Is more better? The impact of extended adjuvant temozolomide in newly diagnosed glioblastoma: a secondary analysis of EORTC and NRG Oncology/RTOG

Abstract

Background:

Radiation with concurrent and adjuvant (6 cycles) temozolomide (TMZ) is the established standard of postsurgical care for newly diagnosed glioblastoma (GBM). This regimen has been adopted with variations, including extending TMZ beyond 6 cycles. The optimal duration of maintenance therapy remains controversial.

Methods:

We performed pooled analysis of individual patient data from 4 randomized trials for newly diagnosed GBM. All patients who were progression free 28 days after cycle 6 were included. The decision to continue TMZ was per local practice and standards, and at the discretion of the treating physician. Patients were grouped into those treated with 6 cycles and those who continued beyond 6 cycles. Progression-free and overall survival were compared, adjusted by age, performance status, resection extent, and MGMT methylation.

Results:

A total of 2214 GBM patients were included in the 4 trials. Of these, 624 qualified for analysis 291 continued maintenance TMZ until progression or up to 12 cycles, while 333 discontinued TMZ after 6 cycles. Adjusted for prognostic factors, treatment with more than 6 cycles of TMZ was associated with a somewhat improved progression-free survival (hazard ratio [HR] 0.80 [0.65–0.98], P = .03), in particular for patients with methylated MGMT (n = 342, HR 0.65 [0.50–0.85], P < .01). However, overall survival was not affected by the number of TMZ cycles (HR = 0.92 [0.71–1.19], P = .52), including the MGMT methylated subgroup (HR = 0.89 [0.63–1.26], P = .51).

Conclusions:

Continuing TMZ beyond 6 cycles was not shown to increase overall survival for newly diagnosed GBM.

glioblastoma, temozolomide, adjuvant, maintenance, treatment duration

Importance of the study

Following biopsy or resection, radiation (RT) with concurrent and adjuvant (6 cycles) TMZ has been established as standard of care for newly diagnosed GBM. This regimen has been widely adopted; however, in practice the recommended duration of adjuvant or maintenance TMZ varies greatly. Although the original description of the regimen calls for up to 6 cycles of TMZ, many investigators prolong treatment for up to 12 or even more cycles in nonprogressive patients. The optimal duration of maintenance therapy remains a matter of debate. We present a retrospective secondary analysis of a pooled database of trials by the European Organisation for Research and Treatment of Cancer and the NRG Oncology/Radiation Therapy Oncology Group in newly diagnosed GBM patients and compare the outcome of nonprogressive patients being treated with TMZ chemotherapy for a prolonged period beyond the established 6 cycles.

Surgical tumor resection to the extent safely feasible, followed by radiotherapy (RT), has been the backbone of therapy for patients with newly diagnosed glioblastoma (GBM) for over a decade. In 2005 concomitant chemoradiotherapy with temozolomide (TMZ) followed by 6 cycles of maintenance TMZ (TMZ/RT→TMZ) became the standard-of-care therapy following tumor resection or biopsy.¹ The duration of adjuvant TMZ chemotherapy was arbitrarily established at 6 cycles, in analogy to many other cancer protocols, but the optimal number of cycles has never been formally evaluated in a comparative fashion. This regimen has been uniformly adopted and the FDA approved the treatment regimen of 6 adjuvant or maintenance cycles of TMZ. In daily practice, however, the prescribed number of cycles for patients without tumor progression after 6 months varies greatly. In the United States, patients commonly receive up to 12 or even more cycles of maintenance TMZ following chemoradiation, despite lack of data.² Many previous and ongoing trials of chemotherapy for malignant glioma prescribe maintenance TMZ for up to 12 months.

We set out to investigate whether prolonged maintenance TMZ administration confers a survival advantage. We compared the pooled data from 4 large randomized trials where prolonged maintenance treatment for up to one year was allowed as per local practice.

