Chemotherapy in glioma

SUMMARY 

The treatment of glial brain tumors begins with surgery, and standard adjuvant treatment at the end of the past millennium for high-grade glioma and high-risk low-grade glioma was radiotherapy and chemotherapy was given at recurrence. However, over the past 10 years much has changed regarding the role of chemotherapy in gliomas and it is now clear that chemotherapy has a role in the treatment of almost all newly diagnosed diffuse gliomas (WHO grade II–IV). This is the result of several prospective studies that showed survival benefit after combined chemoradiotherapy with temozolomide in glioblastoma (WHO grade IV) or after procarbazine, CCNU (lomustine) and vincristine chemotherapy in diffuse low-grade (WHO grade II) and anaplastic (WHO grade III) glioma. The current standard of treatment for diffuse gliomas is described in this overview and in addition some attention is given to targeted therapies.

Practice points .

• Chemoradiotherapy with temozolomide is currently the standard of care for newly diagnosed glioblastoma.

• Temozolomide chemotherapy is more effective than radiotherapy alone in elderly glioblastoma patients with a methylated MGMT promoter.

• Treatment with lomustine or retreatment with temozolomide can be considered in recurrent glioblastoma.

• Bevacizumab combined with lomustine may play a role in recurrent glioblastoma.

• Other ‘targeted’ therapies have shown poor results in diffuse glioma until now.

• Adjuvant PCV (procarbazine, CCNU [lomustine] and vincristine) chemotherapy is currently the standard of care in newly diagnosed anaplastic oligodendroglioma with combined loss of 1p/19q.

• Adjuvant PCV clearly improves survival in newly diagnosed high-risk low-grade glioma.

• Chemotherapy is the standard for recurrent low-grade WHO grade II and high-grade gliomas grade III.

• Temozolomide is commonly used instead of nitrosoureas containing regimen (e.g., PCV) due to better tolerance.

• It is unlikely that outcome in glioma can be further enhanced with the currently available cytotoxic drugs and new (targeted) drugs are needed.

The primary brain tumors include astrocytomas, oligodendrogliomas and mixed oligoastrocytomas (Figure 1). These tumors arise from the supporting tissues of the brain, the glia, and are called glioma. Besides this classification into different types, these tumors are graded according to the presence of anaplastic features in low-grade (or WHO grade II) and high-grade tumors (either WHO grade III or anaplastic and WHO grade IV or glioblastoma). Grade I gliomas are well-demarcated gliomas, and if completely resectable can be cured. The grade II–IV gliomas are among the so-called diffuse gliomas for which a curative treatment is not available. Over the past 10 years, much has changed regarding the role of chemotherapy in glial brain tumors. The purpose of this overview is to describe the current standard of treatment for gliomas, and in addition some attention is given to targeted therapies, for example, angiogenesis inhibitors. Currently, there is a lot of attention for this category, but the role for this class of agents in diffuse gliomas has by no means been defined.

Figure 1. . Classification of glioma in types and WHO grades.

Pathophysiology

Most brain tumors occur sporadically and no obvious environmental causes of brain tumors are known. Occasionally, gliomas arise as a part of rare genetic syndromes (e.g., neurofibromatosis type 1, Li-Fraumeni syndrome and Turcot syndrome type I) [1]. The various types of diffuse glioma are characterized by different molecular abnormalities, such as loss of chromosome 10, EGF receptor (EGFR) amplification, PTEN mutations and TERT mutations in glioblastoma, combined 1p/19q loss and ATRX mutations in oligodendroglial tumors, TP53 mutations and TERT mutations in grade II and III astrocytic tumors and mutations in isocitrate dehydrogenase (IDH) genes in grade II and III diffuse glioma and secondary glioblastoma (see Figure 2) [2–4]. Epigenetic factors such as methylation of the MGMT gene and other genes are involved in the pathogenesis, but their role is far from understood.

Figure 2. . Possible genetic pathways in glioma.

Glioblastomas (WHO grade IV) can be differentiated into de novo glioblastomas (primary) or secondary glioblastomas, which originate from low-grade astrocytomas (WHO grade II) directly or via malignant transformation from anaplastic astrocytomas (WHO grade III). It has been demonstrated that the two glioblastoma pathways show different genetic alterations. Mutations of IDH1/2 are almost exclusively present in WHO grade II and III gliomas and are early events occurring before codeletion of chromosomes 1p and 19q in oligodendrogliomas or TP53 mutation in astrocytomas.

CIMP: CpG island hypermethylated phenotype.

