Treatment of glioblastoma in elderly patients

SUMMARY: 

Although glioblastoma occurs mostly in elderly patients, there is a paucity of trials addressing patients older than 70 years of age. Age, by itself, constitutes an unfavorable prognostic factor, which is probably due to unpropitious genetic features, but also due to iatrogenic defeatism. However, many retrospective studies report a survival benefit achieved by aggressive surgical resection seeking gross total removal of contrast-enhancing tumor according to preoperative MRI. Combined radiochemotherapy with concomitant and adjuvant temozolomide has not been investigated in prospective trials. Numerous retrospective studies and a meta-analysis suggest benefit from combined treatment. Prospective randomized trials only evaluated either temozolomide or radiotherapy. Single-treatment hypofractionated radiotherapy performed superior to conventional fractionation. In patients with methylated MGMT promoter, first-line dose-dense temozolomide facilitates prolonged survival. However, there is no comparison with combined radiochemotherapy as the standard-of-care in adult patients. Comorbidity is more frequent in elderly patients, but does not correlate with preterm termination of temozolomide treatment. This review article compiles data proposing a straightforward glioblastoma treatment, irrespective of age.

PRACTICE POINTS.

• In elderly patients, the treatment strategy should strive for surgical gross total tumor resection.

• If a gross total resection is not feasible, diagnosis should be confirmed by stereotactic biopsy.

• Elderly patients may profit from concomitant and adjuvant temozolomide radiochemotherapy.

• Prospective randomized trials have shown that dose-dense temozolomide is superior to radiotherapy only in glioblastomas with methylated MGMT promoter.

• When temozolomide is not practicable, hypofractionated radiotherapy should be preferred to conventional fractionation.

• Comorbidity is more frequent in elderly patients, limiting temozolomide initiation, but temozolomide toxicity is comparable with adult patients.

In adult patients, glioblastoma is the most frequent malignant primary brain tumor, with a peak in the 1970s [1]. Every third patient with newly diagnosed glioblastoma appears to be older than 65 years of age [2]. According to the aging population, we will increasingly deal with glioblastomas in patients of advanced age. However, elderly patients are frequently excluded from clinical trials. But, any cut-off age to consider patients ‘elderly’ lacks any rationale. The age of 70 years is arbitrary and not based on any biological events [3]. Thus, considering a patient ‘elderly’ is inconstant and assigned to the ages of 50, 55, 60, 65, and, most often, 70 years. The therapy rationale in elderly patients often differs from younger fellows (subsequently called adults) with this diagnosis. Glioblastoma patients of an older age might have a distinct accumulation of genetic changes, naturally reduced life expectancy and less resistance to the tumor or its therapy, causing iatrogenic restriction of treatment intensity [4,5]. The latter constitutes a severe bias in retrospective series. In this review, data will be compiled that pre-empt the need for a stigma of the elderly in terms of decision-making.

Prognostic factors

Age by itself constitutes a dismal prognostic factor in glioblastoma patients [5]. This prognostic role remains within the elderly patients [6], with a median overall survival of just 2.5 months in patients over the age of 75 years [7] or a hazard ratio of 2.64 for shortened survival, when patients exceed the age of 80 years [8].

In glioblastoma patients, methylation of MGMT constitutes a strong prognostic factor in terms of time-to-progression and survival [9,10]. In elderly patients, the MGMT promoter is methylated in 43–57.5%, depicting about the same ratio of methylated MGMT promoters compared with patients under the age of 70 years. Deletion of chromosome 10 is associated with dismal prognosis [11] and is specifically found in glioblastoma patients over the age of 70 years [12].

Glioblastomas of elderly patients exhibit frequent alteration in genes regulating VEGF [12]. VEGF expression correlates this WHO grade in gliomas, but does not constitute a prognostic factor within glioblastomas [13]. However, VEGF overexpression serves as treatment target by the anti-VEGF antibody bevacizumab and therefore might constitute a positive predictive feature in the future.