Materials and Methods

Individual patient data from 4 randomized prospective phase III^1,3,4 and phase II⁵ trials were pooled. The latest data update was used—that is, data up to 2007 for the European Organisation for Research and Treatment of Cancer/National Cancer Institute of Canada (EORTC/NCIC) trial with a median follow-up of 5.14 years⁸; for Radiation Therapy Oncology Group (RTOG) 0525/EORTC 26052-Intergroup as of 2011 (median follow-up of 2.95 y); and for CENTRIC/CORE as of 2013 (median follow-up of 2.44 y/1.90 y, respectively). In these trials patients received TMZ/RT either as experimental treatment (EORTC/NCIC 26981-22981/CE.3)¹ or as the control therapy in subsequent trials aiming to further improve outcome (EORTC 26071-22072-CENTRIC³; EMD Serono-CORE⁶ and RTOG 0525/EORTC 26052-Intergroup⁷). Inclusion criteria for all 4 trials consisted of histologically confirmed newly diagnosed GBM and no prior therapy other than surgery or biopsy. World Health Organization (WHO) performance status (PS) of ≤2 or Karnofsky performance status ≥70 (in trials using KPS, performance was translated into WHO PS for the purpose of uniform analysis); adequate bone marrow, liver, and renal function; and absence of significant comorbidities were required. Patients were required to be on a stable or decreasing dose of steroids. The 2 trials with cilengitide included only patients with a methylated (CENTRIC³) or unmethylated (CORE⁶) promoter of O⁶-DNA methylguanine-methyltransferase (MGMT). The RTOG 0525 trial⁷ mandated tumor debulking in order to determine the MGMT promoter status prior to randomization.

Treatment prescription for chemoradiation followed by TMZ was comparable for all 4 trials. The trials stratified patients by either geographic region (CENTRIC: Europe, North America, and the rest of the world); treatment center (EORTC 26981); or MGMT methylation status (RTOG 0525). The duration of the adjuvant/maintenance TMZ treatment for nonprogressing patients was defined as up to 6 cycles in the original EORTC/NCIC protocol. For the subsequent protocols, in order to allow for variability of local practice, extension of maintenance treatment up to 12 cycles was left to the investigator’s discretion. In the United States, continuation of TMZ for one year is common practice in nonprogressing patients who have tolerated 6 cycles without undue toxicity, while in most European centers, maintenance TMZ is discontinued after completion of 6 cycles.

All patients who were progression free 28 days after the start of cycle 6 of maintenance TMZ (ie, at the time a seventh cycle would usually be administered) were included in the analysis. Two populations of patients were compared: those who discontinued TMZ therapy without progression after the end of cycle 6 (6C; Group 1) and those who continued maintenance TMZ after 6 cycles until tumor progression or for up to 12 cycles (>6C; Group 2), based on local practice and at the discretion of the investigator. Additional analyses were made for a small subgroup, who continued adjuvant TMZ treatment beyond 12 cycles. Residual survival was defined as start of cycle 6 + 28 days (theoretical start of cycle 7) until death, progression, or censor at last follow-up, and computed using the Kaplan–Meier method.

The primary endpoint for this analysis was the residual overall survival (OS), with residual progression-free survival (PFS) a secondary endpoint. Exploratory subgroup analyses of the impact of MGMT promoter methylation status were included.

For multivariable analyses, the Cox proportional hazards model (with forward stepwise selection) was used and adjusted for age, performance status, resection extent, and MGMT promoter methylation status at baseline (at start of TMZ/RT→TMZ). As MGMT methylation status could not be assessed in all patients, analyses were repeated for patients with and without known MGMT status. Sensitivity analyses were done with MGMT data imputation. All analyses were performed at 5% significance.

All protocols were approved by the local ethics committees or institutional review boards and competent authorities, and patients provided written informed consent for trial participation. Trials were conducted in accordance with the Declaration of Helsinki.⁹

Results

A total of 624 patients qualified for inclusion, corresponding to 32% of the patients included in the respective treatment arms of the trials (Table 1). The time from initial diagnosis (surgery date) until end of cycle 6 was almost superimposable among the 4 studies, with a median interval of 9.1 months (Table 1); there was also no difference in the time from diagnosis until original study registration (27 days). Of these 624 patients, 333 received 6C, and 291 received >6C of adjuvant/maintenance TMZ. Twenty-four patients (4%) received more than 12 cycles of adjuvant TMZ (range 13–32 cycles; median number of cycles 17, mean 17.43).