Diagnosis & prognosis of primary brain tumors

The MRI scan is sensitive for the diagnosis of a brain tumor, but it is nonspecific for the type of tumor. The diagnosis of glial tumors is therefore based on the histology of the tumor after biopsy or resection. The histological distinction between different types of glial tumors, such as astrocytomas, oligodendroglial tumors and mixed oligoastrocytoma (with characteristics of both oligodendrogliomas and astrocytomas) is difficult and is subject to ‘interobserver variation’ [5]. Specific markers for these tumors do not exist. The hallmark of glioblastoma is the presence of endothelial proliferation and/or necrosis in an astrocytic tumor. The prognosis of brain tumors depends on tumor grade and type: high-grade tumors and astrocytic tumors have a worse prognosis than low-grade tumors and oligodendrogliomas. Also, genetic defects have an important influence on the prognosis. For example, the presence of combined 1p/19q loss and/or IDH1 mutations are correlated with a favorable prognosis. Other non-treatment-related prognostic factors are the age of the patient and the clinical condition: the prognosis is worse in older patients in a poor condition (usually expressed as performance status). The median survival of the most aggressive glioma, the glioblastoma, in patients treated with combined chemoradiotherapy, is approximately 15 months [6]. For grade III tumors without 1p/19q loss, the median survival is 2–3 years, and for anaplastic oligodendrogliomas with combined 1p/19q loss, the median survival is more than 12–14 years. The prognosis of patients with low-grade glioma varies in most series between 5 and 15 years.

Outcome assessment in glioma studies

Response rate, progression-free survival (PFS) and survival are used for the assessment of outcome in studies on chemotherapy in gliomas. The objective response or response rate (ORR) is the percentage of patients with a complete or partial response, usually measured using MRI of the brain in conjunction with clinical assessment and corticosteroid dose. In the USA, conditional approval may be granted based on response rate only, since this is expected to reflect the activity of the investigational agent. In such studies as a rule a confirmatory scan made 4 weeks after the first scan documenting the response is required.

PFS is the time between the start of treatment (or randomization) until progression or death. PFS is particularly used in Phase II studies on recurrent glioblastoma as the primary end point in which it is usually expressed as the percentage of patients that are alive and do not have progression after 6 or 12 months (6mo-PFS or 12 moPFS, respectively). However, PFS is not accepted by US regulators as an end point in uncontrolled trials, as PFS may be subject to prognostic factors and scanning intervals and does not only reflect treatment activity. The use of PFS as the primary end point in randomized Phase III trials on newly diagnosed glioblastomas has become controversial since the Phase III trials on bevacizumab and an increase in PFS should at least be supported by a clear survival signal [7,8].

Both the ORR and PFS are largely dependent on the MRI findings, but changes in the enhancing lesion on an MRI or CT scan do not necessarily reflect changes in tumor activity. The phenomenon of pseudoprogression after chemoradiotherapy in glioblastoma patients is now well established [9,10]. Similarly, pseudoresponses occur after treatment with VEGF receptor signaling pathway inhibitors [11]. Both these phenomena confuse the assessment of ORR and PFS in clinical trials [12].

Because of the above, alternative end points and response criteria have been developed by an international working group (Response Assessment in Neuro-Oncology or RANO group) [13,14].

Survival or overall survival (OS) is defined by the time between the start of treatment until death, in trials usually measured from the time of randomization. OS is not influenced by imaging. In oncology, OS is usually seen as the golden end point, although salvage treatments given at progression may have an influence on this end point.

Treatment

The treatment begins with surgery, to resect as much of the tumor as safely possible. The surgery of brain tumors performs three functions: obtaining a histological diagnosis, improving the condition of the patient by a rapid reduction in tumor volume and finally improving the prognosis. If no meaningful resection is possible, a biopsy must be performed to obtain tissue for diagnosis. Although no randomized studies exist showing a survival benefit from extensive resection, uncontrolled studies have suggested that more extensive resection contributes to survival [15]. Safe extensive resections are more often possible with advanced surgical and imaging techniques.

Radiotherapy was the first treatment proven to be effective for patients with high-grade gliomas in randomized trials. Postoperative radiotherapy of 60 Gy in 30–33 fractions of 1.8–2.0 Gy improved the median OS from 3–5 months to 8–10 months for glioblastoma (see Table 1) [16,17]. A higher radiotherapy dose, either with conventional techniques or high-dose techniques (i.e., interstitial brachytherapy or stereotactic radiotherapy), has not been proved to be more successful. Many centers give a short radiotherapy schedule (e.g., 14 fractions of 3 Gy; category 1C) to older patients with a poor prognosis (aged over 65–70 years, moderate clinical condition, only a biopsy) [18]. In very poor prognosis patients (e.g., older patients with a poor performance status), when it is assumed that the burden of treatment does not outweigh the very limited survival benefit, a palliative supportive strategy is often followed.

Table 1. . Early studies in WHO grade III and IV gliomas on various postoperative adjuvant treatments.

