Neuro-Oncology Neuro-Oncology 18(3), 303– 305, 2016 doi:10.1093/neuonc/now012
TTFields: where does all the skepticism come from? Wolfgang Wick Neurology Clinic, University of Heidelberg and Clinical Cooperation Unit (CCU) Neurooncology, German Cancer Consortium (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany (W.W.) Corresponding Author: Wolfgang Wick, MD, Neurology Clinic & National Centre for Tumour Disease, University of Heidelberg, Im Neuenheimer Feld 400, D-69120 Heidelberg, Germany ([email protected]).
Tumor-treating fields (TTFields) are low-intensity, intermediatefrequency (200 kHz) alternating electric fields that are delivered locoregionally via transducer arrays. The treatment, delivered through the shaved scalp, is regarded an antimitotic treatment that interferes with cell division and organelle assembly. There are no adjustments for size, or depth of the tumor in the brain. The data from the current EF-14 trial of this therapy1 are based on an interim analysis of 315 patients with a data lock on September 5, 2014. The analysis was pre-specified and led to a premature closure of the trial for success. In fact, at closure 695 out of 700 patients had been randomized. Progression-free survival (PFS) showed an impressive hazard ratio (HR) of 0.62 with a 98.7% confidence interval (CI) of 0.43–0.89. Overall survival (OS) in the intention-to-treat group is likewise impressive with a HR of 0.74 [95%CI: 0.56–0.98]. The HRs for PFS and OS, which translate to a median improvement of about 3 months, are supported by a favorable safety profile. After a decade of antiangiogenic or targeted therapy trials that have not fulfilled their promise, the current study offers a true success story and a shift in treatment paradigm. EF-14 investigated a novel treatment concept and provided positive data from a phase III trial with a survival benefit and limited extra toxicity. Although the data, recently published in the Journal of the American Medical Association,1 have now been in the public domain for more than one year after the initial presentation at the Society for Neuro-Oncology annual conference in 2014, the run for tickets has not happened. The enthusiasm among experts, patients, and caregivers is limited; on the contrary, and there remains significant skepticism about the TTFields device. It may be valuable to understand the reasons for this. First, the trial was initiated based on a negative trial for patients with progressive glioblastoma.2 Despite the lack of benefit over standard-of-care therapy, TTFields received regulatory approval in the United States and other countries and is recognized in the National Comprehensive Cancer Network guidelines, thus far without a significant impact on the standard of practice. Second, the lack of a clear mechanism of action is a factor. Although preclinical studies provide descriptive data for the proposed mode of action at a cellular level,3 there is a lack of understanding of the mechanism of action in complex and
variably localized tissues, at multiple tissue interfaces, and in conjunction with chemoirradiation, given the limited singlemodality efficacy of TTFields. Third, the decision to perform an open-label, not shamcontrolled trial is understandable given the level of complexity for designing a sham device and the maintenance of TTFields; on the other hand, this lack of blinding combined with the challenging primary endpoint, PFS, obviously triggers a serious debate on the validity of the observed effects. Fourth, who are the trial patients? These patients constitute an unprecedented trial population in many aspects. Implantation of carmustine wafers was allowed.4 Treatment with the TTFields device was allowed after the primary endpoint, progression, was reached, although we do not know whether this was relevant or not. Patients had already experienced a favorable course of disease as the median interval between diagnosis and randomization was 3.8 months. The importance of timing is underscored by the fact that 82 of 1019 patients experienced progression in the (short?) interval between registration and randomization and were not randomized. In the EORTC trial5 coordinated by the same lead investigator, PFS was also measured from randomization but at a time close to diagnosis. At an interval of more than 3.5 months, more than 30% of patients had already progressed. Data from the Radiation Oncology Therapy Group 0525 trial, which also randomized the good-risk patients who were stable after radiotherapy, were used as a comparator. In this study, there was an almost 2-month difference in OS between registered and randomized patients.6 This time lag in the TTFields trial impacts the generalizability of the present data. As in other contemporary trials,7,8 it is not possible to decipher the risk factors for early progression. Age, resection, and MGMT status are comparable to trials in which randomization was closer to diagnosis. Thus, it will be difficult to define the group of patients that may derive benefit from present treatment or the exact starting time. Fifth, the data presented are for the groups of patients included in the interim analysis at an intention-to-treat level. Here, the remarkable HR for PFS and OS, buttressed by a favorable safety profile, translate to a median improvement of about 3 months. It is unfortunate that we do not get a sense of the maturity of the data; censored patients are indicated but no