Table 1

Studies included in pooled analysis and outcomes. Description of study schema

Study Name Ref. / Particular Characteristics	Treatment Arm	Survival	n	Interval Diagnosis to Post Cycle 6 †	# Patients (%) Progression Free Post Cycle 6	# Patients Included in the Analysis
						6 cy	>6 cy
EORTC 26981/ NCIC CE3 [ref. 1, 8]	RT→vs	HR 0.63 (95% CI: 0.52–0.75), P < .001	286		N/A	N/A	NA
	TMZ/RT→TMZ		287*	Median 9.4 mo (IQR: 9.1–9.8)	92/287 (32)	92	0
EORTC 26071-CENTRIC. Methylated tumors [3]	TMZ/RT → TMZ vs	No difference, HR 1.01	273	Median 9.1 mo (IQR: 8.9–9.3)	108/273 (40)	77	31
	Cil + TMZ/RT→TMZ		272		115/272 (42)	96	19
EMD-CORE. Unmethylated tumors [5]	TMZ/RT →TMZ vs	Randomized phase II trial. Overall comparable outcome	89	Median 9.1 mo (IQR: 8.9–9.3)	15/89 (17)	13	2
	Cil (2d/wk/) TMZ/ RT→TMZ		88		23/88 (26)	18	5
	Cil (5d/wk/) TMZ/ RT→TMZ		88		28/88 (32)	22	6
RTOG 0525/ EORTC 26052 [7]	TMZ/RT → TMZ	No difference, HR 1.03	411	Median 9.0 mo (IQR: 8.7–9.3)	115/411 (28)	5	110
	TMZ/RT → dd TMZ		420		128/420 (30)	10	118
Total number of patients enrolled			2214	Median 9.1 mo (IQR: 8.8–9.4)	624/1928 (32)	333	291

Abbreviations: Cil, cilengitide; d, day; dd, dose-dense; IQR: Q1-Q3 interquartile range.

* Only the TMZ/RT->TMZ arm was used for pooled data. Both TMZ/RT->TMZ and experimental arms were used for the other 3 trials to maximize the sample size for the analyses, since none of the experimental arms demonstrated significant benefit over TMZ/RT->TMZ.

† Interval diagnosis—post cycle 6: Time interval from initial histological diagnosis (surgery date) until inclusion in current analysis (=day 1 of cycle 6 + 28 days).

Known prognostic factors between 6C versus >6C groups were balanced except for performance status, steroid usage, and MGMT promoter methylation status (Table 2).

Table 2

Baseline characteristics of patients receiving 6 cycles (Group 1) versus more than 6 cycles (Group 2) of adjuvant temozolomide

Characteristic	Group 1 (6C) (N = 333) No (%)	Group 2 (>6C) (N = 291) No (%)	All (N = 624) No (%)
Age, y*
<50	122 (36.6)	99 (34.0)	221 (35.4)
≥50	211 (63.4)	192 (66.0)	403 (64.6)
Gender
Male	191 (57.4)	162 (55.7)	353 (56.6)
Female	142 (42.6)	129 (44.3)	271 (43.4)
WHO performance status*
0	179 (53.8)	214 (73.5)	393 (63.0)
≥1	154 (46.2)	77 (26.5)	231 (37.0)
Extent of resection*
Incomplete resection	145 (43.5)	142 (48.8)	287 (46.0)
Complete resection	186 (55.9)	149 (51.2)	335 (53.7)
Missing	2 (0.6)	0 (0.0)	2 (0.3)
Corticosteroid therapy*
No	189 (56.8)	82 (28.2)	271 (43.4)
Yes	144 (43.2)	209 (71.8)	353 (56.6)
MGMT methylation status*
Unmethylated	76 (22.8)	132 (45.4)	208 (33.3)
Methylated	203 (61.0)	139 (47.8)	342 (54.8)
Unknown	54 (16.2)	20 (6.9)	74 (11.9)

*Enrollment (baseline) values.