Therapy	Median survival (months)	1-year survival (%)	2-year survival (%)	Ref.
Supportive care	3	3	0	[16]
BCNU	4	12	0
RT	8	24	1
RT and BCNU	7.5	32	5
RT	–	40	15	[19]
RT and chemotherapy	–	46	20
RT	11.6	49.6	–	[20]
RT + BCNU wafer	13.9	59.2	–

BCNU: Carmustine; RT: Radiotherapy.

Early studies investigating the role of adjuvant chemotherapy (mostly nitrosoureas) in newly diagnosed high-grade gliomas (WHO grade III and IV) showed at best a small and statistically insignificant increase in survival from adjuvant chemotherapy (see Table 1) [16–17,19,21–24].

Consequently, in most areas in Europe adjuvant chemotherapy was not considered to be part of the standard treatment of newly diagnosed high-grade glioma. In Europe, until quite recently, chemotherapy was mainly used for recurrent high-grade (WHO grade III and IV) tumors after radiotherapy in which case temozolomide or a combination of procarbazine, CCNU (lomustine) and vincristine (PCV) chemotherapy was used. Recent results of several trials are discussed below, showing an early role for classical cytotoxic chemotherapy in glioma.

In recent years, inhibitors of blood vessel formation (angiogenesis inhibitors) have been extensively evaluated in glioblastoma, in particular, inhibitors of the VEGF signaling pathway. VEGF is expressed by hypoxic tumor cells and released into the bloodstream, leading to angiogenesis in the tumor. Especially, bevacizumab (a humanized monoclonal antibody against circulating VEGF) and cediranib (receptor tyrosine kinase inhibitor of VEGF receptor) have been extensively investigated.

Temozolomide

Temozolomide is an alkylating cytostatic, which has the same active metabolite as dacarbazine (DTIC). It is a relatively small molecule (194 Da) and it penetrates the blood–brain barrier easily due to its lipophilic character. Unlike dacarbazine, temozolomide does not require hepatic metabolism: a spontaneous conversion to the active ingredient MTIC finds place at a physiological pH. After oral administration, the bioavailability is almost 100%, minimally (10%) affected by taking food [25]. Elimination occurs in the liver and excretion is renal. Temozolomide is registered for use in gliomas. The standard schedule of temozolomide when used as monotherapy is 150–200 mg/m² on days 1–5 every 4 weeks. There are many alternative ‘dose dense’ regimens in use, resulting in a twofold greater dose intensity.

Temozolomide is generally well tolerated. Side effects are mainly nausea and vomiting (well avoidable with 5HT3 antagonists), bone marrow depression (in particular leukopenia and thrombocytopenia; nadir after 21–28 days), lymphopenia (with a decrease in the CD4⁺ population) and hepatotoxicity. Sporadically allergic skin reactions are seen with the use of temozolomide. Pregnancy and lactation are contraindications. In the continuous dosing schedules a relative lymphopenia occurs, with a risk of opportunistic infections, especially Pneumocystis carinii (PCP) infections. PCP prophylaxis, for example cotrimoxazole, is indicated in these dose dense schedules.

Lomustine

Due to their good penetration across the blood–brain barrier, the nitrosoureas (especially CCNU [lomustine], BCNU [carmustine], ACNU [nimustine] and fotemustine) have been investigated early on for their efficacy in brain tumors [17]. Lomustine is also used as part of the PCV combination regimen. PCV was widely used in the 1990s, especially in oligodendroglial tumors. The PCV schedule is currently used less frequently due to the more favorable side effect profile and easier dosing schedule of temozolomide.

Recently, a number of studies in recurrent glioblastoma used single-agent lomustine as a control arm and proved that lomustine is at least as effective as temozolomide in that setting. This has led to widespread use in recurrent glioblastoma patients. For single-agent use, the recommended dose is 110–130 mg/m² with a maximum of 200 mg, after combined chemoradiotherapy with temozolomide the maximum recommended dose of lomustine is 110 mg/m². One of the main side effects of lomustine is cumulative bone marrow suppression, with a relatively late nadir occurring at 4–6 weeks. Therefore, nitrosourea are given in 6–8-week cycles. Other side effects are nausea, vomiting and hepatotoxicity. The use of prophylactic 5HT3 antagonist is recommended with lomustine. The 6-weekly PCV combination regimen consists of lomustine 110 mg/m² on day 1, procarbazine (60 mg/m² on days 8–21) and intravenous vincristine (1.4 mg/m², maximum 2 mg on days 8 and 29). The added value of vincristine in this schedule is questionable, and other combination schedules have been used. Other side effects of the PCV chemotherapy schedule are the vincristine-associated peripheral neurotoxicity, loss of appetite, weight loss and malaise symptoms, including fatigue. These side effects seem largely due to procarbazine. Furthermore, the myelosuppression of the PCV schedule is more prominent compared with that of lomustine alone. Pregnancy and lactation are contraindications.

