A Cochrane review by Tsao et al (2012) assessed the effectiveness of whole brain radiotherapy (WBRT) either alone or in combination with other therapies in adult participants with newly diagnosed multiple brain metastases.45 The review updated a previous 2006 Cochrane review. Nine new RCTs involving 1420 participants were added to the updated review. The updated review included a total of 39 trials involving 10,835 participants with mixed primary tumours. Results are presented under stereotactic radiosurgery (SRS), WBRT, altered fractionations and radiosensitizers.45
The Tsao et al 2012 Cochrane review compared WBRT plus radiosurgery versus WBRT.45 Three trials (Andrews 2004,13 Chougule 2000, Kondziolka 199946) examining the use of WBRT with or without radiosurgery boost for up to four brain metastases (469 participants in total) were included. Two trials were fully published (Andrews 2004,13 Kondziolka 199946) and included populations of mixed primary tumours. In addition, five retrospective analyses reporting results of radiosurgery in patients with central nervous (CNS) metastases were identified; three are discussed here.14,15,83
Pooled results of two randomised controlled trials comparing WBRT plus radiosurgery and WBRT alone (Andrews 200419 and Kondziolka 199946 showed no difference in six-month survival (p=0.24).45 The Andrews trial reported improved survival (p=0.0393) for a subset of patients with a single, surgically un-resectable brain metastasis treated with WBRT and radiosurgery (6.5 months) compared to WBRT alone (4.9 months).
Kased et al 2009 retrospectively reviewed 176 patients who underwent gamma knife stereotactic radiosurgery (SRS) for brain metastases from breast cancer.83 Among the 95 patients with newly diagnosed brain metastases, median survival time was 16 months. Among the 81 patients treated for recurrent brain metastases, median survival time was 11.7 months. In patients treated with SRS alone initially, survival was 17.1 months compared to 15.9 months for patients treated with SRS and upfront WBRT (p=0.20). Factors associated with longer survival included age less than 50 years, primary tumour control, ER positivity, and HER2-positive disease.83
Akyurek 2007 retrospectively reviewed 49 breast cancer patients who underwent SRS for brain metastases; 34 patients as primary treatment, and 15 as salvage treatment following prior WBRT.15 The median overall survival for patients receiving SRS as primary treatment was 25 months and 14 months for patients receiving SRS as salvage treatment.
The Tsao et al 2012 Cochrane review pooled data for local brain control at one year from two studies (Andrews 2004 and Kondziolka 1999).13,46 A statistically significant improvement in local brain control favouring WBRT and radiosurgery boost compared to WBRT alone was observed (RR 1.20, p=0.003).
The Kondziolka 1999 trial also reported median time to local brain failure of 6 months for patients receiving WBRT alone in comparison to 36 months after WBRT and radiosurgery.46 The Andrews trial found no statistically significant difference (p=0.1278) regarding overall time to intracranial tumour progression.13
A retrospective analysis of breast cancer patients (Akyurek 2007) reported on local brain tumour control.15 One year local control rates were 79% for patients receiving SRS as initial treatment, and 77% for patients receiving salvage SRS after WBRT treatment. Two-year local control rates were 49% for the group receiving SRS alone and 46% for the SRS salvage treatment group. The difference was not statistically significant (p=0.99).15 Akyurek 2007 also reported distant brain metastases-free survival; at one year, the survival rate was 69% in the 49 patients receiving either initial SRS alone or salvage SRS.15 The one-year distant brain metastases-free survival rate was 64% among the group receiving initial SRS compared to 57% for the group receiving SRS salvage treatment. The difference was not statistically significant (p=0.62).
Kased et al 2009 reported on 176 patients who underwent SRS for brain metastases from brain cancer.83 No significant difference was observed in freedom from progression between SRS alone and SRS with WBRT in the newly diagnosed patients. The median freedom from new brain metastases was 14.8 months for patients treated with SRS alone, compared to 11.3 months for patients in the SRS and WBRT arm (p=0.83).83
Andrews 2004 reported that KPS was improved at six months in 13% of patients treated with WBRT and radiosurgery, compared to 4% of patients treated with WBRT alone (p=0.0331).13 Mental status as measured using a mini-mental status examination did not show a significant difference.
