Intracranial control and radiographic changes with adjuvant radiation therapy for resected brain metastases: whole brain radiotherapy versus stereotactic radiosurgery alone
- 623 Downloads
The aim of this study was to compare outcomes of postoperative whole brain radiation therapy (WBRT) to stereotactic radiosurgery (SRS) alone in patients with resected brain metastases (BM). We reviewed records of patients who underwent surgical resection of BM followed by WBRT or SRS alone between 2003 and 2013. Local control (LC) of the treated resected cavity, distant brain control (DBC), leptomeningeal disease (LMD), overall survival (OS), and radiographic leukoencephalopathy rates were estimated by the Kaplan–Meier method. One-hundred thirty-two patients underwent surgical resection for 141 intracranial metastases: 36 (27 %) patients received adjuvant WBRT and 96 (73 %) received SRS alone to the resection cavity. One-year OS (56 vs. 55 %, p = 0.64) and LC (83 vs. 74 %, p = 0.31) were similar between patients receiving WBRT and SRS. After controlling for number of BM, WBRT was associated with higher 1-year DBC compared with SRS (70 vs. 48 %, p = 0.03); single metastasis and WBRT were the only significant predictors for reduced distant brain recurrence in multi-variate analysis. Freedom from LMD was higher with WBRT at 18 months (87 vs. 69 %, p = 0.045), while incidence of radiographic leukoencephalopathy was higher with WBRT at 12 months (47 vs. 7 %, p = 0.001). One-year freedom from WBRT in the SRS alone group was 86 %. Compared with WBRT for patients with resected BM, SRS alone demonstrated similar LC, higher rates of LMD and inferior DBC, after controlling for the number of BM. However, OS was similar between groups. The results of ongoing clinical trials are needed to confirm these findings.
KeywordsBrain metastases Whole brain radiation therapy Stereotactic radiosurgery Frameless radiosurgery Post-operative
In 2012, 10–30 % of all adult cancer patients developed brain metastases (BM) in the United States . Nearly 80 % of these cases originated from three histologic primaries: lung, breast, and melanoma . Recent advances in systemic therapy have demonstrated significant increases in control of extracranial disease for these patients [2, 3]; however with limited penetration of the blood–brain barrier by many of these agents, there has been a concomitant rise in BM and related surgical intervention . As a result, BM remains an important cause of morbidity and mortality in this population.
Surgery for BM is generally indicated to establish a pathologic diagnosis or for the treatment of neurologic symptoms from local mass effect from the tumor. Although very effective for symptomatic relief, in a recent randomized trial, Kocher et al.  demonstrated that surgical resection alone results in a high local failure rate of 59 % at 2 years. Consequently, adjuvant treatment regimens such as whole brain radiation therapy (WBRT) and/or stereotactic radiation therapy (SRS) have been used to improve local control after surgical resection.
For patients with intact BM, randomized controlled studies comparing WBRT and SRS versus SRS alone have demonstrated comparable survival [5, 6]; Chang et al. , additionally demonstrated that the addition of WBRT to SRS significantly increased the rate of neurocognitive decline. As a result, SRS with careful imaging follow-up is a standard of care for patients with limited BM. For patients with resected BM, there are a number of non-comparative series in the literature describing the outcomes of SRS alone following surgical resection (Supplemental Table 1). In general these studies have demonstrated adequate local control with SRS and the ability to salvage distant brain recurrence (DBF) with further local therapies [8, 9, 10, 11, 12]. Given the potential deleterious effect of higher intracranial relapse rates with SRS alone after resection of BM, we undertook this institutional review to compare intracranial control, rates of radiographic leukoencephalopathy, and survival associated with adjuvant WBRT compared to SRS alone for patients with resected BM.
Methods and materials
Following IRB approval, we retrospectively reviewed the records of 136 consecutive patients with BM who underwent surgical resection followed by radiation therapy from 2003 to 2012. A total of 96 patients with 99 cavities underwent SRS to the resection cavity, and 36 patients with 42 cavities underwent postoperative WBRT. Treatment choice was based on physician preference and year of presentation, with SRS to the resection cavity first offered as a treatment option in 2007. Patients with radiosensitive tumors (e.g. small-cell cancer, lymphoma, seminoma) were excluded. Patient who had received prior WBRT were also excluded.