Finally, there are genetic markers associated with a favorable prognosis in glioblastoma patients, which are virtually absent in elderly patients. Mutation of IDH 1/2 is associated with secondary glioblastomas and a prolonged progression-free survival [14], but only present in 2% of elderly patients [15]. A H3F3A G34 mutation, which, in contrast to a H3F3A K27 mutation, is associated with the supratentorial glioblastomas, also constituted a favorable marker, typically found in pediatric but not elderly-type glioblastomas [15,16]. Glioma-CpG island methylator phenotype positivity is associated with an overall survival of 150 weeks in contrast to 54 weeks in Glioma-CpG island methylator phenotype-negative glioblastoma patients [17], while the median age of the patients is 36 and 59 years, respectively.

Likewise, the adult patients the Karnofsky performance score (KPS) remains a strong prognostic factor in the elderly patients as well [6,9,18,19]. General condition and age both influence the treatment decision [20,21], leading to the tendency of less treatment in elderly patients. Thus, the dismal prognosis in elderly patients is prone to be a result of a self-fulfilling prophecy due to defeatism of the physician in charge.

Clinical appearance

Symptoms and signs of glioblastoma differ according to age. Elderly patients more often present with confusion, memory loss and personality changes, which possibly cause delay of tumor diagnosis [5,6]. Seizures, as an initial symptom of malignant brain tumors, appear to be rare in elderly patients, which might be due a dominance of primary glioblastomas from 71 to 80 years of age [22]. Seizure is more common in WHO grade II and III gliomas [23], which then progress to secondary glioblastoma, and were more likely to occur from 51 to 60 years of age. According to genetics, secondary glioblastoma are associated with TP53 and icocitratdehydrogenase 1 or 2 mutations. The latter, which is associated with seizure as an initial symptom in low-grade glioma [24], is rare in the elderly, providing an explanation for fewer seizures [15]. Concerning the general condition, elderly patients initially present with a lower KPS compared with adult patients [25].

Surgery

Despite advances in tumor resection, there is no surgical cure for glioblastoma patients, which is due to infiltrative growth of these tumors. However, maximum cytoreductive surgery has been shown to prolong survival in glioblastoma patients [26,27]. For complete resection of the contrast-enhancing tumor, fluorescence-guided resection has proven to be beneficial, being the current gold standard [28,29]. Thus, fluorescence-guided resection often exceeds the contrast-enhancing tumor [30]. However, in some series, elderly patients receive more biopsies than resection [25,31], revealing a strong confounder due to individual treatment aggressiveness in various centers or trials.

To elucidate the role of resection in elderly patients suffering from glioblastoma, a variety of retrospective studies have been conducted. In an analysis of 40 matched pairs of elderly patients receiving either resection or biopsy, the median survival was 5.7 and 4.0 months, respectively [32]. This results in a significant survival benefit due to the more aggressive treatment, but with equal perioperative morbidity. Of note, the relatively short survival time in this study might be due the fact that only 10% of the patients received temozolomide. A median survival of 17.4 and 11.4 months could be achieved in elderly patients receiving gross and subtotal resection, respectively [25]. Although 40% of the patients underwent combined radiochemotherapy, the brisk difference in survival of the resected patients by almost 1 year recalls the obvious strong bias in retrospective studies. Ewelt and coworkers also report a median survival of 13.9 months after gross total resection, 7.0 months after partial resection and 2.2 months following biopsy [7]. A multivariate analysis including age and performance status confirmed the prognostic value of resection, but, considering the extreme survival difference, a treatment confounder must be assumed again. Likewise, survival times of 18, 16 and 6 months according to gross total resection, partial resection and biopsy, respectively, are reported in another study [33]. In all these publications, the median survival of patients receiving a biopsy were only 6 months at best, raising the question of whether biopsy should have been performed at all. However, there are good reasons to overcome the defeatism that omits any diagnostic efforts when patients are not candidates for gross total tumor resection. First, despite typical appearance of glioblastoma, differential diagnosis, such as metastasis, lymphoma or rare entities, should be kept in mind when treatment decisions in elderly patients will be made [5,34]. Second, stereotactic biopsy is a very low-risk procedure, which can easily be performed under local anesthesia [35]. Third, even genetic markers, such as MGMT promoter methylation, can be determined on serial stereotactic specimens [36].