Tsao 2009 and the updated review in 2012 reported adverse effects identified in the Andrews 2004 trial.84 In this trial, early and late toxic effects did not differ greatly for patients receiving WBRT alone compared to WBRT plus radiosurgery. More patients in the WBRT plus radiosurgery arm experienced acute grade three and four toxicity (4 of 160 patients), compared to those receiving WBRT alone (0 of 166 patients). More patients in the WBRT plus radiosurgery arm experienced late grade three and four toxicity (6 of 160 patients) compared to those receiving WBRT alone (3 of 166 patients).13
WHOLE BRAIN RADIOTHERAPY (WBRT)
The Cochrane review by Tsao et al (2012) identified three RCTs comparing radiosurgery alone to radiosurgery plus WBRT.45 The trials (Aoyama 2006,14 Chang 2009,50 Kocher 201116) included patients with up to three or four brain metastases from mixed primary tumours.
The systematic review by Kalkanis et al 201085 identified one RCT (Patchell et al 1998)51 and three retrospective cohort studies that addressed the question of surgery alone versus surgery plus WBRT for the initial management of a single brain metastasis. The RCT by Patchell et al (1998) randomised patients to postoperative WBRT (50.4 Gy over 5 ½ weeks (n=49) or no further treatment (observation, n=46). The study included patients with various primary tumours including nine (9%) breast cancer patients.51
Gaspar 1997 was not identified in the systematic review as it was published prior to the search period of 2001 to 2012, but was included for additional background information on WBRT.47 A prospective study was identified (Yaneva) that assessed the effect of palliative radiotherapy on quality of life in 65 patients with brain metastases from various primary tumours (33 breast cancer, 50.8%).52
Pooled data from two trials identified in the Tsao et al Cochrane review, found no difference in overall survival (Aoyama 2006 and Chang 2009), (HR 0.98, 95% CI 0.71 to 1.35, P = 0.88).45
Aoyama et al 2006 undertook a randomised controlled trial comparing WBRT plus SRS with SRS alone for the treatment of 132 patients with one to four brain metastases from various primary tumours (7% breast cancer patients).14 Median survival time in the WBRT plus SRS group was 7.5 months compared to 8.0 months for the SRS alone group (p=0.42). Death was attributed to neurologic causes in 13 patients in the WBRT plus SRS group, and in 12 patients in the SRS alone group (p=0.64).14
Patients with one to three brain metastases from various solid tumours (12% breast cancer patients) treated with SRS or surgery were randomised to WBRT or observation in the European Organisation for Research and Treatment of Cancer (EORTC) 22952-26001 study.16
Tsao 2012 reported that for the Kocher 2011 trial, overall survival for the radiosurgery alone arm versus WBRT and radiosurgery boost could not be isolated.45 Kocher et al 2011 reported that overall survival did not differ between the two arms, with a median survival of 10.9 months for patients who had observation only (surgery alone or radiosurgery alone), compared to 10.7 months for patients treated with WBRT (surgery and WBRT or radiosurgery and WBRT) (p=0.89). Malignant disease was the dominant cause of death in both arms.16 The Patchell et al 1998 study included patients who had undergone complete surgical resection for a single brain metastasis, comparing patients randomly assigned to post-operative WBRT or no further treatment.51 Overall survival rates did not differ significantly. Among the 49 patients who received WBRT, median length of survival was 48 weeks, compared to 43 weeks for the 46 patients who did not have further treatment (p=0.39). Patients who had WBRT were more likely to die of systemic disease rather than neurologic progression (p=<0.001).51
Rades et al 2007 retrospectively investigated whether SRS alone improved outcomes compared with WBRT for patients with one to three brain metastases.86 Only patients in recursive partitioning analysis (RPA) classes 1 and 2 were included in the study. Median survival was 7 months for patients receiving WBRT compared to 13 months for patients receiving SRS.
Gaspar 1997 reviewed prognostic factors of patients with brain metastases in three RTOG trials conducted between 1979 and 1993, testing dose fractionation and radiation sensitisers.47 The majority of included patients had a lung primary tumour (61%) with patients with a breast cancer primary comprising 12% of the population. Improved survival was associated with age less than 65 years, a Karnofsky Performance Status (KPS) of at least 70, and controlled primary tumour with the brain the only site of metastases. Shorter survival was observed among patients with a KPS of less than 70.