The decision for initial surgical resection was made by the treating neurosurgeon on the basis of tumor size, location, and associated symptoms. Neurosurgical resection was then performed by experienced surgeons using standard technique; gross total resection of each lesion was achieved whenever possible. Gross total resection was defined as removal of all enhancing tumor on brain imaging obtained within 48 h post-operatively.
Technical details of SRS have been previously described [10, 13]. In brief, SRS was performed using a linear accelerator with 6-megavoltage (MV) photon energies. The MRI defined cavity constituted the gross tumor volume (GTV). The GTV was expanded by 1 mm to create the clinical target volume (CTV) which was expanded by another 1 mm to generate the planning target volume (PTV). The resection tract was not purposefully included in the GTV or CTV. Cavities up to 20 mm in diameter were typically treated to 21 Gy, 21–30 mm in diameter to 18 Gy, and 31–40 mm in diameter to 15 Gy. Patients with large resection cavities (typically >40 mm in diameter) were treated with fractionated radiosurgery over 3–5 fractions using a frameless radiosurgery technique .
WBRT was performed using a linear accelerator, using opposed lateral fields and 6 MV photons. Multileaf collimators were used for beam shaping and shielding the globes, lenses, and pharynx. The most common fractionation regimens were 30 Gy in 10 fractions over 2 weeks and 37.5 Gy in 15 fractions over 3 weeks. The later fractionation was more commonly used in younger patients with fewer intracranial and extracranial metastases in order decrease potential neurocognitive sequelae.
Baseline patient characteristics
SRS (96 patients, 99 cavities)
WBRT (36 patients, 42 cavities)
Median age (range)
Active systemic disease
Extent of resection
Number of metastasis
Median cavity size
Median days to chemotherapy
Follow up after radiation treatment generally consisted of a clinical examination and MRI of the brain with and without contrast 4–6 weeks later, followed by clinical exam and MRI brain imaging every 3 months thereafter, unless clinically indicated at an earlier time point. Patients were followed until death or loss to follow up.
Patient characteristics were compared between those treated with WBRT and SRS by Wilcoxon rank-sum test for continuous variables, and Chi square test or Fisher’s exact test for categorical variables, where appropriate.
Time-to-event analyses were measured from the date of initiation of radiation therapy. Local recurrence (LR) was defined as the presence of new progressive nodular enhancement involving the resection cavity seen on MRI with contrast. If there was a question of the nodular enhancement representing radiation necrosis additional functional imaging was obtained (e.g. MR perfusion, MR spectroscopy, or brain positron emission tomography [PET]) to evaluate this further. Distant brain failure (DBF) was defined as presence of new enhancing lesions distinctly outside the prior treatment zone. For patients with multiple lesions, recurrence or progression at unresected sites was considered DBF. Leptomeningeal disease (LMD) was defined as new abnormal leptomeningeal enhancement of the brain, spinal cord, or cauda equina and focal or diffuse enhancement of the dura or subarachnoid space. For the LR, DBF, and LMD analyses, patients were censored at time of last brain imaging or salvage WBRT, whichever came first. For OS, death from any cause was defined as the event, and patients were censored at time of last follow-up. Clinical radiation necrosis was defined on the basis of two radiographic features: the development of contrast enhancing mass within prior radiation treatment fields and conventional imaging features, including soap bubble appearance ; advanced imaging, such as decreased perfusion on dynamic susceptibility contrast MRI perfusion was also utilized. Leukoencephalopathy is a radiographic diagnosis defined using the National Cancer Institute definition of common terminology criteria for adverse events: the development of either ventriculomegaly, increased in MRI T2/FLAIR hyperintensity within the peri-ventricular white matter, or increase in subarachnoid space .
Survival functions were calculated by the Kaplan–Meier product-limit method and the log-rank test was used to assess the difference in estimated rates between patients treated with WBRT or SRS . Figures were truncated when <10 % of patients were remaining at risk. Multivariate analysis (MVA) was performed using the Cox proportional hazards model. All potentially prognostic covariates were initially entered and a backwards stepwise selection method was employed with p = 0.1 entry and p = 0.2 removal criteria . All analyses were carried out using the SPSS version 20.0 statistical software package (IBM Inc., Armonk, NY, USA). All statistical analyses were 2-sided, and p values <0.05 were considered statistically significant.