The MGMT promoter methylation status was more frequently unknown in the earlier trials, as this marker had only recently been established as a prognostic factor. MGMT was methylated more frequently in the 6C group compared with patients receiving >6C (203/279, 73% vs 139/271, 51%), and subgroups were analyzed separately (see below and Fig. 2A and B).

Baseline performance status was WHO PS of 0 in 74% of the >6C group, compared with 54% of the 6C cohort. In the >6C group, 72% received baseline steroids compared with only 43% in the 6C group, although the prognostic impact of baseline steroid usage in this context is unknown, and might not even be relevant. Higher use of steroids at baseline in the >6C group may be due to differences in practice in the USA compared with other parts of the world. Steroid use was not significantly associated with survival in univariate analysis and hence was not included in multivariable analyses. Overall survival was impacted by age, MGMT methylation, extent of resection, and performance status.

The details and individual outcomes of the 4 trials have been published previously and are summarized in Table 1.¹⁰ EORTC 26981 TMZ/RT→TMZ improved survival and constituted the standard arm of all subsequent trials.

In the NRG Oncology/RTOG 0525/EORTC 26052-Inter group study, dose intensification of TMZ with a dose-dense TMZ schedule failed to improve outcome; 63% (125 of 197) of the participating institutions, mainly from the USA, treated nonprogressive patients with more than 6 cycles of adjuvant TMZ and contributed a large number of the >6C group. The median follow-up for the 179 surviving patients at the time of the last database update used for this analysis (2011) was 33.7 months from registration (range: 2.0 to 53.3). Of the 833 randomized patients, 243 (29%) were progression free at the end of cycle 6, out of which 228 (94%) received more than 6 cycles of adjuvant TMZ. The majority of the 243 patients included in this study were from the US or Canada. Ninety-six percent of those patients were treated with more than 6 cycles of adjuvant TMZ.

The number of patients who were progression free after 6 cycles of TMZ was comparable between the EORTC 26981 (32%) and RTOG 0525 (29%) trials; substantially larger in CENTRIC (41%; MGMT-methylated tumors only); and as expected, lower in CORE (25%; MGMT-unmethylated tumors only), explaining the unequal distribution of MGMT promoter methylation.

Twenty-four patients (4%) in our present cohort received more than 12 cycles of adjuvant TMZ.

Ninety-one (27.3%) of the 333 patients treated with 6 cycles of adjuvant TMZ were retreated with TMZ at recurrence. Thirty-six (12.4%) of the 291 patients treated with >6 cycles of adjuvant TMZ were treated with TMZ at recurrence. Of 15 patients who stopped adjuvant TMZ treatment at 12 cycles, 12.2% were retreated with TMZ at recurrence. Thirteen percent of the patients treated with more than 12 cycles of adjuvant TMZ received retreatment with TMZ.

Median residual OS from inclusion for this analysis (day 1 of cycle 6 + 28 days) was 24.9 months (95% CI: 19.9–28.7) versus 27.0 months (95% CI: 21.54–30.9) for the 6C (n = 333) and >6C (n = 291) groups, respectively. The 2-year survival rates were 50.9% (95% CI: 44.7–56.7) versus 53.9% (95% CI: 47.6–59.7), corresponding to an adjusted HR of 0.92 (P = .52) (Fig. 1A). More importantly, no significant differences were seen in OS between the 2 cohorts when separated into subgroups by MGMT methylation status (methylated adjusted HR = 0.89 [95% CI: 0.63–1.26], P = .51; unmethylated adjusted HR = 0.85 [95% CI: 0.58–1.25], P = .41) (Fig. 2A and B). No OS benefit was seen for extending TMZ beyond 12 cycles.

Fig. 1

(A, B) Overall and progression-free residual survival for patients treated with 6C vs >6C. Survival was calculated from start of cycle 6 + 28 days. Overall survival curves of the 6C vs >6C groups demonstrated no difference. Although median PFS was the same between the groups, 2-year PFS was superior for patients receiving >6C vs 6C. Adjusted P-values are displayed.