Carmustine wafers

BCNU-containing wafers are registered in Europe for use in newly diagnosed high-grade gliomas and in recurrent glioblastoma [20,26]. However, the possible survival benefit is limited and not statistically significant in newly diagnosed glioblastoma.

Mechanism of action

The nitrosoureas cause in particular chloroethyl-adducts at the O⁶ position of guanine. This results in N¹-guanine, N³-cytosine DNA interstrand cross-links which are cytotoxic. Temozolomide causes single-and double-stranded DNA breaks by adding methyl groups to N⁷ guanine (70% of total number of adducts), N³ adenine (9%), and O⁶ guanine (5%). The cytotoxic effects of temozolomide are mainly attributed to the O⁶-methylguanine adducts. The DNA repair protein O⁶-methylguanine-DNA methyltransferase (MGMT) removes both methyl and 2-chloroethyl adducts from the O⁶ position of guanine, and is an important mechanism of resistance against these agents [27]. Removing the MGMT protein, for example, from O⁶-benzylguanine leads to a higher cytotoxicity of nitrosourea and temozolomide. Expression of the MGMT gene is affected by epigenetic changes such as DNA methylation of the promoter gene, thus inhibiting the expression of MGMT. The presence of MGMT promoter methylation results in an increased effectiveness of temozolomide chemotherapy in glioma and perhaps also of PCV chemotherapy [28,29]. Once O⁶-methylguanine adducts are present, an intact mismatch repair (MMR) system is necessary for the induction of apoptosis. Temozolomide is not effective in cells with MMR deficiency [30]. A defect in the MMR system results in microsatellite instability and tolerance to O⁶-methylguanine DNA adduct mismatch. More than 80% of the lesions induced by temozolomide are N-methylated bases, recognized by DNA glycosylases and not by MGMT. Therefore, the resistance to temozolomide is also determined by the base excision repair (BER) system. The enzyme PARP, a nuclear enzyme that recognizes both double- and single-stranded DNA breaks and plays a central role in the activity of the BER system and the removal of methylated N³ and N⁷ adducts. Possibly, PARP inhibitors may disrupt temozolomide resistance by blocking BER through which N³- and N⁷-methyl adducts are still becoming cytotoxic. The combination of temozolomide and PARP inhibitors is therefore potentially interesting.

Chemotherapy in diffuse gliomas

• Newly diagnosed glioblastoma

The European Organization for Research and Treatment of Cancer (EORTC) 26981/NCIC CE3 randomized Phase III study in 563 glioblastoma patients showed that radiotherapy combined with temozolomide chemotherapy (60 Gy radiotherapy in 30 daily fractions of 2 Gy in combination with daily 75 mg/m² temozolomide followed by adjuvant six cycles of temozolomide at a dose of 150–200 mg/m² on days 1–5 every 4 weeks) significantly improves survival compared with treatment with radiotherapy alone (see Table 2) [6]. A small randomized Phase II study showed a similar result [31]. The long-term follow-up of the EORTC study confirmed the results of the initial publication (hazard ratio [HR] for death with chemotherapy: 0.6; 95% CI: 0.5–0.7]; see Table 2) [32]. Combined chemoradiotherapy increases 2-year survival to 27 versus 11% after radiotherapy alone. The 5-year survival was 10% in the combination arm versus 2% in the radiotherapy only arm. With these results chemoradiation with temozolomide became the world-wide accepted standard treatment for glioblastoma patients.

Table 2. . Survival in months measured from the date of randomization in EORTC study 26981 of the combined chemoradiotherapy and temozolomide in glioblastoma .

Therapy	n	Survival
		Median, months (95% CI)	2-year, % (95% CI)	3-year, % (95% CI)	5-year, % (95% CI)
Radiotherapy	286	12.1 (11.2–13.0)	10.9 (7.6–14.8)	4.4 (2.4–7.2)	1.9 (0.6–4.4)
Chemoradiotherapy	287	14.6 (13.2–16.8)	27.2 (22.2–32.5)	16.0 (12.0–20.6)	9.8 (6.4–14.0)

The methylation status of the MGMT gene promoter is of prognostic significance for survival in the EORTC 26981 study and there is a suggestion of predictive significance but the study was not powered to show this. In tumors without methylation of the MGMT gene promoter the 2-year survival was 14.8% whereas it was 49% in the MGMT-methylated tumors (and 24% after radiotherapy alone) [33]. Predictive value of MGMT promoter methylation status was found in two studies on elderly glioblastoma patients (see below). The survival benefit from the combined chemoradiotherapy was less in patients with a poorer prognosis (only biopsy, older patients, poorer performance status) [34].