On the other hand, performing a tumor resection provides the opportunity for additional local treatment. In a retrospective matched-pair study of elderly patients receiving intraoperative placement of carmustine wafers, the patients had a longer survival compared with equally treated patients without interstitial chemotherapy (8.7 vs 5.5 months, respectively) [37]. However, the benefit seems to be restricted to only complete resections in patients with a good KPS [38]. As the carmustine wafer implantation harbors an up to fivefold-increased risk for severe complications such as edema, postoperative seizures, cerebrospinal fluid leakage or infections in a dose-dependent manner [39], this treatment must be critically evaluated for the individual patient.

Radiotherapy

Radiotherapy was the former standard therapy in elderly glioblastoma patients and is superior to supportive care [40]. Patients are mostly assigned to conventional radiotherapy with 30 fractions of 1.8–2 Gy. Alternatively, abbreviated protocols have been introduced to meet the needs of patients with expected shortened survival. Due to hypofractionation, the radiation was completed within 2, 3 or 4 weeks, applying 30, 45 and 50 Gy, respectively. Hypofractionation revealed equal survival compared with the conventional 60 Gy in 6 weeks [41–43]. In the Nordic elderly trial, temozolomide with 200 mg/m² was administered as a third arm and performed superiorly to conventional but equally to hypofractionated radiotherapy [43]. Likewise for NOA-08, the scope of the Nordic elderly trial was less aggressive treatment considering the poorer prognosis in elderly patients. Thus, until recurrence, neither concomitant nor sequential treatment was offered. In fact, the treatment was mostly discontinued due to tumor progression rather than toxicity.

Concomittant & adjuvant chemotherapy

The EORTC trial 26981/22981 introducing a combined concomitant and adjuvant administration of temozolomide provided class 1 evidence for survival benefit [44,45]. However, due to the conviction that elderly patients could not be referred to such intensive treatment, patients older than 70 years of age were excluded from the trial. Likewise, resent prospective trials for elderly glioblastoma patients sought less intense and toxic treatment without offering a combined treatment arm [43,46].

Doubt regarding an effect of concomitant and adjuvant temozolomide in elderly patients was raised when recursive partitioning analysis of EORTC 26981/22981 was performed. Those older than 50 years qualify for group IV or V, the latter group showing a nonsignificant benefit from combined treatment. Disregarding the fact that no resection had been performed and that a poor performance status was also a criterion for group V, it was presumed that older age tended to be a predictive factor for combined treatment being less effective [47]. However, a later long-term follow-up subgroup analysis of the same study cohort reports a survival benefit of patients older than 60 years of age treated with combined therapy (Table 1) [45]. These data are in line with retrospective matched-pair data revealing a benefit from combined radiochemotherapy, disregarding age as a restricting factor, with a hazard ratio of 3.1 [19]. Likewise, prospective data of patients of ages between 60 and 75 years reported a prolonged median survival of 16.3 months when treated with radiochemotherapy compared with 11.2months with radiation only [48]. The scope of this study was the safety of fluorescence-guided resection; therefore, none of the patients had biopsies only. Furthermore, patients in altered general condition were excluded. This selection might explain the superior survival data. Vast retrospective data come from 13,932 patients with glioblastoma compare the era before and after the introduction of radiotherapy combined with temozolomide [49]. However, there were no records of the chemotherapy actually administered, so the treatment was assumed according to the era the treatment took place. On the other hand, before temozolomide became standard treatment, nitrosurea was commonly administered in patients with high-grade glioma. However, in elderly patients, temozolomide proved to be superior compared with radiotherapy alone or adjuvant procarbacin, vincristine and lomustine treatment, as shown by a prospective series with consecutively treated patient groups [18].

