchyma, delivering more conventional partial-brain radiation, increasing the dose of SRS, or using fractionated SRS. In addition, novel techniques for WBRT could minimize the cognitive impact of WBRT while providing surgical bed and distant brain control. Recently NRG CC001 showed substantial improvement in cognitive outcome in patients treated with hippocampal-avoidant WBRT and memantine as compared to conventional WBRT with memantine.¹¹ In reviewing retrospective results of altered treatment regimens, one should be reminded that across multiple retrospective series the reported 1-year surgical bed control of adjuvant single-fraction SRS was beyond 80%.^12–14 Proponents of more cumbersome and expensive treatment regimens should bear the burden of producing prospective comparative evidence.

Before adopting fractionated stereotactic radiotherapy (FSRS), we should consider the most reasonable explanation for the local control rates observed with adjuvant SRS and proceed in a rational, coordinated manner.

Fractionated Stereotactic Radiotherapy View (Drs Palmer and Brown)

The management of resected brain metastasis has evolved over the years, and local tumor bed control is crucial to successful management.⁸ Several modern brain metastasis trials have demonstrated that escalating therapy with the incorporation of WBRT to obtain higher local and distant brain tumor control does not improve overall survival but does come at a cost of cognitive deterioration.^4,6 Thus the question at hand is how to obtain the best local control of the resected tumor bed without sacrificing cognitive function.¹⁵

In the 2 largest randomized prospective trials, using SRS alone for resected brain metastasis demonstrated poor local control.^6,8 In the MD Anderson study, doses of 12-14 Gy were used for large brain metastasis (>10.1 cc). The 12-month freedom from local recurrence following SRS in the surgical bed for tumors measuring 2.5-3.5 cm was 33% and for tumors larger than 3.5 cm was 72%.⁸ In the Alliance trial, larger tumors (>14.4 cc) received 12-15 Gy. The 1-year surgical bed control rate was 60.5%.⁶ The relatively inferior local control rates found in these clinical trials have many potential causes, including tumor volume delineation, radiation dose, and surgical type. In the Alliance trial, 36% to 44% of patients had piecemeal resections that may have led to localized tumor bed seeding and in some cases seeding along the meningeal lining.⁶ In the MD Anderson trial, 26% of patients had large tumors greater than 3.5 cm, whereas in the Alliance trial 40% had cavities larger than 3 cm.⁸ These large cavities due to historical dose selection for single-fraction radiosurgery require dose deescalation for safety.¹⁶ This is a very likely cause of poor tumor control. In contrast, the local tumor bed control in the MD Anderson trial was 100% at 1 year with a dose of 16 Gy in a single fraction for smaller cavities. This trial very clearly demonstrated that larger cavities have a higher risk of local recurrence than smaller cavities.

FSRT has demonstrated excellent local control in multiple retrospective studies. The dose and fractionation schedules vary but typical regimens include 27 Gy in 3 fractions and 30 Gy in 5 fractions. These regimens have higher biologic equivalent doses (BEDs) than the single-fraction SRS doses used for large brain metastasis cavities in trials and in practice. This is possible thanks to the improved safety profile of fractionation on the normal brain with relatively low rates of radiation necrosis, while not sacrificing dose to the tumor. The BED₁₀ for single-fraction radiosurgery doses (BED₁₀ in parentheses) are 12 Gy (26.4), 14 Gy (33.6), 15 Gy (37.5), 16 Gy (41.6), 18 Gy (50.4), and 20 Gy (60), whereas the BED₁₀ for 27 Gy in 3 fractions is 51.3 and 30 Gy in 5 fractions is 48. The Alliance trial demonstrated that the conventional palliative WBRT dose significantly improved 12-month tumor bed control by 20 points (61% to 81% with WBRT), which may be due to more generous treatment of the tumor bed (fewer marginal failures or less nodular leptomeningeal disease) or due to the fractionated radiation and higher BED (39) with WBRT for tumors larger than 14.4 cc, which would receive 12-15 Gy (26.4-37.5) in a single fraction as per the Alliance trial.

To summarize, optimal management of resected brain metastasis includes local control with radiosurgery because of the high rates of local failure with observation.⁸ The use of postoperative WBRT has improved tumor bed control but comes at the cost of worse cognitive function. Single-fraction radiosurgery has been used as the standard of care but requires dose deescalation for larger tumor resection cavities. This dose deescalation for larger tumors seems illogical but is required for safety. Conversely, the use of FSRT allows for preservation of dose intensity needed for tumor control while maintaining safety.