• Newly diagnosed glioblastoma in elderly patients

Elderly patients in a poor clinical condition are often treated with a short radiotherapy schedule [18,35]. The main consideration for this choice is the decreased survival in this group compared with younger patients. Subgroup analyses of EORTC 26981 show that even in elderly patients (>70 years) with favorable prognostic factors the median survival is more than 12 months. Therefore, a longer radiotherapy schedule, combined with chemotherapy, can be considered in patients over 60 with favorable prognostic factors (macroscopic total resection and with a good performance status [KPS >70, MMSE >27]). Two trials have investigated treatment with temozolomide alone in an elderly population. The Nordic trial showed that a long-schedule radiotherapy leads to shorter survival compared with a short radiotherapy schedule or temozolomide alone [36]. In both the Nordic and the NOA-08 trials a better survival was found after temozolomide monotherapy compared with a short-schedule radiotherapy in patients with a methylated MGMT gene promoter, whereas MGMT promoter methylation status did not impact outcome to radiotherapy [36,37]. Therefore, temozolomide can be considered in elderly patients with MGMT promoter methylation instead of a short radiotherapy schedule. However, neither the Nordic nor the NOA-08 trial investigated the combination of a short-schedule radiotherapy and temozolomide chemotherapy. Such a trial was conducted by the EORTC/NCI (ClinicalTrials.gov [38] identifier: NCT00482677). In this trial on elderly glioblastoma patients hypofractionated radiotherapy is compared with hypofractionated radiotherapy with temozolomide. Results are expected by the end of 2015.

• Recurrent glioblastoma

In Europe, a randomized Phase II study in recurrent glioblastoma led to the registration of temozolomide. In this study, a significant improvement in 6mo-PFS was found after treatment with temozolomide compared with the procarbazine-treated control group (see Table 3; HR: 1.47; 95% CI: 1.11–1.95) [39]. Although the median survival in the temozolomide arm was 1.5 months longer, this difference was not significant. A single-arm study with temozolomide showed similar results [40]. The 6mo-PFS of 19–21% also shows that temozolomide chemotherapy in this situation is only marginally effective, with an ORR in glioblastoma of 5–10%. Although most relapsed glioblastoma patients will have received prior treatment with temozolomide, a Canadian study has shown that temozolomide may still be effective for recurrent glioblastoma with a temozolomide therapy-free interval of >3 months [41].

Table 3. . Studies with chemotherapy in patients with recurrent glioblastoma .

Chemotherapy	n	6mo-PFS	Median PFS (months)	Median OS (months)	Ref.
Temozolomide	112	21%	3	–	[39]
Procarbazine	113	8%	2	–
Temozolomide	138	19%	2.1	–	[40]
Lomustine	84	19%	1.6	7.1	[42]
Lomustine	65	25%	2.7	9.8	[43]
Lomustine/cediranib	129	35%	4	9.4
Cediranib	131	16%	3	8.0
PCV	224	–	3.6	6.7	[44]
Temozolomide	223	–	4.7	7.2
PCV	63	29%	3	7.5	[45]
Bevacizumab	85	43%	4.2	9.2	[46]
Bevacizumab/irinotecan	82	50%	5.6	–

6mo-PFS: Percentage of patients free of progression 6 months after the start of the chemotherapy; OS: Overall survival; PCV: Procarbazine, CCNU (lomustine) and vincristine; PFS: Progression-free survival.

Recent Phase III studies using lomustine as a control arm showed that nitrosoureas may be a useful treatment alternative in recurrent glioblastoma, with 20–30% 6mo-PFS (see Table 3) [42,43]. A large British randomized Phase III study in recurrent glioblastoma patients after radiotherapy alone showed no difference between temozolomide and the PCV regimen (HR: 0.89; 95% CI: 0.73–1.08) [44].

• Dose-intensified temozolomide schedules

Continuous or ‘dose-dense’ temozolomide dosing regimens were based on theoretical considerations to overcome MGMT-dependent resistance to temozolomide [47]. Some uncontrolled studies suggested relatively favorable results with these intensified regimens for recurrent glioblastoma [48]. Many studies, however, failed to show such promising results [49–51]. In line with the latter studies was a British randomized Phase III study in chemotherapy naive patients with a recurrent glioblastoma. This study showed a better survival of the standard 1–5 days every 4 weeks temozolomide schedule compared with temozolomide given in a 3-weeks-on/1-week-off dose intensified regimen (HR: 1.32; 95% CI: 0.99–1.75) [44]. Furthermore, a large international RTOG lead study showed no survival benefit of a dose-intensified adjuvant schedule in newly diagnosed glioblastoma after radiotherapy, regardless of the MGMT promoter status [52]. With the currently available data it can be concluded that there is no proven efficacy for dose-intensive temozolomide regimens, and that retreatment with temozolomide can be considered in carefully selected patients.