Patient and treatment characteristics
A total of 132 patients with 141 treated cavities were included in the analysis. Ninety-nine cavities (70 %) were treated with SRS alone. The median imaging follow-up period for patients without local recurrence was 9.7 months (range 0.2–96.6 months). The median overall follow-up period for alive patients was 19 months (range 1.5–96.6 months). Median time from surgery to post-operative SRS and WRBT was similar: 30 (range 14–86) and 24 days (range 13–55) respectively (p = 0.23). Median post-operative cavity size was also similar between both groups: 7.19 (range 0.90–35.70) and 7.56 cm2 (range 0.88–30.68) (p = 0.85) respectively.
Patients receiving WBRT or SRS had similar baseline characteristics except for the number of BM (Table 1). The SRS cohort had a significantly higher proportion of patients with single BM than the WBRT cohort (71 vs. 39 %, p = 0.001); Median SRS dose was 18 Gy (range 13–30 Gy). Median WBRT dose was 30 Gy (range 18–37.5 Gy).
Of the 96 patients who were treated with post-operative SRS, 15 developed a LR and 48 developed a DBF. Of the local failures, 4 patients were treated with repeat SRS, 4 were treated with surgery alone, 2 had surgery followed by SRS, and 5 had salvage WBRT. Of the 48 DBF events, 36 were treated salvage SRS, 11 were treated WBRT, and 1 was managed with hospice care only. Overall, 85 (86 %) out of the 96 patients who underwent post-operative SRS had not required salvage WBRT at one year (Supplemental Fig. 1). On MVA, multiple brain metastases (HR 3.41, 95 % CI 1.02–11.74, p = 0.047) and presence of extracranial metastases (HR 9.81, 95 % CI 1.22–78.8, p = 0.023) were significantly associated with use of salvage WBRT after SRS alone.
Radiographic leukoencephalopathy was observed in 28 (21 %) patients; the median time to developing leukoencephalopathy was 9.8 months (range 0.8–39.5 months). The 1 year rate of radiographic leukoencephalopathy was significantly higher with WBRT compared to SRS (47 vs. 7 %, p = 0.001) (Supplemental Fig. 2).
The rate of any clinical radiation necrosis were higher in patients receiving SRS compared with WBRT (27 vs. 0 %, p = 0.001). Symptomatic radiation necrosis, defined as patients requiring steroids or other interventions, was also more commonly observed in patients receiving SRS relative to WBRT (13 vs. 0 %, p = 0.001). For the 12 patients with symptomatic radiation necrosis, treatment was steroids for 7, surgery for 3, Bevacizumab for 1, and hyperbaric oxygen therapy for 1.
Post-operatively, the standard of care for resected BM has been WBRT. Due to potential neurocognitive decline from WBRT, many institutions have demonstrated the safety and efficacy of post-operative SRS alone to the resection cavity [8, 9, 10, 11, 12, 19]. However, to the best of our knowledge, no studies published to date have directly compared this treatment approach to the standard of care (WBRT). This current study uniquely compares the outcomes between patients treated with post-operative WBRT or post-operative SRS alone at a single institution.
We found that WBRT and SRS alone have similar rates of cavity LC, but with significantly higher risk of DBF with SRS alone after adjusting for number of brain metastases. Despite higher intracranial failure with SRS alone, there was no detriment in OS with SRS alone. These results are in line with randomized phase III studies in the intact BM setting [4, 6, 7].
Both treatment groups were similarly balanced for potential prognostic factors except for the number of BM, which is related to the inherent selection bias governing the use of WBRT. Additionally, RPA class and extent of resection did demonstrate trends toward difference between the SRS and WBRT cohorts. To best adjust for these group imbalances, we used MVA in order to adjust analyses by significant prognostic factors.
The one-year actuarial LC for SRS (74 %) and WBRT (83 %) were similar. Several prospective studies independently support these LC rates: Kocher et al.  demonstrated with post-operative WBRT, a one-year LC rate of 79 % in patients with 1–3 metastasis; post-operative SRS demonstrated a one-year LC rate of 78 % in patients with 1–2 brain metastasis treated in a prospective, phase II clinical trial at Memorial Sloan Kettering . Additionally, prior retrospective studies also demonstrate similar one-year LC with post-operative SRS alone (Supplemental Table 1) [8, 9, 10, 11, 12].