Pooled data in three trials (Aoyama 2006; Chang 2009; Kocher 2011) identified in the Tsao 2012 Cochrane review found the addition of WBRT to radiosurgery significantly improves locally treated brain metastases control (HR 2.61, 95% CI 1.68 to 4.06, P < 0.0001) and distant brain control (HR 2.15, 95% CI 1.55 to 2.99, P < 0.00001).45
In Aoyama et al 2006 comparison of WBRT plus SRS with SRS alone, multivariate analysis shows that WBRT plus SRS was associated with a reduced risk of recurrence (p=<0.001).14 Twenty-three patients in the WBRT plus SRS group experienced either distant or local brain tumour recurrence, compared to 40 in the SRS alone group. The 12 month brain tumour recurrence rate was 46.8% in the WBRT plus SRS group and 76.4% in the SRS alone group (p=<0.001). Twenty-one patients in the WBRT plus SRS group had new brain metastases at distant sites compared with 34 in the SRS alone group. The 12-month actuarial rate of developing new brain metastases was 41.5% in the WBRT plus SRS group and 63.7% in the SRS-alone group (P=0.003). Salvage treatment for progression of brain metastases was required significantly more frequently in patients receiving SRS alone compared to the group receiving SRS plus WBRT (29 patients vs. 10 patients respectively, p=<0.001).14 Salvage WBRT was used for 11 of the patients who had SRS alone, but not used for any patient in the WBRT plus SRS group. Salvage SRS was used for 19 patients in the SRS alone group, and for 9 patients in the WBRT plus SRS.
Recurrence outcomes from the EORTC 22952-26001 study comparing WBRT to observation were reported in Kocher et al 2011.16 Median progression-free survival was slightly longer in patients receiving WBRT (4.6 months) compared to patients in the observation arm (3.4 months) (p=0.20). Extra-cranial progression was reported at similar rates; in 64% of patients in the observation arm and 66% of patients in the WBRT arm. Progression at intracranial sites occurred significantly more frequently in patients in the observation arm compared to patients in the WBRT arm (p=<0.001). After surgery and SRS, WBRT reduced the probability of relapse at initial sites and at new sites. Salvage therapies for intracranial relapse were used more frequently in patients following observation (51%) compared to patients treated with WBRT (16%).16 Thirty-one per cent of patients in the observation arm required salvage WBRT, compared to 3% in the WBRT arm.
In the Patchell et al 1998 study, patients who had WBRT after surgical resection to a single brain metastasis had significantly lower rates of tumour recurrence anywhere in the brain, compared to patients who had no further treatment after surgery (18% vs. 70%, p=<0.001).51 The time to any brain recurrence was also significantly longer (p=<0.001). Multivariate analysis showed only post-operative radiotherapy lessened the risk of brain recurrence (p=<0.001).
Aoyama et al 2006 reported systemic functional preservation rates at 12 months were 33.9% in the WBRT plus SRS group, and 26.9% in the SRS alone group (p=0.53).14
Chang et al 2009 reported on patients with one to three brain metastases from various primary tumours, assigned to SRS plus WBRT or SRS alone. The trial was halted early based on the probability of decline in neurological function among patients in the SRS plus WBRT group compared to the SRS alone group.50
Quality of life
Patients with one to three brain metastases from various solid tumours treated with SRS or surgery were randomised to WBRT or observation in the EORTC 22952-26001 study.17 Soffietti et al 2013 reported on quality of life findings, using the Health-related Quality of Life (HRQOL) scale. A statistically significant and clinically meaningful difference in global HRQOL mean scores was detected at 9 months follow-up, in favour of patients who had observation alone (p=0.0148). No differences were found at any other time points. Patients in the observation only group had better mean scores in physical, role and cognitive functioning.17
Yaneva et al 2006 evaluated the influence of WBRT on quality of life and neurologic symptoms.52 Patients with various primary tumours were included, including 50.8% breast cancer patients. All patients had a KPS above 70. After radiotherapy, all patients showed improvement in their clinical status and functioning including physical, role, emotional, cognitive and social functioning. Fatigue, pain, nausea, insomnia and appetite loss also improved significantly after WBRT.52
Aoyama et al 2006 reported that symptomatic acute neurological toxicity was observed in four of the 65 patients in the WBRT plus SRS arm, and in eight of the 67 patients in the SRS alone arm (p=0.36). Symptomatic late neurologic radiation toxic effects were observed in seven patients in the WBRT plus SRS group, and in three patients in the SRS alone group (p=0.20). Toxic effects were experienced for a median of 15.6 months in the WBRT plus SRS group and 6.2 months in the SRS alone group.14
The EORTC 22952-26001 study reported sixteen serious adverse events; 13 among patients in the WBRT arm compared to three in patients who underwent observation. Acute toxicity of WBRT was reported as mild.16
ALTERED FRACTIONATION WBRT
The Cochrane review by Tsao et al (2012) addressed altered WBRT schedules.45 A total of nine published reports involved participants randomised to altered WBRT dose-fractionation schedules compared to standard 30 Gy in 10 daily fractions (Borgelt 1980,56 Borgelt 1981,56 Chatani 1985,87 Chatani 1994,88 Haie-Meder 1993, Harwood 1977,25 Kurtz 1981,26 Murray 1997,89 Priestman 199654). One study (Haie-Meder 1993) was excluded because the trial design did not include a standard WBRT dose-fractionation arm (30 Gy in 10 fractions or 20 Gy in five fractions). Eight studies therefore met the inclusion criteria for the review (3645 participants with mixed primary tumours).