# The Author(s) 2016. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].
303
Editorial
counts are provided for the interim analysis, nor is the analysis of the final data set shown in the supplement. It will be important for the expert community to critically compare the published data from the interim analysis with the mature data from the full dataset at a future time point. In considering the results as a whole, it is necessary to challenge the relevance of the data. The subgroup analyses show a remarkable uniformity of benefit independent of age, MGMT status, resection status, or Karnofsky performance status. This means that, unlike in the early temozolomide days when patients were convinced to take the drug despite a numerically small effect on the median PFS and OS, there is no MGMTequivalent status, or long-term benefit that can be extrapolated from the “tail of the curve,” that can be used to convince understandably skeptical patients to take on yet another burden. After ten years of contemplation, many in the field regard the term “temozolomide/MGMT status”,9 ie neuro-oncology entering precision medicine, to be the real success story, not the general standard-of-care.10 It may be difficult to accept another one-size-fits-all approach with overall modest impact on this extremely challenging disease. The very proactive interpretation of the relevance of the trial results by the study group and the company developing TTFields, while understandable, has been provocative to the community, and especially many major academic centers, not well acquainted with the device. To proclaim “A new standard of care for patients with GBM has been established”,11 implies that EF-14 investigators, but not guideline committees or the discussion in the community including health care professionals’ and patients’ adoption of a treatment modality, sets a standard. A formal definition of a standard of care is difficult. One could probably agree that “standard of care” refers to the degree of attentiveness, caution, and prudence that a reasonable person would exercise under the circumstances. This suggests, however, that “failure to meet the standard” is negligent, and the person who fails to meet the standard is liable for any damages caused by such negligence. Applied to our subject, TTFields may be an available option for all patients with newly diagnosed glioblastoma with successful completion of chemoradiotherapy and stable disease at potential initiation of the device treatment. What about elderly patients who may potentially be receiving a shorter course of chemoradiotherapy or chemo- or radiotherapy alone? What about the significant number of patients for whom progression may not be excluded at the first or second MRI after chemoradiation, or who had an interruption of temozolomide due to toxicity? Standard of care would influence the “standard arm” in future trials. This most likely will not happen because of the practical aspects that have been discussed and, most importantly, since its incorporation in the standard arm requires widespread adoption. The present TTFields trial obviously is not providing more, but also not less than an option for patients with newly diagnosed glioblastoma. Using the analogous experience in lung cancer, adjuvant chemotherapy is “a new standard of care, but not necessarily the only standard of care.”12 The uptake of TTFields will depend on the burden-benefit assessment that each patient will have to make, with the help of their multidisciplinary neuro-oncology team. There are a considerable number of patients with understandable reservations. The sentiment is that doubt on the physician’s side has a major impact on the
304
patient’s decisions. The cost of the treatment plays a role, not only for payers, but also physicians and patients interested in a responsible allocation of resources. The TTFields therapy, like no other regimen, will alter the relationship between patients and their neuro-oncology team. A company employee will see the patient much more often than the medical team. Who will make the decision whether or not to continue TTFields at progression? The technician with a relevant conflict of interest? Finally, in addition to the concerns framed above, several issues remain: † What will be the outcome, both for PFS and more importantly OS, of an analysis with .95% of events? This is especially important, since it is conceivable that trials with an early closure due to success overestimate the real efficacy. † What if the experts declare a new standard of care, but patients do not buy it? How many patients need to be informed in order to generate one EF-14 trial participant? Would that ratio be different today? According to the definition provided by the National Cancer Institute of the United States a standard of care is a “treatment that is accepted by medical experts as a proper treatment for a certain type of disease and that is widely used by healthcare professionals“ (NCI Definition, accessed January 1, 2016). † What if the indisputable higher level of care in the TTFields arm, or the lack of blinding, or the imprecise definition of the patient population generated a bias? The analogy to the lung cancer trial, in which early supportive care produced a median OS benefit of .2 months,13 seems much closer than to the negative CENTRIC trial with twice-weekly cilengitide infusions,8 since patients may really benefit from this additional care at home but not another set of intravenous drug administrations. Given the low magnitude of PFS gain in the bevacizumab trials7,14 and the controversy about the difficulty of assessment of progression, a new treatment resulting from one trial based mainly on PFS poses a challenge to all of us. Since there are relevant groups of patients that would not be able to even receive the device according to the current data, and considering the skepticism in the field, a confirmatory trial done in an academic setting in a group of elderly patients, with the yet-to-be-defined standard of care as a comparator, which included a sham device, would yield much more than just adherence to good scientific practice.