With regard to DBF, WBRT and SRS demonstrated no statistically significant difference in univariate analysis. However, the treatment cohorts were imbalanced at baseline in terms of the proportion of patients with multiple BM. The presence of multiple BM is a known prognostic factor for both OS and increased risk of distant brain failure [20, 21] After adjusting for the number of brain lesions, WBRT was associated with significantly improved distant brain control relative to SRS (p = 0.03). These results are also in line with randomized trials of local therapy versus local therapy with WBRT in the intact or resected BM setting [4, 6, 7].
In addition to lower rates of DBF, WBRT, when combined with SRS, has been shown to decrease rates of LMD compared to SRS alone in the intact BM setting . Our retrospective series illustrated that the association between WBRT and reduced LMD rates in the post-operative setting as well. Rates of LMD with SRS alone and with WBRT were also similar to prior retrospective studies [22, 23].
In patients with unresected brain metastases, predictors for radiation necrosis include volume of high dose irradiated tissue, dose per fraction, and the number of fractions . Utilizing standard SRS doses that take into account these parameters , we observed a symptomatic radiation necrosis rate of 13 %. This rate is similar to Brennan et al. , who prospectively observed a 17 % rate of symptomatic radiation necrosis in resection cavities treated with similar SRS doses. With low dose fractionated therapy WBRT, we observed no cases of radiation necrosis; these low rates are comparable to the trial of Kocher et al.  experience, where similar doses of WBRT were used post-operatively.
As part of the paradigm of SRS alone, patients are followed by close imaging surveillance. For patients with a limited volume of DBF identified with surveillance, we also sought to describe the use of salvage SRS alone and WBRT. For patients with DBF after SRS alone, the majority (80 %) were effectively salvaged with SRS alone. Overall, 86 % of patients treated with SRS alone were free from salvage WBRT at one year. Our rates are consistent with the study by Aoyama et al. , which demonstrated an 83.6 % freedom from salvage WBRT rate in patients with unresected lesions receiving SRS alone. We also demonstrated that rates of leukoencephalopathy, a form of white matter damage correlated with neurocognitive dysfunction , are significantly higher at 1 year in patients with WBRT vs. SRS (47 vs. 7 %, p = 0.001). Taken together, these findings in the post-resection setting support the SRS alone treatment paradigms meant to spare the excess detrimental neurocognitive sequelae associated with WBRT.
Limitations of this study include its retrospective design, potential for selection bias due to the non-randomized treatment cohorts, difference in time periods of patient treatment, with most WBRT treatments occurring prior to 2007, potential difference in the extent/burden of extra-cranial metastases, and lack of neurocognitive testing and quality of life/neurocognitive data for comparison between arms. Strengths of this study include its relatively large patient numbers, homogenous patient treatment methods and follow-up/surveillance schedule, and use of multivariate analysis to adjust for potential confounding variables.
Resection of BM followed by SRS to the surgical cavity produced similar local control rates to WBRT. After controlling for number of brain metastases, WRBT was associated with a significantly reduced risk of DBF compared with SRS alone. However, this benefit in intracranial control did not translate into improved OS, similar to results from randomized trials in the intact BM setting. Treatment with SRS was associated with a significantly increased rate of symptomatic radiation necrosis and LMD occurrence relative to WBRT. With no detriment in overall survival despite these recurrence differences, we conclude that post-operative SRS alone is a viable treatment option for patients with resected BM. However, proper patient selection is necessary to minimize risk of radiation necrosis and LMD. We believe that close imaging surveillance is essential to allow for salvage SRS for DBF and is an important part of this treatment paradigm. We eagerly await the results of the ongoing randomized phase III trial, NCCTG-N107C, which includes patients with 1–4 BM, of which at least 1 is resected, and will report on oncologic and neurocognitive outcomes of patients receiving post-operative WBRT alone versus post-operative SRS alone .
Conflict of interest
No conflicts of interest exist for any of the authors except Constantinos Hadjipanayis, MD, Consultant, Meditech Incorporated.
