Whole Brain Radiotherapy Improves Overall And Intracranial Progression-free Survivals In Breast Cancer Patients With Brain Metastasis In The Setting Of Systemic Treatments

Objective: To evaluate the effects of whole brain radiotherapy(WBRT)on overall(OS)and intracranial progression-free survival(PFS)in breast cancer(BC)patients with brain metastasis(BM)in setting of systemic treatments.Methods:A retrospective analysis of 105 female patients consecutively admitted for BM from BC between 2013 and2015 at the Fourth Hospital of Hebei Medical University(China)was conducted.All BM diagnoses were supported with brain CT and/or MRI.Patient,disease,treatment characteristics and survival data till December 1 2016 were either collected or calculated.Brain radiotherapy(RT)modalities were classified:(1)none,<30,30-39 and ≥40 Gy of WBRT;(2)none,<50 and ≥50 Gy of largest BM tumor boost dose accumulated;(3)use or not use of stereotactic radiotherapy/stereotactic surgery(SRS/SRT).Cox proportional hazard regression models were used to estimate the independent effects of WBRT regimen on OS and PFS.Beside other RT modalities,the analyzed covariates included BMBC age,PS,ECOG,GPA,number of brain lesions(1,2,3 & ≥4),meningeal involvement(yes vs.not),extracranial metastasis(yes vs.not),any other treatments on BM including surgery,chemotherapy,endocrine therapy,and target drug use.The covariates also covered BC age,primary tumor surgery types,TNM,pathology,ER,PR,HER2,Ki-67,local and systemic treatments on BC which consisted of chemotherapy,post-operative RT,and endocrine therapy.Statically,mean(standard deviation)and median were summarized for continuous variables.ANOVA was used for their comparison among groups.Chi-square and Fisher’s tests(if applied)were conducted to compare the distribution ofcategorical variables.Kaplan-Meier curves and log rank tests were used for OS and PFS’s descriptions and comparisons among groups.Hazard ratios(HR)and its confidence interval(CI)from Cox models were presented along with p values.Two-sided p< 0.05 was treated as significant.Statistical tool was SAS9.20.Results:1 Description of overall study populationThe study population size was 105.BM mean age(range)were 53(27-82)years old.By number of BM sites as 1 to ≥4,their distributions were42.9%(45 case),11.4%(12),1.9%(2),and 43.8%(46),respectively.ECOG’s distribution by score 0/1 to 3/4 were 38.1%(40),32.4%(40),29.5%(31),and respectively.Meningeal involvement was 17.1%(18).Presence of any BM symptom was 66.7%(70).Patients with extracranial metastases were 17.1%(18).Patients combined with CVD,hyponatremia,and infection were30.5%(32),58.1%(61),and 17.1%(18),respectively.The median(95% CI)of OS since BM diagnosis was 11.3(2.0-22.9)months(1 month = 30 days).The overall cumulative mortality and treatment failure rate on intracranial metastatic tumor(death or relapsed)were 76.2%(80/105)and 81%(85/105),respectively.The distributions by brain RT schema listed above were(1)48.6%(51),10.5%(11),12.4%(13),and 28.6%(30)of WBRT regime(2)64.8%(68),4.8%(5),and 30.5%(32)of total RT dose categories through locally or boosted RT(3)7.6%(8)and 92.6%(97)by use or not use of SRS/SRT.Percents of patients with surgical resection,endocrine therapy,chemotherapy and target therapy on BM sites were 8.6%(9),23.8%(25),38.1%(40),and14.3%(15),respectively.The distribution of variables related to BC and its treatments were BC mean age(range)50(27-82)years old,mastectomy 89.5%(94),ALND 89.5%(94),and invasive ductal carcinoma(IDC)82.9 %(87).By HR+(either ER or PR positive)status,their distributions were HR+55.2%(58),HR﹣35.2%(37),and HR unrecorded 9.5 %(10).By HER2(IHC)status,their distributionswere HER2+ 34.3%(36),HER2﹣41.9%(44),and HER2 unassigned 23.8%(25).By Ki-67 status(defined <14% as low)listed above,their distributions were 10.5%(11),66.7%(70)and unrecorded 22.9%(24),respectively.According to TNM stage I-II,III,IV and unrecorded,their distributions were21.9%(23),52.4%(55),10.5%(11),and 15.2%(16),respectively.By other treatments on BC as of chemotherapy,endocrine therapy,target therapy,their distributions were 96.2%(101),47.6(50),and 8%(4),respectively.2 Patient comparison among WBRT subgroupsThere were statistically significant of local boost dose reached,BM with chemotherapy,TNM stage,Ki-67 level among WBRT subgroup(yes vs.no).Some of those important statistics were(1)by I-II,III,IV stage among WBRT patients,they were 34.8%(8/23),52.7%(29/55),and 63.6%(7/11),respectively,(2)similarly among those non-WBRT patients,they were65.2%(15/23),47.3%(26/55),and 36.4%(4/11),respectively.In general,there were some tendency of WBRT patients were given more frequently to the BC patients with advanced stage.Some denovo Stage IV with BM could explain this link.No statistical differences(p>0.05)were detected between WBRT subgroups with HR and HER2 status.3 Analyses on OS among WBRT subgroupsFor all patients,the crude mortalities by WBRT subgroups ordered above were 76.5%(39/51),81.8%(9/11),76.9%(10/13),and 73.3%(22/30).Their median of OS month were 5.2,3.7,13.8,and 13.3 months,respectively.Compared to none,the univariate COX analysis on OS of other WBRTs did not show their significant differences.However,the multivariable Cox analyses show that the 30-39 Gy and≥40Gy