Conflict of interest interest statement. W.W. has participated in a speaker′ s bureau for MSD, has received research funding from Apogenix, Boehringer Ingelheim, Genentech Roche, MSD and Pfizer and has a consultant relationship with Genentech/ Roche and his institution received a compensation for an advisory board from Novocure.
References 1.
Stupp R, Taillibert S, Kanner AA, et al. Maintenance therapy with tumor-Treating fields plus temozolomide vs temozolomide alone for glioblastoma: a randomized clinical trial. JAMA. 2015;314(23): 2535– 2543.
Editorial
2.
Stupp R, Wong ET, Kanner AA, et al. NovoTTF-100A versus physician’s choice chemotherapy in recurrent glioblastoma: a randomised phase III trial of a novel treatment modality. Eur J Cancer. 2012;48(14):2192–2202.
3.
Kirson ED, Dbaly´ V, Tovarys F, et al. Alternating electric fields arrest cell proliferation in animal tumor models and human brain tumors. Proc Natl Acad Sci USA. 2007;104(24):10152– 10157.
4.
Westphal M, Hilt DC, Bortey E, et al. A phase 3 trial of local chemotherapy with biodegradable carmustine (BCNU) wafers (Gliadel wafers) in patients with primary malignant glioma. Neuro Oncol. 2003;5(2):79–88.
5.
Stupp R, Mason WP, van den Bent MJ, et al. European Organisation for Research and Treatment of Cancer Brain Tumor and Radiotherapy Groups; National Cancer Institute of Canada Clinical Trials Group. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005; 352(10):987 –996.
6.
Gilbert MR, Wang M, Aldape KD, et al. Dose-dense temozolomide for newly diagnosed glioblastoma: a randomized phase III clinical trial. J Clin Oncol. 2013;31(32):4085– 4091.
7.
Chinot OL, Wick W, Mason W, et al. Bevacizumab plus radiotherapy-temozolomide for newly diagnosed glioblastoma. N Engl J Med. 2014;370(8):709 –722.
8.
Stupp R, Hegi ME, Gorlia T, et al. European Organisation for Research and Treatment of Cancer (EORTC); Canadian Brain
Neuro-Oncology
Tumor Consortium; CENTRIC study team.: Cilengitide combined with standard treatment for patients with newly diagnosed glioblastoma with ethylated MGMT promoter (CENTRIC EORTC 26071-22072 study): a multicentre, randomised, open-label, phase 3 trial. Lancet Oncol. 2014;15(10):1100 –1108. 9.
Hegi ME, Diserens AC, Gorlia T, et al. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med. 2005; 352:997 – 1003.
10.
Wick W, Weller M, van den Bent M, et al. MGMT testing--the challenges for biomarker-based glioma treatment. Nat Rev Neurol. 2014;10(7):372 –385.
11. Stupp R, Wong ET, Scott CB, et al. Interim analysis of the EF-14 trial: A prospective, multicenter trial of NovoTTF-100A together with temozolomide compared to temozolomide alone in newly diagnosed glioblastoma. SNO 2014 late-breaking abstract, November 15, 2014. 12.
Diasio RB. Adjuvant chemotherapy for adenocarcinoma of the lung--is the standard of care ready for change? N Engl J Med. 2004;350(17):1777 –1779.
13.
Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med. 2010;363(8):733–742.
14. Gilbert MR, Dignam JJ, Armstrong TS, et al. A randomized trial of bevacizumab for newly diagnosed glioblastoma. N Engl J Med. 2014;370(8):699–708.