- 5.Kocher M, Soffietti R, Abacioglu U, Villa S, Fauchon F, Baumert BG, Fariselli L, Tzuk-Shina T, Kortmann RD, Carrie C, Ben Hassel M, Kouri M, Valeinis E, van den Berge D, Collette S, Collette L, Mueller RP (2011) Adjuvant whole-brain radiotherapy versus observation after radiosurgery or surgical resection of one to three cerebral metastases: results of the EORTC 22952-26001 study. J Clin Oncol 29:134–141. doi:10.1200/JCO.2010.30.1655 PubMedCentralPubMedCrossRefGoogle Scholar
- 6.Aoyama H, Shirato H, Tago M, Nakagawa K, Toyoda T, Hatano K, Kenjyo M, Oya N, Hirota S, Shioura H, Kunieda E, Inomata T, Hayakawa K, Katoh N, Kobashi G (2006) Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: a randomized controlled trial. JAMA 295:2483–2491. doi:10.1001/jama.295.21.2483 PubMedCrossRefGoogle Scholar
- 7.Chang EL, Wefel JS, Hess KR, Allen PK, Lang FF, Kornguth DG, Arbuckle RB, Swint JM, Shiu AS, Maor MH, Meyers CA (2009) Neurocognition in patients with brain metastases treated with radiosurgery or radiosurgery plus whole-brain irradiation: a randomised controlled trial. Lancet Oncol 10:1037–1044. doi:10.1016/s1470-2045(09)70263-3 PubMedCrossRefGoogle Scholar
- 8.Hartford AC, Paravati AJ, Spire WJ, Li Z, Jarvis LA, Fadul CE, Rhodes CH, Erkmen K, Friedman J, Gladstone DJ, Hug EB, Roberts DW, Simmons NE (2013) Postoperative stereotactic radiosurgery without whole-brain radiation therapy for brain metastases: potential role of preoperative tumor size. Int J Radiat Oncol Biol Phys 85:650–655. doi:10.1016/j.ijrobp.2012.05.027 PubMedCrossRefGoogle Scholar
- 9.Mathieu D, Kondziolka D, Flickinger JC, Fortin D, Kenny B, Michaud K, Mongia S, Niranjan A, Lunsford LD (2008) Tumor bed radiosurgery after resection of cerebral metastases. Neurosurgery 62: 817–823; discussion 823–814 doi:10.1227/01.neu.0000316899.55501.8b
- 10.Prabhu R, Shu HK, Hadjipanayis C, Dhabaan A, Hall W, Raore B, Olson J, Curran W, Oyesiku N, Crocker I (2012) Current dosing paradigm for stereotactic radiosurgery alone after surgical resection of brain metastases needs to be optimized for improved local control. Int J Radiat Oncol Biol Phys 83:e61–66. doi:10.1016/j.ijrobp.2011.12.017 PubMedCrossRefGoogle Scholar
- 16.Institute NC (2009) Common terminology criteria for adverse events, v4.0Google Scholar
- 18.Cox DR (1972) Regression models and life tables. J R Stat Soc 34:187–220Google Scholar
- 19.Brennan C, Yang TJ, Hilden P, Zhang Z, Chan K, Yamada Y, Chan TA, Lymberis SC, Narayana A, Tabar V, Gutin PH, Ballangrud A, Lis E, Beal K (2014) A phase 2 trial of stereotactic radiosurgery boost after surgical resection for brain metastases. Int J Radiat Oncol Biol Phys 88:130–136. doi:10.1016/j.ijrobp.2013.09.051 PubMedCrossRefGoogle Scholar
- 20.Sawrie SM, Guthrie BL, Spencer SA, Nordal RA, Meredith RF, Markert JM, Cloud GA, Fiveash JB (2008) Predictors of distant brain recurrence for patients with newly diagnosed brain metastases treated with stereotactic radiosurgery alone. Int J Radiat Oncol Biol Phys 70:181–186. doi:10.1016/j.ijrobp.2007.05.084 PubMedCrossRefGoogle Scholar
- 23.Atalar B, Modlin LA, Choi CY, Adler JR, Gibbs IC, Harsh GRt, Li G, Nagpal S, Hanlon A, Soltys SG (2013) Risk of leptomeningeal disease in patients treated with stereotactic radiosurgery targeting the postoperative resection cavity for brain metastases. Int J Radiat Oncol Biol Phys 87:713–718. doi:10.1016/j.ijrobp.2013.07.034 PubMedCrossRefGoogle Scholar
- 25.Shaw E, Scott C, Souhami L, Dinapoli R, Bahary JP, Kline R, Wharam M, Schultz C, Davey P, Loeffler J, Del Rowe J, Marks L, Fisher B, Shin K (1996) Radiosurgery for the treatment of previously irradiated recurrent primary brain tumors and brain metastases: initial report of radiation therapy oncology group protocol (90-05). Int J Radiat Oncol Biol Phys 34:647–654PubMedCrossRefGoogle Scholar