Single-Fraction Radiosurgery Reply (Drs Greenspoon and Roberge)

Drs Palmer and Brown certainly make an interesting case for FSRT; however, the hypothesis that the SRS dose is the limiting factor to the effectiveness of SRS, at this point, does not have enough data to support it. Is higher dose responsible for the improved local control seen with smaller metastases? The 2-year actuarial local control of a metastasis 2.5 cm or less was 77% with a dose of 0 Gy. If measured by the improvement in local control, the lower doses of radiosurgery used in tumors 3.5 cm or larger were actually more effective than the higher doses—reducing the 1-year local recurrence rate from 78% to 28%.

Although this is not the most likely culprit in local recurrences, if dose were the issue, it may be safe to simply escalate the dose of SRS for large cavities based on the 2 postoperative RCTs discussed above.^6,8 Grade 2 or higher radiation necrosis was reported in only 4% of patients in the study by Brown et al⁶ and 0% that by Mahajan and colleagues.⁸ With these low rates of SRS-related toxicity, one could systematically evaluate whether dose was the major cause of poorer than expected local control in the SRS arm of N107C/CEC.3 by escalating the single-fraction SRS dose to larger cavities. As well FSRS is unlikely to reduce the need for WBRT (the use of WBRT in the trial by Mahajan et al⁸ was similar in both arms), and the safety of WBRT post-FSRS may be compromised. On the other hand, the safety of WBRT combined with SRS is well documented.

As Drs Palmer and Brown point out, it is likely that inadequate volume contributed to local failures. It stands to reason that if there are tumor cells remaining in the adjacent dural meninges that were not identified as tissues at risk, these cells could contribute to local progression. Larger tumors would also be more likely to have larger areas of affected meninges as part of the craniotomy, therefore potentially undertreating tissues at risk. Adopting various alternate adjuvant regimens will impede important research. Offering FSRS outside clinical trials will significantly impede accrual to any upcoming trial comparing SRS with FSRS.¹⁷

At this point it is unclear whether the poorer than expected local control of the surgical cavity is due to dose or volume treated, with reasonable explanations for either or both contributing. Rational, prospective trials should be used to inform patient care, and until these data exist, SRS based on N107C/CEC.3 should remain the standard form of adjuvant radiosurgery.

Fractionated Stereotactic Radiotherapy Reply (Drs Palmer and Brown)

Though we agree that SRS has been used as the standard for postoperative resection bed treatment, we show below that in multiple large studies the local tumor bed control is superior with FSRT vs SRS (Fig. 1).^{6,8,12,17–28} In addition, a meta-analysis analyzing local control for single-fraction vs fractionated radiosurgery demonstrated that postoperative tumors larger than 3 cm had a 37% relative local control benefit with the use of FSRT with similar rates of radiation necrosis.¹⁸ The use of FSRT allows for tumoricidal doses while maintaining safety to the normal brain. It also allows for using a larger margin on the dura and normal brain that could potentially decrease recurrence rates in our current patient. The current CTVs used were defined as the surgical cavity alone with a 1 or 2 mm margin. This does not typically include the adjacent dural meninges but for our patient a more generous dural margin is recommended to decrease the risk of nodular or marginal recurrence.¹⁹

Fig. 1.

Comparison of Local Tumor Bed Control for Stereotactic Radiosurgery (SRS) vs Fractionated Stereotactic Radiotherapy (FSRT) at 12 Months; Average of 79% for SRS and 91% for FSRT

Discussion

The controversy regarding SRS vs FSRT for treatment of resected brain metastases fits a common pattern in neuro-oncology, the choice between a therapy of established efficacy and one with potential advantages and promising early data but without the same track record of rigorous evaluation in clinical trials. Both sides in this debate agree on the efficacy of SRS for treatment of small metastases and resection cavities, and also agree that the efficacy of SRS for treatment of large resection cavities is less than for small ones; the difference comes in the response to this issue. Are we to modify the dose or treatment volume of SRS, building on established practice and maintaining the convenience of single-fraction therapy, or take advantage of the larger safe BED delivery made possible by fractionation?

Ultimately, the question of the optimal treatment regimen can be definitively answered only by a randomized trial. We eagerly await the next Alliance study, A071801, which will randomly assign patients with a resected brain metastasis and 1-4 unresected brain lesions to SRS vs FSRT. This study will randomly assign at least 208 patients in a 1:1 fashion, with a primary outcome of surgical bed tumor control, and secondary endpoints including overall survival, quality of life, safety, and pattern of failure.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest statement. JDP reports grant funding from Varian Medical Systems and speaking fees from DePuy Synthes outside the submitted work. PDB reports personal fees from UpToDate (contributor) outside the submitted work. DR has had speaking fees and/or research from Elekta, Varian Medical Systems, Accuray, BrainLab, and Siemens Healthineers. JG and DRJ have nothing to declare.