305
TTFields: where does all the skepticism come from? Wolfgang Wick Neurology Clinic, University of Heidelberg and Clinical Cooperation Unit (CCU) Neurooncology, German Cancer Consortium (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany (W.W.) Corresponding Author: Wolfgang Wick, MD, Neurology Clinic & National Centre for Tumour Disease, University of Heidelberg, Im Neuenheimer Feld 400, D-69120 Heidelberg, Germany ([email protected]).
Tumor-treating fields (TTFields) are low-intensity, intermediatefrequency (200 kHz) alternating electric fields that are delivered locoregionally via transducer arrays. The treatment, delivered through the shaved scalp, is regarded an antimitotic treatment that interferes with cell division and organelle assembly. There are no adjustments for size, or depth of the tumor in the brain. The data from the current EF-14 trial of this therapy1 are based on an interim analysis of 315 patients with a data lock on September 5, 2014. The analysis was pre-specified and led to a premature closure of the trial for success. In fact, at closure 695 out of 700 patients had been randomized. Progression-free survival (PFS) showed an impressive hazard ratio (HR) of 0.62 with a 98.7% confidence interval (CI) of 0.43–0.89. Overall survival (OS) in the intention-to-treat group is likewise impressive with a HR of 0.74 [95%CI: 0.56–0.98]. The HRs for PFS and OS, which translate to a median improvement of about 3 months, are supported by a favorable safety profile. After a decade of antiangiogenic or targeted therapy trials that have not fulfilled their promise, the current study offers a true success story and a shift in treatment paradigm. EF-14 investigated a novel treatment concept and provided positive data from a phase III trial with a survival benefit and limited extra toxicity. Although the data, recently published in the Journal of the American Medical Association,1 have now been in the public domain for more than one year after the initial presentation at the Society for Neuro-Oncology annual conference in 2014, the run for tickets has not happened. The enthusiasm among experts, patients, and caregivers is limited; on the contrary, and there remains significant skepticism about the TTFields device. It may be valuable to understand the reasons for this. First, the trial was initiated based on a negative trial for patients with progressive glioblastoma.2 Despite the lack of benefit over standard-of-care therapy, TTFields received regulatory approval in the United States and other countries and is recognized in the National Comprehensive Cancer Network guidelines, thus far without a significant impact on the standard of practice. Second, the lack of a clear mechanism of action is a factor. Although preclinical studies provide descriptive data for the proposed mode of action at a cellular level,3 there is a lack of understanding of the mechanism of action in complex and
variably localized tissues, at multiple tissue interfaces, and in conjunction with chemoirradiation, given the limited singlemodality efficacy of TTFields. Third, the decision to perform an open-label, not shamcontrolled trial is understandable given the level of complexity for designing a sham device and the maintenance of TTFields; on the other hand, this lack of blinding combined with the challenging primary endpoint, PFS, obviously triggers a serious debate on the validity of the observed effects. Fourth, who are the trial patients? These patients constitute an unprecedented trial population in many aspects. Implantation of carmustine wafers was allowed.4 Treatment with the TTFields device was allowed after the primary endpoint, progression, was reached, although we do not know whether this was relevant or not. Patients had already experienced a favorable course of disease as the median interval between diagnosis and randomization was 3.8 months. The importance of timing is underscored by the fact that 82 of 1019 patients experienced progression in the (short?) interval between registration and randomization and were not randomized. In the EORTC trial5 coordinated by the same lead investigator, PFS was also measured from randomization but at a time close to diagnosis. At an interval of more than 3.5 months, more than 30% of patients had already progressed. Data from the Radiation Oncology Therapy Group 0525 trial, which also randomized the good-risk patients who were stable after radiotherapy, were used as a comparator. In this study, there was an almost 2-month difference in OS between registered and randomized patients.6 This time lag in the TTFields trial impacts the generalizability of the present data. As in other contemporary trials,7,8 it is not possible to decipher the risk factors for early progression. Age, resection, and MGMT status are comparable to trials in which randomization was closer to diagnosis. Thus, it will be difficult to define the group of patients that may derive benefit from present treatment or the exact starting time. Fifth, the data presented are for the groups of patients included in the interim analysis at an intention-to-treat level. Here, the remarkable HR for PFS and OS, buttressed by a favorable safety profile, translate to a median improvement of about 3 months. It is unfortunate that we do not get a sense of the maturity of the data; censored patients are indicated but no