Table 1. . Controlled studies of elderly patients treated first-line for glioblastoma by combined concomittant and adjuvant radiochemotherapy.

Study (year)	Design	Age restriction (years)	n	Median survival/age of radiotherapy group (years)	Median survival/age of combined treatment group (years)	HR (95% CI)/p-value (log rank) for comparison with adult patients	Ref.
Behm et al. (2013)	Retrospective	>65	90	3.6/71.9	8.7/70.9	3.1 (1.92–4.99)/p < 0.0001	[19]
Subgroup	>70	62	3.2/73.6	7.5/73.0	4.5 (2.42–8.14)/p < 0.0001
Matched pair	>75	24	3.2/77.4	9.2/76.2	9.0 (3.04–26.9)/p < 0.0001
Stupp et al. (2009)	Prospective, randomized, multivariate	>60–70	170	11.8/nr	10.9/nr	1.4 (1.03–2.00)/nr	[45]
Stummer et al. (2011)	Prospective, subgroup, multivariate	≥60–75	130 res	11.2/66.0	16.3/68.0	nr/p = 0.0145	[48]
Johnson and O’Neill (2012)	Retrospective	≥60–69	2334	10.2†/nr	13.0†/nr	nr/p < 0.0001	[49]
	Historical, multivariate	≥70–79	1549	7.9†/nr	9.3†/nr	nr/p < 0.0001
Ewelt et al. (2011)	Retrospective, selected	>65	72	4.5/70.6	15.0/68.5	nr/nr	[7]
Arrigo et al. (2012)	Prospective	>70	129 res	4.4/nr	5.5/nr	nr/nr	[50]
	Subgroup, multivariate		89 bx	5.8/nr	13.3/nr	nr/nr

^†Treatment assumption due to historical split.

bx: Biopsy; HR: Hazard ratio; nr: Not reported; res: Resection.

Other retrospective smaller series report a benefit of combined treatment compared with radiotherapy only in elderly patients. However, these studies restricted inclusion criteria to good general state, excluded patients who only received biopsies and allowed a selection bias by omitting multivariate or matched-pair analysis [21,51]. As a consequence, a meta-analysis of these studies addressing the benefit of combined radiochemotherapy compared with radiotherapy alone proclaims class 2a evidence for clear survival benefit due to combined treatment [52].

Prospective controlled trials dealing solely with elderly patients randomized radiotherapy alone versus dose-dense temozolomide as single first-line treatment (NOA-08 trial) [46]. For the entire study, the temozolomide treatment cohort appeared not to be inferior. Benefit was shown for the subgroup of patients with methylated MGMT promoter receiving temozolomide. However, concerning the actual treatment recommendation in elderly patients, NOA-08 raises questions:

• NOA-08 has had no arm offering combined radiochemotherapy, which is actually the standard therapy in adult patients. Therefore, even in patients with methylated MGMT promoter receiving temozolomide, omitting concomitant radiotherapy could result in undertreatment;

• MGMT data are mostly missing in patients who received stereotactic biopsy, which is comparatively frequent in elderly patients;

• In many retrospective reports, combined treatment revealed a clear benefit without stratification for MGMT status [7,9,19,33,40,48,50,51].