Fig. 2

(A, B) Residual OS in MGMT methylated and unmethylated tumors. No difference in OS was seen among MGMT methylated or unmethylated patients receiving 6C vs >6C. Adjusted P-values are displayed.

Median residual PFS was comparable for the 6C versus >6C groups: 10.4 months (95% CI: 8.7–11.9) and 12.2 months (95% CI: 9.4–14.0), respectively. The PFS comparison reached statistical significance (adjusted HR = 0.80, 95% CI: 0.65–0.98; P = .03; Fig. 1B). The PFS probability at 2 years was 28.9% (95% CI: 23.5–34.5) for the >6C group versus 19.4% (95% CI: 14.8–24.5) for the 6C group. This benefit for >6C seemed to be enhanced in the MGMT-methylated group (adjusted HR = 0.65 [95% CI: 0.50–0.85], P = .0019) but was actually lost in the MGMT-unmethylated group (HR = 0.88 [95% CI: 0.64–1.21], P = .43). Analyses with MGMT data imputation provided similar results for all comparisons (data not shown).

In order to evaluate whether occult tumor progression may have influenced the decision to continue treatment beyond cycle 6, we further performed a sensitivity analysis excluding patients who progressed (n = 24, C6; n = 15, >C6, n = 9) or were censored (n = 1, >6C) within 4 weeks of the effective or theoretical start of cycle 7. The results of this sensitivity analysis did not differ from the overall results.

Discussion

In this pooled individual patient data-based analysis of 4 prospective, randomized trials we evaluated the impact of extending the duration of maintenance TMZ chemotherapy beyond the formally established 6 cycles, with 333 patients receiving 6C compared with 291 patients receiving >6C. Our analysis showed no prolongation in OS when patients continued TMZ for up to 12 adjuvant or maintenance cycles. We acknowledge the possibility that unrecognized bias could have influenced the distribution of more favorable patients in the >6C category. However, the time from diagnosis to completion of 6 cycles of adjuvant chemotherapy was almost identical among the 4 studies, as was the percentage of patients eligible for this subgroup analysis in the 2 studies without molecular enrichment. This consistency strongly suggests comparability of the study populations and solidity of the data. The residual survival for patients who remained progression free and completed at least 6 cycles of maintenance TMZ is over 2 years, accounting for a prognostically favorable subgroup of patients with an estimated median survival from diagnosis of close to 3 years.

MGMT promoter methylation has been shown to be the strongest prognostic and predictive biomarker of outcome for GBM when patients are treated with TMZ chemotherapy. Even in the more sensitive MGMT promoter methylated subgroup, we could not demonstrate a survival benefit for prolonged TMZ administration, although PFS may have been increased in this population. The numbers of patients at risk in our analysis of long-term PFS are too small to allow for definitive conclusions regarding this endpoint. Our findings are in accordance with RTOG 0525. In that trial—driven by a different type of “more is better” hypothesis—an attempt was made to overcome MGMT-mediated treatment resistance by a dose-dense TMZ regimen, that is, doubling the dose intensity of maintenance TMZ by using a continuous 21/28-day regimen rather than the standard 5/28 days. Neither PFS nor OS were improved when analyzing all patients pooled or grouped by MGMT promoter methylation status.⁷

In solid-tumor oncology, maintenance chemotherapy has been investigated repeatedly. For most tumor types, prolonged treatment with cytotoxic agents has not translated into improved outcome. In advanced colorectal cancer, intermittent chemotherapy with planned treatment interruptions and re-initiation of the same therapy at progression resulted in similar OS compared with continuous chemotherapy administration.¹¹ Only in non-small-cell lung carcinoma did maintenance chemotherapy with pemetrexed demonstrate a consistent prolongation of PFS and OS (a result not replicated for other cytotoxic agents or for bevacizumab alone).^12,13