• Newly diagnosed grade II & III gliomas

Anaplastic tumors including oligodendrogliomas

In the late eighties of the past century the first reports emerged that recurrent anaplastic oligodendrogliomas were sensitive to PCV chemotherapy with response rates in the 55–65% range and a median duration of response of 12–18 months [53–55]. In particular, tumors with combined 1p/19q loss were very chemotherapy responsive, with 90–100% of these patients showing a response to the PCV regimen or temozolomide [56,57]. These studies in recurrent tumors were the reason to investigate whether outcome would improve if the chemotherapy was given early in the treatment of these patients. Two open randomized controlled Phase III trials (EORTC: 368 patients, RTOG: 289 patients) investigated the value of adjuvant PCV immediately before or after radiotherapy in newly diagnosed anaplastic oligodendroglial tumors compared with radiotherapy only and chemotherapy for progression [58,59]. Both had survival measured from the time of randomization as their primary end point. At the time of their first report both studies did not show an overall survival benefit from early adjuvant PCV chemotherapy although both studies showed a significant improvement in PFS after adjuvant chemotherapy [58,59]. However, the long-term results (median follow-up >11 years) of both studies showed survival benefit in anaplastic oligodendroglial tumors with combined 1p/19q loss after adjuvant PCV chemotherapy (see Table 4) despite crossover treatment with chemotherapy at the time of progression in 70% of patients randomized to radiotherapy only [60,61]. Both studies have shown that the presence of combined 1p/19q loss in oligodendrogliomas is of important prognostic significance and identifies patients with more benefit from adjuvant PCV chemotherapy. The median survival is over 10–14 years in patients with tumors with combined 1p/19q loss as opposed to 2–3 years for tumors without combined 1p/19q loss. Correlative side studies conducted within the context of these clinical trials have suggest that the CpG island hypermethylated phenotype (CIMP), a methylated MGMT gene promoter and mutations of IDH in anaplastic oligodendroglioma can also be used to identify patients that may benefit from adjuvant PCV chemotherapy [29,62].

Table 4. . Median overall survival (months) and median progression-free survival (months) in relation to the 1p/19q status.

Chromosomal status	Median OS, months (95% CI)		Median PFS, months (95% CI)
	RT/PCV	RT	RT/PCV	RT
Combined 1p/19q loss
EORTC	NR	111.8 (75.7–134.3)	156.8 (68.1–NR)	49.9 (27.8–101.8)
RTOG	176.4	87.6	100.8	34.8
No 1p/19q loss
EORTC	25 (18–37)	21 (18–29)	14.8 (9.9–21.1)	8.7 (7.1–11.7)
RTOG	31.2	32.4	14.4	12.0

EORTC study 26951 and RTOG 9402 study with (neo)adjuvant PCV chemotherapy in anaplastic oligodendroglial tumors [60,61].

EORTC: European Organization of Research and Treatment of Cancer; NR: Not reached; OS: Overall survival; PCV: Procarbazine, CCNU (lomustine) and vincristine; PFS:Progression-free survival; RT: Radiotherapy; RTOG: Radiotherapy and Oncology Group.

A randomized German study in 318 patients with grade III gliomas compared initial treatment with radiotherapy with initial treatment with chemotherapy, and randomized patients in the chemotherapy arm between PCV and temozolomide [63]. While previous studies consistently showed a better survival when high-grade tumors were treated with radiotherapy compared with chemotherapy, this first report of the study with still rather immature survival data showed no difference in PFS between patients treated with one line of chemotherapy followed at progression by radiotherapy and patients treated with radiotherapy followed at progression by one line of chemotherapy. This shows that in grade III tumors there is no difference whether the treatment is started with chemotherapy or radiotherapy, as long as patients are being surveyed carefully and further treatment is given at the time of progression. This study and the EORTC study 26951 in anaplastic oligodendrogliomas revealed that grade III tumors have a hypermethylation of the MGMT gene promoter in 70–80% of the cases, and that this is also a favorable prognostic factor after radiotherapy treatment. This was subsequently found to be related to hypermethylation of CpG islands (CIMP) in IDH-mutated tumors. In these tumors, methylation is induced by metabolic alterations that are the consequence of the altered substrate affinity of the IDH1 mutation product resulting in an increased 2HG glutarate production [64–67]. This shows that the clinical significance of MGMT promoter methylation is determined by the molecular background of the tumor. An ongoing randomized trial (CATNON) investigates the value of combined chemoradiotherapy with temozolomide in grade III tumors without combined 1p/19q loss (anaplastic astrocytomas).

Low-grade gliomas

The role of chemotherapy in newly diagnosed low-grade gliomas is slowly being clarified. Uncontrolled studies with up-front temozolomide and PCV showed favorable results, especially in tumors with combined 1p/19q loss. In a study of 149 patients treated with up-front temozolomide for a low-grade glioma, the median time to progression was 28 months and significantly longer in the group with combined 1p/19q loss [68]. Similar observations have been made with PCV [69]. A first and still early analysis of the EORTC study 22033 presented at the ASCO meeting in 2013 suggests that PFS does not differ between patients with 1p loss receiving upfront temozolomide versus patients receiving radiotherapy, but radiotherapy may provide a superior PFS in patients without 1p loss (p = 0.06) [70]. The median OS was not yet reached in that study.