This suggests a high effect of combined treatment, which could be even stronger when restricted to patients with methylated MGMT promoter, anticipating a better effect of combined treatment compared with temozolomide only. This was shown in a large multicenter cohort study, where 233 elderly patients with glioblastoma were treated with radiotherapy, temozolomide or combined treatment [10]. In tumors with methylated MGMT promoter, application of temozolomide was correlated with prolonged survival. Combined treatment in nonmethylated MGMT promoter glioblastoma did not show a significant survival benefit. However, the latter analysis was based on 66 patients only. As the methylation status was not blinded for treatment decision, there could have been a bias by omitting temozolomide in nonmethylated patients [53]. In fact, the survival curves of methylated and nonmethylated MGMT promoters do not cross over [10], leaving the possibility of type II error on the basis of a slight temozolomide effect. A meta-analysis of these trials, offering only temozolomide or radiotherapy as first-line treatment, concludes with a recommendation of temozolomide treatment for patients with methylated MGMT promoter. Although randomized trials are welcome, as they provide high-class evidence, there are caveats due to the trial protocols, as already critically reviewed by Fiorentino et al. [54]:

• The Nordic trial included patients from the age of 60 years, conflicting initial EORTC data;

• NOA-08 also included patients with a very low KPS, providing a serious confounder;

• NOA-08 also includes anaplastic gliomas;

• Temozolomide regimes differed from standard high dose over 5 days to intermediate daily dose, applied one week on/one week off;

• Salvage treatment was different, with crossover designs, not giving a true control treatment.

According to the meta-analysis of first-line single treatment, interestingly, temozolomide performed slightly superiorly in all patients, with a hazard ratio of 0.86 (95% CI: 0.75–1.0). Thus, temozolomide might not be without an effect in nonmethylated MGMT promoter glioblastomas. Besides methylation of the promoter, MGMT activity can also be altered due to loss of heterozygosity of the MGMT gene or the entire deletion of chromosome 10q. The latter is highly associated with glioblastomas in the elderly [12]. As mentioned above, 10q deletion is linked to dismal prognosis [11]; however, when treated with temozolomide, these patients benefit significantly [55]. Interestingly, the favorable impact of 10q deletion and temozolomide treatment was emphasized in elderly patients [55]. This underlines the role of temozolomide in elderly patients. A meta-analysis of studies comparing temozolomide with radiotherapy as a single first-line treatment favors temozolomide, with a borderline improvement of overall survival [56].

The combination of temozolomide with hypofractionated radiotherapy could accomplish both an intense treatment and a concise radiation protocol. This has been prospectively addressed by Muni et al., reporting a survival advantage in poor-performing elderly or dependent patients [57]. Adjuvant temozolomide was added to hypofractionated radiation, reaching a median survival of 9.4 compared with 7.3 months, when only radiation was performed. Currently, EORTC 26062/22061 addresses the issue of combined hypofractionated treatment, enrolling elderly patients only. Preliminary data applying this protocol exist from a series of 31 patients, of which 16 were older than 70 years [58]. Tolerance was good, with unexpectedly low toxicity. Hypofractionatad radiotherapy with 40 Gy in 15 fractions with concomitant temozolomide followed by adjuvant cycles, including 71 patients over the age of 70 years, was reported by Minniti et al. [59]. The median overall and progression-free survival was 12.4 and 6 months, respectively. However, efficiency data of concomitant temozolomide in abbreviated radiotherapy do not exist and might prove temozolomide to only be equal to adjuvant therapy, as retrospective data suggests [60]. Perhaps lowered total concomitant temozolomide dose due to the abbreviated radiotherapy hampers additional efficiency. Other than for hypofractionated radiotherapy, there is no trial planned addressing elderly patients receiving the current standard-treatment protocol.