Theoretical considerations and preclinical data suggest that extended treatment with TMZ may even be detrimental and confer resistance to tumor cells.¹⁴ Exposure to therapy with TMZ results in mutational changes in the tumor that may connote resistance to ongoing alkylating therapy. In the absence of MGMT-mediated DNA repair, inactivation of the mismatch repair (MMR) pathway is associated with dramatically increased resistance to alkylating agents based on in vitro studies.¹⁵ The acquisition of mutations in the MMR gene mutS homolog 6 (MSH6) after treatment with alkylating agent therapy has been reported in association with a hypermutator phenotype.¹⁶ Recent studies have substantiated this finding in a larger dataset of TMZ-treated GBM: 15% had acquired a mutator phenotype that was associated with MGMT-methylated GBM. Mutations of MSH6 were confined to hypermutated samples and were present in the majority of these tumors at recurrence, suggesting a functional relationship between loss of function of MSH6 and genomic hypermutation in GBM.¹⁷ Decreased expression levels of several MMR proteins including MSH2, MSH6, and PMS2 (postmeiotic segregation increased 2) have been reported in GBM at recurrence.^15,18

It is conceivable that prolonged treatment with TMZ could be counterproductive and in fact promote development of resistant and more aggressive tumor clones in patients who extend therapy beyond 6 months. A trend for PFS benefit as seen in our study may be abrogated by this process of negative selection, eventually rendering the tumor more resistant to subsequent salvage therapies.^14,19,20 A phase II study evaluating the efficacy of salvage continuous dose-intense TMZ in recurrent malignant glioma suggested that patients receiving fewer doses of adjuvant TMZ or who had a treatment-free interval benefited most from salvage TMZ therapy, compared with patients who progressed while receiving extended adjuvant TMZ beyond 6 months.²¹

There are numerous reasons to avoid prolonged chemotherapy. Although many adverse effects related to TMZ may be idiosyncratic (hepatic toxicity, aplastic anemia)²² and are usually seen early in the treatment course, there is a degree of cumulative treatment-related fatigue.^7,23 Additional toxicities such as nausea and anorexia, albeit mild to moderate, have been reported in up to 44% and 14%, respectively.²⁴ Many patients are unable to return to work or their normal routines while on active treatment with TMZ. Treatment-related fatigue in patients treated with cytotoxic chemotherapies may not be adequately captured by standard toxicity evaluations but can have a profound effect on the patient’s well-being and quality of life.²⁵ Some studies suggest an increasing risk of myelodysplasia and leukemia associated with increasing cumulative doses of TMZ.²⁶ While the cumulative toxicities during treatment with TMZ are typically not as severe as those seen with other alkylating chemotherapies, such as nitrosoureas (namely myelosuppression),²⁷ the overall bone marrow reserve may nonetheless be decreased and compromise the possibility for subsequent salvage treatments. Prolonged therapy with TMZ may chronically suppress the immune system,^28,29 which may in itself shorten survival.

There are inherent limitations to our study. It is an unplanned retrospective meta-analysis of patients inclu ded in clinical trials spanning over one decade. In our dataset the majority of US and Canadian sites extended adjuvant treatment with TMZ beyond 6 months; arbitrary physician preference or international practice pattern influenced treatment decisions.

Although the patients were included in randomized prospective trials, this analysis is a post-hoc, nonrandomized comparison subject to occult biases and confounding factors. As only approximately 30%–40% of patients remained progression free at the end of the standard 6 cycles of TMZ maintenance therapy, the number of patients potentially analyzable became relatively small despite the initial large sample size. While OS is a robust endpoint, analyses of PFS are subject to variability due to frequency and timing of imaging, interpretation of the MRI, and potential for pseudoprogression.