An updated report from an American randomized study (RTOG 9802) in 251 patients with newly diagnosed low-grade gliomas with a relatively unfavorable prognostic profile (less than gross total resection and/or over 40 years of age) showed that adjuvant PCV chemotherapy increased both PFS and OS [71]. With 55% of the patients having died, the median survival was 7.8 years after radiotherapy alone and 13.3 years after radiotherapy followed by PCV (HR: 0.59; p = 0.002) [72]. This implies that radiotherapy followed by adjuvant chemotherapy should now perhaps be standard therapy in high-risk low-grade glioma. Unfortunately, molecular data from RTOG 9802 are not available yet and it is not clear whether all molecular subgroups benefit equally from combined treatment. Another unsolved question is whether temozolomide will provide the same OS advantage as PCV. Although the tolerability of temozolomide is better than that of PCV, no comparable trials with temozolomide have been performed. This suggest that some fundamental questions on how to best select patients for adjuvant PCV chemotherapy in this population may remain unanswered.

Another important and unanswered question is whether it is safe to use only up-front chemotherapy in low-grade glioma patients with a large chemotherapy sensitive tumor (e.g., 1p/19q loss), in which is a large radiation field is required. Such a policy is used by some to delay radiotherapy as this has been associated with cognitive deficits [73,74].

• Recurrent grade II & III gliomas

Recurrent grade II and III tumors and in particular the anaplastic oligodendroglioma with 1p/19q loss are significantly more sensitive to chemotherapy than the glioblastoma (see Table 5). In oligodendrogliomas with combined loss of 1p/19q a response is seen in 90–100% of the tumors, with a median response duration of 1–2 years. However, even in these tumors the response to a second line of chemotherapy is limited, with an ORR in 20–25% of cases and a 6-month PFS of approximately 50% [75,76]. Both temozolomide and nitrosoureas or PCV apply as the standard treatment in this situation.

Table 5. . Studies with first-line temozolomide or procarbazine, lomustine and vincristine chemotherapy in patients with recurrent grade II and III gliomas.

Histology	n	Drugs	Response (%)	6mo-PFS (%)	12mo-PFS (%)	mOS (months)	Ref.
AII	70	TMZ	47	63	–	13	[77]
AA/AOA	65	TMZ	43	50	–	11.5	[78]
AOD/AOA	67	TMZ	46	–	50	31	[79]
AOD/AOA	38	TMZ	54	71	40	NR	[55]
AO/AOA	52	PCV	63	–	50	20	[54]
AOD	37	PCV	59	72	52	30.7	[80]

6mo-PFS: Progression-free survival at 6 months; 12mo-PFS: Progression-free survival at 12 months; AA: Anaplastic (grade III) astrocytoma; AII: Low-grade (grade II) astrocytoma; AOA: Anaplastic oligoastrocytoma; AOD: Anaplastic oligodendroglioma; mOS: Median overall survival; PCV: Procarbazine, CCNU (lomustine) and vincristine; Response: Percentage of patients with an objective response (partial and complete response); TMZ: Temozolomide.

Angiogenesis inhibitors, integrin inhibitors & other targeted agents

In recent years there is much interest in treatment of glioblastoma with substances inhibiting the signal transduction by the VEGF system, either by catching circulating VEGF (such as bevacizumab, a humanized monoclonal antibody against circulating VEGF) or by blocking the receptor (such as cediranib, an orally tyrosine kinase inhibitor of VEGF receptors 2 and 3). Uncontrolled Phase II studies showed high response rates and 6mo-PFS with bevacizumab (see Table 3) [46,81]. Based on the randomized Phase II BRAIN trial (bevacizumab vs bevacizumab with irinotecan), the US FDA registered bevacizumab for use in recurrent glioblastoma (‘conditional approval’) in the USA. The lack of a control arm without bevacizumab was the reason for the European Medicines Agency to reject the registration of bevacizumab. Nonetheless, bevacizumab is currently used in many European countries for this indication.

An unsolved issue of the BRAIN trial is the use of imaging end points: the percentage of patients showing an objective response and/or 6mo-PFS [82]. Inhibition of the VEGF signaling system results also in a normalization of the abnormal permeability of the tumor blood vessels, reducing contrast enhancement on MRI scans even in the absence of a true tumor response [83]. The occurrence of these ‘pseudoresponses’ makes the interpretation of the MRI scans troublesome: it is questionable whether any reduction of contrast uptake on MRI scans in anti-VEGF agents treated glioblastoma patients indeed reflects a real tumor reduction or only a decrease in contrast uptake and edema. Indeed, a large placebo-controlled Phase III study (REGAL trial) in 325 patients showed no survival benefit of cediranib alone or in combination with lomustine compared with only lomustine single agent in recurrent glioblastoma, despite a >20% response rate in an uncontrolled Phase II study [43,84]. While the PFS in this trial was slightly better in patients treated with cediranib, this did not translate into a survival benefit. Though initial reports suggested that inhibition of tumor angiogenesis might induce an infiltrative growth pattern by co-opting existing blood vessels, some recent studies failed to find evidence for this [85].