It was assumed that older patients reveal enhanced toxicity when treated with alkylating agents. In a series of 19 glioblastoma patients older than 65 years of age, severe myelosuppression was seen in 15% [61]. The total toxicity rate was reported at 42%, including inappropriate ADH secretion and fatigue, which is not necessarily related to temozolomide administration. Likewise, only 14% of the elderly patients had to terminate temozolomide due to myelosuppression [19], similar to 12% of the patients after concomitant therapy in the EORTC trial [44]. A series of 43 patients older than 65 years of age receiving concomitant temozolomide reports termination of temozolomide due to myelosuppression in only one patient (2%) [33]. Three further patients stopped temozolomide due to pneumonia or rash, resulting in a toxicity rate of 9%. No patient had to stop temozolomide due to toxicity in a prospective trial omitting radiation and applying temozolomide only on days 1–5 in a 28-day cycle [31]. There was also no additional toxicity in elderly patients reported by Oszvald and coworkers [25]. CTC grade 3 and 4 thrombocytopenia or neutropenia were reported in 12.5 and 5.4% of patients over the age of 65 years, and 11.5and 3.8% over the age of 70 years, respectively [62]. These findings were no different to those of glioblastoma patients under the age of 65 years [62]. These data, compiled in Table 2, report a comparable toxicity with that known in patients younger than 65 years of age [44,63]. Of note, similar toxicity was considered as serious in the trials omitting radiotherapy, questioning the role of temozolomide [56].

Table 2. . Hematologic toxicity in patinets treated by concomittant and adjuvant radiochemotherapy.

Study (year)	Age restriction (years)	n	Neuropenia CTC grade 3/4, % (patients [n])	Thrombopenia CTC grade 3/4, % (patients [n])	Toxicity-related temozolomide termination (%)	Ref.
Fiorentino et al. (2013)	>65	35	0	11.4 (4)	nr	[64]
Behm et al. (2013)	>70	56	1.7 (1)	14.3 (8)	14	[19]
Sijben et al. (2008)	>65	19	15.7 (3)	15.7 (3)	26	[61]
Combs et al. (2008)	>65	43	2.3 (1)	2.3 (1)	11.6	[33]
Oszvald et al. (2012)	>65	58	nr	nr	8.6	[25]
Minniti et al. (2008)	>70	32	12.5 (4)	21.8 (7)	6.5	[65]
Lee et al. (2013)	>70	20	10 (2)	10 (2)	nr	[66]
Balducci et al. (2012)	>65	56	1.7 (1)	3.4 (2)	nr	[62]

CTC: Comon toxicity criteria; nr: Not reported.

Obviously comorbidity appears more likely in elderly than adult patients. This comorbidity by itself is accountable for worse prognosis. However, when comorbidity did not provide contraindication for temozolomide initiation, there was no association with toxicity-related termination of temozolomide [67].

Quality of life

Besides survival and toxicity, the patient’s quality of life and self-dependence must be kept in mind. In EORTC 26981/22981, the rate of patients reporting fatigue was increased from 26 to 33% when radiotherapy was combined with temozolomide [44]. The increase of fatigue also occurred in concomitant temozolomide applying an abbreviated radiotherapy over 3 weeks [59]. Besides fatigue, the latter study also reported improvement in all functioning and symptom scores. However, as only half of the patients completed the questionnaire after 6 months, selection must be assumed. In elderly patients, radiotherapy implies the risk of cognitive decline. Compared with temozolomide treatment, patients suffered from communication deficit [46] and a potential decline of cognitive, physical and role functioning [43]. On the other hand, tumor progression is the major cause for decline in quality of life [68], thus offering the most effective treatment might provide quality of life maintenance as well. Thus, elderly patients must be closely watched and considered for sparing from intense treatment, such as completing radiation, in therapy-related deterioration.

Conclusion & future perspective

In general, patients with glioblastoma should be treated irrespective of age with gross total resection and combined radiochemotherapy. The predictive role of MGMT promoter methylation must be kept in mind when patients reveal risk factors for temozolomide application. The role of first-line bevacizumab treatment is a matter of current trials. As proangiogenic proteins such as VEGF seem to play a paramount role in glioblastoma in elderly patients, bevacizumab is likely to give additional benefit. However, the patient’s exposure to extra risk must be critically evaluated. In the future, clinical trials should omit an upper age exclusion criterion to reveal the age-related reality in this tumor entity.