Progression-free survival may not be a reliable proxy for OS, and its clinical significance is questionable. While second-line treatments after initial tumor progression in our cohort are incompletely documented, we do have data showing that more than twice the percentage of patients receiving 6 cycles (27.3% vs 12.4% of the group treated with >6 cycles) received rechallenge with TMZ. This imbalance could have affected the survival outcomes in favor of the 6-cycle group. However, while theoretically a potential survival advantage could be missed due to retreatment at recurrence of those patients who initially received only 6 cycles, data from EORTC 26981 show otherwise: despite 60% of patients receiving TMZ after progression, OS was significantly improved for the group receiving TMZ as first-line and adjuvant therapy (opposed to RT only first-line, and TMZ at recurrence). Furthermore, 2 recent large trials assessing the role of bevacizumab first-line in newly diagnosed GBM showed a statistically significant PFS signal in the experimental arms, but no translation of that apparent advantage to OS.^4,30

Missing data and small numbers of molecular subgroups do not allow for firm conclusions. Still, a formal comparative randomized phase III trial is unlikely to ever be conducted, as this would require recruitment of over 2000–3000 patients and demonstrate, at best, an incremental to no impact on outcome. As such, at present, our analysis is the closest possible simile to a prospective comparative trial.

In conclusion, our data do not support the practice of extending adjuvant or maintenance temozolomide beyond 6 cycles. Increased duration of adjuvant therapy with TMZ was not correlated with increased OS in newly diagnosed patients with GBM. Patients’ quality of life; chronic treatment-associated toxicity, albeit mild; concerns regarding teratogenesis, fertility, and the need for contraception; and ultimately cost to the health care system are considerations which favor limiting the treatment duration of adjuvant TMZ beyond proven efficacy.^31–33

Individual decisions to prolong treatment beyond 6 months in patients with ongoing evidence of radiologic and clinical benefit need to be carefully weighed against toxicity; late consequences regarding the resistance of recurrent tumor and limitations for salvage therapy; immunosuppression; quality of life; and ultimately economic impact. Future research should investigate justification of continued treatment in individual “dynamic” patients actively responding to chemotherapy with TMZ after 6 cycles, for whom a continuous decrease of tumor is seen on serial imaging. Dynamic imaging (PET/metabolic or perfusion dynamic contrast enhanced MRI)^34–36 may help discern whether indeed a remaining enhancing area represents active disease. However, these methods have not been validated with surgical confirmation.

Funding

This work was supported by the EORTC Cancer Research Fund from Belgium and by grants U10CA180868 (NRG Oncology Operations) and U10CA180822 (NRG Oncology SDMC [Statistics and Data Management Center]) from the National Cancer Institute.

Conflict of interest statement. D.T.B. serves on the advisory board for Vascular Biogenics (VBL), Inc. and has received travel support from Roche and Bristol-Myers Squibb (BMS). M.M.K. receives research funding from EpicentRx, Inc. M.P.M. has served as a consultant to Varian, Agenus, and Insys; he serves on the DSMB of Monteris and has received institutional grant funding from Novocure. B.G.B. received travel support from Merck & Co (MSD) (unrelated to this manuscript); NOXXON pharmaceutical company for consulting fees; and the Falk Foundation for support of scientific presentation. M.W. has served as a paid consultant, received honoraria, or had advisory board participation for BMS, Celldex, Immunocellular Therapeutics, Isarna, Magforce, MSD, Merck & Co, Northwest Biotherapeutics, Novocure, Pfizer, Roche, Teva, and Tocagen. R.S. has served on advisory boards for Abbvie, Merck, KGaA, Merck & Co/MSD (outside the submitted work), Novartis, Novocure, Pfizer, and Roche, fees (when applicable) to institution. All other authors have no declared potential conflict of interest.

Acknowledgments

The authors wish to thank the patients and their families for their participation in these clinical trials. We also acknowledge the contributions of hundreds of physicians and local investigators contributing to patient care, data collection, and trial participation making such an analysis possible. We also extend our thanks to the collaborators and staff at the EORTC and NRG headquarters and at Merck KGaA/EMD Serono for their support and sharing of the data.

Author notes

Corresponding Authors: Deborah T. Blumenthal MD, Neuro-oncology Service / Oncology Division Tel Aviv Medical Center, 6 Weizmann Street, Tel-Aviv, Israel 64239 ([email protected]); and Roger Stupp, MD, Brain Tumor Institute & Robert H. Lurie Comprehensive Cancer Center, Northwestern University Feinberg School of Medicine, Northwestern Memorial Hospital, 676 N. St Clair Street, Suite 2210, Chicago, Illinois 60611 ([email protected]).