Overall survival and recurrence
The Tsao et al 2012 Cochrane review identified six trials reporting on overall survival. Three trials (Chatani 199488; Harwood 197725; Priestman 199654) compared a standard dose of 30 Gy in 10 fractions to a lower dose fractionation (20 Gy in 5 fractions, 10 Gy in a single fraction or 12 Gy in 2 fractions). Meta-analysis found a significant difference favouring the control dose of 30 Gy in 10 fractions (p=0.01). Of note, Chatani 1994 and Harwood 1977 reported no statistically significant difference, however Priestman 1996 did.
Four trials (Chatani 198587; Chatani 199488; Kurtz 198126; Murray 199789) compared a standard dose of 30 Gy in 10 fractions to a higher dose (50 Gy in 20 fractions, 54 Gy in 34 fractions). No statistically significant difference was observed in overall survival (p=0.65).45
Three retrospective studies were identified in the Cancer Australia systematic review7 that compared various radiotherapy regimens in patients with CNS metastases; two comparing shorter course WBRT with longer course and the third investigated the potential benefit of dose escalation beyond the standard 30 Gy treatment.
Rades et al (2007) investigated the potential benefit of dose escalation beyond the standard 30 Gy treatment in patients with ≥2 brain metastases from breast (26% of patients), lung and other primaries.49 Two hundred and fifty seven patients who received 30 Gy in 10 fractions (10 fractions of 3 Gy each, with an overall treatment time of 2 weeks) were compared with 159 patients who received higher doses such as 45 Gy in 15 fractions (57 patients) and 40 Gy in 20 fractions (102 patients).49 Rades et al found dose escalation beyond 30 Gy in 10 fractions did not improve survival (p=0.86) or local control (p=0.61).49 Univariate and multivariate analyses of recurrence of brain metastases showed a significant association between breast cancer as the primary tumour and improved local control (p=<0.001 and p=0.012, respectively).
Rades and Lohrynska et al (2007) retrospectively compared survival and local control for short-course WBRT compared with longer programs in breast cancer patients.53 Sixty-nine patients received short course WBRT with 20 Gy in 5 fractions. Long course WBRT with either 30 Gy in 10 fractions or 40 Gy in 20 fractions was given to 138 patients.53 The WBRT schedule was not found to be associated with survival (p=0.254) or local control (p=0.397).
In another retrospective study by Rades et al (2011) shorter course and longer course WBRT were compared for elderly patients (≥ 65 years) treated between 2001 and 2010 for brain metastases.48 The analysis compared 162 patients (23 breast cancer patients, 14%) who received 20 Gy in 5 fractions and 293 patients (53 breast cancer patients, 18%) who received 30 Gy in 10 fractions.48 On univariate analysis, the WBRT regimen of 20 Gy in 5 fractions was significantly associated with improved overall survival (vs. 30 Gy in 10 fractions, p=0.020), however this was not maintained on multivariate analysis (p=0.13). The WBRT regimen was not significantly associated with improved local control (p=0.32). Breast cancer as the primary tumour (vs. lung cancer or other tumours) almost reached statistical significance for improved local control (p=0.054).48
The Tsao 2012 systematic review identified three studies (Borgelt 198055; Borgelt 198156; Kurtz 198126) reporting on neurological function for patients with a baseline neurological function of grade two or three.45 Among these patients there was a statistically significant difference in neurological function improvement favouring those treated with the control dose (30Gy in 10 fractions) compared to a lower dose (OR 1.74, p=0.03). There was no statistically significant difference in rates of neurological function improvement for those treated with higher doses compared to the control dose (OR 1.14, p=0.23).45
The Rades et al 200749 and Rades and Lohrynska et al 200753 retrospective reviews reported no significant differences in grade three toxicity among patients receiving a control dose of 30 Gy in 10 fractions compared to patients on other WBRT regimens.