# The Author(s) 2016. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].
303
Editorial
counts are provided for the interim analysis, nor is the analysis of the final data set shown in the supplement. It will be important for the expert community to critically compare the published data from the interim analysis with the mature data from the full dataset at a future time point. In considering the results as a whole, it is necessary to challenge the relevance of the data. The subgroup analyses show a remarkable uniformity of benefit independent of age, MGMT status, resection status, or Karnofsky performance status. This means that, unlike in the early temozolomide days when patients were convinced to take the drug despite a numerically small effect on the median PFS and OS, there is no MGMTequivalent status, or long-term benefit that can be extrapolated from the “tail of the curve,” that can be used to convince understandably skeptical patients to take on yet another burden. After ten years of contemplation, many in the field regard the term “temozolomide/MGMT status”,9 ie neuro-oncology entering precision medicine, to be the real success story, not the general standard-of-care.10 It may be difficult to accept another one-size-fits-all approach with overall modest impact on this extremely challenging disease. The very proactive interpretation of the relevance of the trial results by the study group and the company developing TTFields, while understandable, has been provocative to the community, and especially many major academic centers, not well acquainted with the device. To proclaim “A new standard of care for patients with GBM has been established”,11 implies that EF-14 investigators, but not guideline committees or the discussion in the community including health care professionals’ and patients’ adoption of a treatment modality, sets a standard. A formal definition of a standard of care is difficult. One could probably agree that “standard of care” refers to the degree of attentiveness, caution, and prudence that a reasonable person would exercise under the circumstances. This suggests, however, that “failure to meet the standard” is negligent, and the person who fails to meet the standard is liable for any damages caused by such negligence. Applied to our subject, TTFields may be an available option for all patients with newly diagnosed glioblastoma with successful completion of chemoradiotherapy and stable disease at potential initiation of the device treatment. What about elderly patients who may potentially be receiving a shorter course of chemoradiotherapy or chemo- or radiotherapy alone? What about the significant number of patients for whom progression may not be excluded at the first or second MRI after chemoradiation, or who had an interruption of temozolomide due to toxicity? Standard of care would influence the “standard arm” in future trials. This most likely will not happen because of the practical aspects that have been discussed and, most importantly, since its incorporation in the standard arm requires widespread adoption. The present TTFields trial obviously is not providing more, but also not less than an option for patients with newly diagnosed glioblastoma. Using the analogous experience in lung cancer, adjuvant chemotherapy is “a new standard of care, but not necessarily the only standard of care.”12 The uptake of TTFields will depend on the burden-benefit assessment that each patient will have to make, with the help of their multidisciplinary neuro-oncology team. There are a considerable number of patients with understandable reservations. The sentiment is that doubt on the physician’s side has a major impact on the
304
patient’s decisions. The cost of the treatment plays a role, not only for payers, but also physicians and patients interested in a responsible allocation of resources. The TTFields therapy, like no other regimen, will alter the relationship between patients and their neuro-oncology team. A company employee will see the patient much more often than the medical team. Who will make the decision whether or not to continue TTFields at progression? The technician with a relevant conflict of interest? Finally, in addition to the concerns framed above, several issues remain: † What will be the outcome, both for PFS and more importantly OS, of an analysis with .95% of events? This is especially important, since it is conceivable that trials with an early closure due to success overestimate the real efficacy. † What if the experts declare a new standard of care, but patients do not buy it? How many patients need to be informed in order to generate one EF-14 trial participant? Would that ratio be different today? According to the definition provided by the National Cancer Institute of the United States a standard of care is a “treatment that is accepted by medical experts as a proper treatment for a certain type of disease and that is widely used by healthcare professionals“ (NCI Definition, accessed January 1, 2016). † What if the indisputable higher level of care in the TTFields arm, or the lack of blinding, or the imprecise definition of the patient population generated a bias? The analogy to the lung cancer trial, in which early supportive care produced a median OS benefit of .2 months,13 seems much closer than to the negative CENTRIC trial with twice-weekly cilengitide infusions,8 since patients may really benefit from this additional care at home but not another set of intravenous drug administrations. Given the low magnitude of PFS gain in the bevacizumab trials7,14 and the controversy about the difficulty of assessment of progression, a new treatment resulting from one trial based mainly on PFS poses a challenge to all of us. Since there are relevant groups of patients that would not be able to even receive the device according to the current data, and considering the skepticism in the field, a confirmatory trial done in an academic setting in a group of elderly patients, with the yet-to-be-defined standard of care as a comparator, which included a sham device, would yield much more than just adherence to good scientific practice.