subgroup have significant differences of OS(P<0.05)compared to patients without WBRT.There was no differences between 30-39 Gy and≥40Gy subgroups(P>0.05).That suggested WBRT does≥30Gy was associated with improved survival among BCBM patietns.For the locally or boosted RT patients listed above,the crude mortalities were 76.5%(52/68),100.0%(5/5),and 71.9%(23/32).Their median monthsof OS were 6.0,9.1,and 14.1months,respectively.The univariate COX analyses showed there was significant differences of OS between patients with0 Gy and ≥50Gy(P<0.05).Further,the multivariable Cox analyses continued to show their significant differences of OS.This results might suggest the local dose ≥50Gy is associated with the improved OS.By SRS/SRT status,the crude mortality rates were 62.5%(5/8)and77.3%(75/97).Their median months of OS were 11.3 and 10.6 months,respectively.The univariate COX analyses did not show the significant differences of OS between two groups.However,the multivariable Cox analyses showed the significant differences of OS between both subgroups.That suggested SRS could independently improve the OS.4 Analyses on PFS among WBRT subgroupsOverall,the treatment failure rates by WBRT subgroups listed above were 84.3%(43/51),81.8%(9/11),84.6%(11/13),and 73.3%(22/30),respectively.Compared with none,the univariate and multivariate COX analyses showed the significant improved PFS of BCBM patients with WBRT≥30Gy.However,no statistical difference of PFS was found between patients with WBRT 30-39 Gy and ones with WBRT ≥40Gy by both analyses.This result suggested that WBRT ≥30Gy be associated with improved PFS of BCBM patients.By locally or boosted RT patients listed above,the treatment failure rates for each subgroup were 82.4%(56/68),100.0%(5/5),and 75.0%(24/32).Both univariate and multivariable COX analyses showed that there were significant differences of improved PFS for patients with <50 Gy and ≥50Gy(P<0.05),compared to the patients with 0 Gy.However,no survival difference was noted between patient with <50 Gy and one with ≥50Gy.This results demonstrated the improved PFS with higher tumor dose(close to or over50Gy).By SRS/SRT presence and absence,their treatment failure rates were87.5%(7/8)and 80.4%(78/97).Their median months of PFS were 11.3 and10.6 months,respectively.Both univariate and multivariable Cox analysesdemonstrated the significant differences of PFS between two group.5 Analyses of OS and PFS among patient subgroups by biomarkersBy HR status ordered above,the crude mortalities were 78.4%(29/37),74.1%(43/58),and 80%(8/10),respectively;their median months of OS were9.0,12.2,and 6.1 months,respectively.By HER2 status ordered above,the crude mortalities were 75%(27/36),79.5%(35/44),and 72%(18/25),respectively;their median months of OS were 11.9,11.3,and 9.0 months,respectively.By Ki-67 subgroup ordered above,the crude mortalities were81.8%(9/11),75.7%(53/70)and 75%(18/24);their median months of PFS were 15.9,11.3,and 9.0 months,respectively.Both univariate and multivariable COX analyses had not showed their significant associations with OS.By the same order of HR status,their treatment failure rates were81.1%(30/37),79.3%(46/58),and 90%(9/10),respectively.By HER2 status listed the same above,their treatment failure rates were 88.9%(32/36),79.5%(35/44),and 72%(18/25),respectively.By Ki-67 status above,their treatment failure rates were 81.8%(9/11),81.4%(57/70)and 9.2%(19/24).The univariate COX analyses did not show the significant differences of PFS among subgroups by HR status although the HR was 1.395(95%CI,0.878-2.217,P=0.159)of HR-,compared to HR+ group(reference).Univariate analysis did not show the significant PFS differences of subgroups by HER2 or Ki-67 status.Please note that the final analyses on OS and PFS have used hormonal treatment(HT)and target therapy on BM,instead of the HR and HER2 status given their very strong associations.Through multivariable Cox analysis,the target therapy demonstrated its statistically significant effect on OS only.Both multivariable analyses on OS and PFS showed the significant role of Ki67 status.6 Comparative analyses of OS and PFS among subgroups by BM age,ECOG and othersThe multivariate analysis indicated that age <35 was no longer associatedwith the significantly poor OS which demonstrated at univariate analysis.Only 2.8 median months of OS was for BCBM patients with age <35.Compared to patients with ECOG 0-1,both univariate and multivariable Cox analyses that patients with other ECOG were associated with poor OS and PFS.Compared to patients with single BM site,the patients with ≥4 BM sites had significantly poor OS and PFS.That was supported by both univariate and multivariable analyses.Last,both univariate and multivariable COX analyses demonstrated the significantly poor survival of BM patients with meningeal involvement.The median months of OS for them was 2.7 months only.Conclusions:1 WBRT associates independently with improved OS and PFS in setting of modern systemic treatments for BCBM patients from BC.WBRT dose ≥30Gy is adequate for both benefits.2 Target therapy on BCBM is associated with improved OS.3 The survival risk factors of BCBM patents include BM age < 35,meningeal involvement,number of BM sites ≥4,ECOG≥2,Ki-67≥14%,HRand HER2+.