Rades et al 2007 reported similar rates of grade 3 acute toxicity among patients receiving 30 Gy (5.8%) and higher doses (5%, p=0.92).49 Neurocognitive dysfunction was noted in six patients treated with 30 Gy in 10 fractions (2.3%) compared to eight patients (5%) treated with higher doses (p=0.24).
Rades and Lohrynska et al 2007 reported that 9% of patients treated with 20 Gy in 5 fractions experienced grade three acute toxicity, compared with 4% of patients receiving 30 Gy in 10 fractions.53 The rates of > grade three late toxicity were less than 5% in each treatment group.
WBRT PLUS RADIOSENSITIZERS
The Cochrane review by Tsao et al (2012)45 identified six published trials (DeAngelis 1989,90 Eyre 1984,91 Komarnicky 1991,92 Mehta 2003,93 Phillips 1995,94 Suh 200695) examining the use of radiosensitizers in addition to WBRT (2016 participants with mixed primary tumours). The radiosensitizers used were lonidamide (DeAngelis 198990), metronidazole (Eyre 198491), misonidazole (Komarnicky 199192), bromodeoxyuridine (BrdU) (Phillips 199594), motexafin gadolinium (Mehta 200393) and efaproxiral (Suh 200695).
The Cochrane review reported that the addition of radiosensitizers in the identified RCTs did not confer additional benefit to WBRT in either the overall survival times (HR 1.08, 95% CI 0.98 to 1.18, P = 0.11) or brain tumour response rates (HR 0.87, 95% CI 0.60 to 1.26, P = 0.46).45
HER2 STATUS AND RADIOTHERAPY
A retrospective study, Wolstenholme et al (2008) assessed whether HER2 status had an effect on outcomes after WBRT.57 A total of 181 patients with known HER2 status were included in the study (88 HER2-positive and 93 HER2-negative).
Dawood et al 2010 retrospectively reviewed the effect of receptor status in 223 women with breast cancer and brain metastases.58 Sixty-seven patients had hormone receptor-positive/HER2-negative disease, 101 had HER2-positive disease, and 54 had triple-negative disease.
Significantly longer survival for HER2-positive compared to HER2-negative patients was reported in these two retrospective studies following WBRT.57,58 Dawood et al 2010 found that the risk of death among women with triple-negative disease was not significantly different from women with hormone receptor-positive/HER2-negative disease (p=0.54).58
Matsunaga et al 2010 reviewed prognostic factors for women undergoing Gamma Knife surgery for brain metastases from breast cancer between 1992 and 2008.96 Of the 101 included patients, 28 had HER2-positive disease, 37 had luminal A or B disease*, 36 had triple-negative disease. Median overall survival for women with HER2-positive disease (25 months) was significantly longer than survival with luminal (12 months) or triple-negative disease (5 months) on univariate and multivariate analyses (p=0.001). The difference in overall survival between patients with luminal disease and triple-negative disease was not statistically significant (p=0.569). There was no statistically significant differences between the three breast cancer subtypes for the incidence of new brain metastases following initial Gamma Knife surgery.96
± Note: the term radiosurgery in these guidelines applies to the use of a single dose (or limited number of doses) of ablative radiotherapy to brain metastases using highly precise immobilisation, dosimetric planning, delivery and verification system and can include (but is not limited to) stereotactic radiosurgery, gamma knife radiosurgery, Cyber knife radiosurgery or radiosurgery delivered using Tomotherapy or IMRT/VMAT.
* Luminal A – hormone receptor-positive/HER2-negative disease; Luminal B – hormone receptor-positive/HER2-positive disease