Conflict of interest interest statement. W.W. has participated in a speaker′ s bureau for MSD, has received research funding from Apogenix, Boehringer Ingelheim, Genentech Roche, MSD and Pfizer and has a consultant relationship with Genentech/ Roche and his institution received a compensation for an advisory board from Novocure.
References 1.
Stupp R, Taillibert S, Kanner AA, et al. Maintenance therapy with tumor-Treating fields plus temozolomide vs temozolomide alone for glioblastoma: a randomized clinical trial. JAMA. 2015;314(23): 2535– 2543.
Editorial
2.
Stupp R, Wong ET, Kanner AA, et al. NovoTTF-100A versus physician’s choice chemotherapy in recurrent glioblastoma: a randomised phase III trial of a novel treatment modality. Eur J Cancer. 2012;48(14):2192–2202.
3.
Kirson ED, Dbaly´ V, Tovarys F, et al. Alternating electric fields arrest cell proliferation in animal tumor models and human brain tumors. Proc Natl Acad Sci USA. 2007;104(24):10152– 10157.
4.
Westphal M, Hilt DC, Bortey E, et al. A phase 3 trial of local chemotherapy with biodegradable carmustine (BCNU) wafers (Gliadel wafers) in patients with primary malignant glioma. Neuro Oncol. 2003;5(2):79–88.
5.
Stupp R, Mason WP, van den Bent MJ, et al. European Organisation for Research and Treatment of Cancer Brain Tumor and Radiotherapy Groups; National Cancer Institute of Canada Clinical Trials Group. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005; 352(10):987 –996.
6.
Gilbert MR, Wang M, Aldape KD, et al. Dose-dense temozolomide for newly diagnosed glioblastoma: a randomized phase III clinical trial. J Clin Oncol. 2013;31(32):4085– 4091.
7.
Chinot OL, Wick W, Mason W, et al. Bevacizumab plus radiotherapy-temozolomide for newly diagnosed glioblastoma. N Engl J Med. 2014;370(8):709 –722.
8.
Stupp R, Hegi ME, Gorlia T, et al. European Organisation for Research and Treatment of Cancer (EORTC); Canadian Brain
Neuro-Oncology
Tumor Consortium; CENTRIC study team.: Cilengitide combined with standard treatment for patients with newly diagnosed glioblastoma with ethylated MGMT promoter (CENTRIC EORTC 26071-22072 study): a multicentre, randomised, open-label, phase 3 trial. Lancet Oncol. 2014;15(10):1100 –1108. 9.
Hegi ME, Diserens AC, Gorlia T, et al. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med. 2005; 352:997 – 1003.
10.
Wick W, Weller M, van den Bent M, et al. MGMT testing--the challenges for biomarker-based glioma treatment. Nat Rev Neurol. 2014;10(7):372 –385.
11. Stupp R, Wong ET, Scott CB, et al. Interim analysis of the EF-14 trial: A prospective, multicenter trial of NovoTTF-100A together with temozolomide compared to temozolomide alone in newly diagnosed glioblastoma. SNO 2014 late-breaking abstract, November 15, 2014. 12.
Diasio RB. Adjuvant chemotherapy for adenocarcinoma of the lung--is the standard of care ready for change? N Engl J Med. 2004;350(17):1777 –1779.
13.
Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med. 2010;363(8):733–742.
14. Gilbert MR, Dignam JJ, Armstrong TS, et al. A randomized trial of bevacizumab for newly diagnosed glioblastoma. N Engl J Med. 2014;370(8):699–708.
305