Two large Phase III trials (AVAglio trial and RTOG 0825) on bevacizumab in newly diagnosed glioblastoma have failed to show an improvement of OS despite a major improvement of PFS [7,8]. Still, a longer PFS could translate into an improved quality of survival, as progression is often accompanied by neurological deterioration. Therefore, in both studies a quality of life assessment was included as a mandatory secondary end point. Surprisingly, the AVAglio trial showed maintenance of the quality of life in the bevacizumab-treated patients prior to progression, whereas the RTOG 0825 trial showed a decreased quality of life despite using similar questionnaires. The present results do not justify the use of bevacizumab in newly diagnosed glioblastoma, and the role of bevacizumab in recurrent glioblastoma is still unclear. The Dutch-randomized Phase II BELOB trial is the only study with a control arm without bevacizumab in recurrent glioblastoma [86]. This trial also failed to show a survival benefit of bevacizumab single agent, although the results within the combination of lomustine and bevacizumab with an OS at 9 months of 59% met the prespecified criterion for further Phase III studies. The ongoing Phase III EORTC trial 26101 in recurrent glioblastoma is designed to demonstrate whether there is indeed survival benefit of the combination bevacizumab/lomustine compared with lomustine alone (ClinicalTrials.gov [38] identifier: NCT01290939).

A near endless list of compounds aiming at established molecular aberrations (e.g., PI3 kinase inhibitors, mTOR antagonists, CDK4 and 6 antagonists, etc.) in particular glioblastoma but also in other diffuse glioma has been investigated. Unfortunately, despite all these efforts, none of these have improved outcome (e.g., CENTRIC, Erlotinib) [87,88].

Conclusion & future perspective

In the last decade, a shift has occurred from chemotherapy for recurrent disease toward the use of adjuvant treatment for newly diagnosed tumors. In addition, there is a better understanding of molecular factors that predict responsiveness to adjuvant chemotherapy.

Chemoradiotherapy with temozolomide is currently the standard of care for newly diagnosed glioblastoma. Two studies have showed that chemotherapy is more effective than radiotherapy alone in elderly glioblastoma patients with a methylated MGMT promoter. Furthermore, it has been shown that bevacizumab added to chemoradiotherapy in newly diagnosed glioblastoma does not improve survival. Treatment with lomustine or retreatment with temozolomide (in patients with a longer temozolomide-free interval) can be considered in recurrent glioblastoma, but with modest effectivity. Additionally, bevacizumab combined with lomustine may play a role in the recurrent glioblastoma setting and is currently being investigated in an adequately powered Phase III trial (EORTC 26101). Until now other ‘targeted’ therapies have shown poor results.

Adjuvant PCV chemotherapy is now part of standard of care in newly diagnosed WHO grade III oligodendroglial tumors at least in the tumors with combined loss of 1p/19q. Adjuvant PCV also clearly improves survival in newly diagnosed high-risk low-grade glioma.

For recurrent WHO grade II and grade III tumors, chemotherapy is the standard and commonly temozolomide is used instead of lomustine or PCV due to better tolerance. Although about half of the patients show an objective response to this treatment, this response however is usually of limited duration, with the exception of oligodendroglial tumors with combined 1p/19q loss.

With the current data available, early chemotherapy is now part of the management of nearly all diffuse glioma.

Ongoing studies investigate whether chemoradiotherapy with concomitant and adjuvant temozolomide also improves outcome in elderly patients with glioblastoma treated with a short schedule of radiotherapy. Furthermore, the value of combined chemoradiotherapy with temozolomide for grade III astrocytoma without 1p/19q loss is currently being investigated, but it will take however many years before the results of this study become available. Clearly, low-grade glioma patients benefit from adjuvant chemotherapy, but the optimal way of identifying these patients (1p/19q loss? IDH mutated? CIMP or MGMT promoter methylated tumors?) remains to be established. Also, whether in large low-grade gliomas, initial treatment with chemotherapy alone can be considered is at present unclear, but the prevailing data suggest that this approach may decrease survival.

Current treatment results in the recurrent glioma are of rather modest effectivity, and it is recommended that these patients are enrolled in clinical trials aiming at the improvement of outcome. Although patient selection can still be improved, it is unlikely that with the currently available cytotoxic drugs outcome can be further enhanced. This implies that for further improvement of outcome new drugs are needed. This should be a high priority.