Department of Medical Oncology, Royal Melbourne Hospital, Victoria, Australia.
Tumours of the central nervous system encompass a large variety of cancers, ranging from slow-growing to rapidly progressive. Although comparatively rare in adults, central nervous system malignancies are relatively common in children, adolescents and young adults, resulting in substantial and ongoing morbidity, significant loss of effective life-years and a heavy burden on family and carers. In fact, more children and adults under the age of 40 die from a brain tumour than from any other cancer. As such, their effect on the community is greater than the apparent low incidence would otherwise indicate. This article focuses on adult brain tumours and in particular glioblastoma. Glioblastoma is rare, but is an example of a disease where treatment has been improved through better understanding of its molecular characteristics, as well as through international clinical trials. We will also discuss some challenges in rare tumours where level one evidence for optimal management is unlikely to ever exist.
Brain tumours are a heterogenous group of diseases, with over 100 types and subtypes, benign and malignant. Figure 1 shows the types of malignant brain tumours operated on most commonly at Royal Melbourne and Melbourne Private Hospitals, both busy tertiary referral centres.
In this paper, we focus on adult malignant brain tumours, accounting for 1.5% of all new cancers diagnosed annually in Australia.1 The most recent Australian Institute of Health and Welfare cancer incidence and mortality data indicate that in 2010, 1680 Australians were diagnosed with a malignant brain tumour and in 2011, over 1200 died from the disease, a sobering observation that incidence is closely matched by mortality.2
For the most part, the aetiology and risk factors for brain cancers remain unknown. Only 5% of brain tumours are attributable to rare familial cancer syndromes such as Turcot or Li-Fraumeni Syndromes.3 Ionizing radiation exposure may increase brain tumour risk, more commonly at least 10-15 years after radiation.4 There is no clear association between mobile phone use and brain tumours, despite some contention in the literature.5-7 As is the case for many other cancer types, histological description of brain tumours is now transitioning to molecular characterisation and, importantly, treatment strategies are being modified accordingly. Table 1 describes some of these molecular markers and their significance and utility.
The management of brain tumours has evolved over the last decade with the advent of improved neuro-surgical and radiation techniques, new systemic therapies, increasing numbers of clinical trials and the introduction of multi-disciplinary care. Among the most important has been the addition of temozolomide chemotherapy to radiotherapy following surgery for glioblastoma (GBM).8 However, beyond chemotherapy and promising developments in targeted therapies, there are several other aspects of neuro-oncology that have developed and strengthened in recent years. This exemplifies optimal management of rare tumours.
In the Australian context, collaboration among the relatively small group of clinicians treating brain tumours has been facilitated by the development of several groups. Examples include: the Cooperative Trials Group for Neuro-Oncology (COGNO), established in 2007 to co-ordinate management of neuro-oncology trials and facilitate discussion of potential investigation into more rare central nervous system (CNS) malignancies; the Clinical Oncology Society of Australia Neuro-oncology Group, established in 2000 which among other activities, has developed comprehensive Australian clinical practice guidelines for brain tumour management; and Cancer Council Victoria’s Clinical Network Neuro-Oncology committee, established in 1999, which has produced several patterns of care studies for Victorian patients in collaboration with the Victorian Cancer Registry.9-11 Co-operative groups in Australia are mirrored overseas with North American, European and Asian groups providing education, scientific and clinical development in the field.
At many hospitals, brain tumours are managed in a multidisciplinary context, with regular multidisciplinary team meetings discussing complex cases, as well as multi-disciplinary neuro-oncology clinics.12 Like other tumour streams, a care co-ordinator is an essential focal point for clinicians and patients.
The traditional standards of care for many brain tumours have been largely dictated by the histopathology of the tumour, coupled with the clinical context (age and performance status). This is beginning to change, albeit slowly although in some contexts, without robust prospective evidence to guide us. Management of several brain tumours is discussed below.
Although rare, glioblastoma (GBM) is the most common CNS malignancy and data from a number of randomised clinical trials is available. The EORTC-NCIC trial reported by Stupp et al, published in 2005, remains the ‘gold standard’ management of patients with GBM under 70 years.8 Several trials have attempted without success to add additional medications in the de novo setting – bevacizumab, cediranib and cilengitide.13-16 In Australia, there is no standard of care for patients with recurrent GBM. Patients often receive single agent carboplatin or lomustine. Enrolment on a clinical trial is appropriate. Bevacizumab is approved by the US Food and Drug Administration, but is not Pharmaceutical Benefits Advisory Committee approved in Australia and is only available on an access program.
The most pressing clinical issues in the management of GBM include: the treatment of de novo disease in patients aged over 70 years or those with poorer performance status; the management of recurrent GBM; and the management of patients who have a non-methylated O6-Methylguanine-DNA methyltransferase (MGMT), a DNA repair enzyme. This group of patients has a poorer prognosis and is less likely to respond to current therapies including temozolomide.17
Grade 3 glioma
This term encompasses pure astrocytomas, oligodendrogliomas and mixed oligoastrocytomas. Until recently, these tumours were grouped as one and standard management was surgery, radiotherapy, then temozolomide chemotherapy at disease progression. However, the prognostic and predictive value of 1p/19q co-deletion – seen in up to 70% of patients with an oligodendroglial component,18 − is now recognised and two randomised trials reported striking overall survival improvements for those 1p/19q co-deleted patients who received PCV chemotherapy in addition to radiotherapy. In one study, the benefit was 14.7 years versus 7.3 years in and in the other study: 123 versus 23 months.19,20 Early post-radiotherapy chemotherapy is now considered routine for patients with 1p/19q codeletions. Isocitrate dehydrogenase (IDH) mutations may also confer benefit from chemotherapy compared with non-mutated tumours.21 Of interest, these two studies evaluated PCV (procarbazine, lomustine and vincristine) chemotherapy, an old fashioned and complex regimen. Many believe that temozolomide is likely to provide equivalent results to PCV with less toxicity.22
Low grade glioma
Low grade glioma (LGG) is much less common than GBM in adults, and tends to affect young adults. After surgery, management options include watching and waiting or up-front radiotherapy, with chemotherapy traditionally reserved until progression. However, updated data from RTOG 9802, comparing radiotherapy + PCV chemotherapy, versus radiotherapy alone, reported a striking overall survival benefit (13.3 versus 7.8 years, HR 0.59, p=0.03) with the combination arm.23 The data are not yet available for 1p/19q co-deletion status, but given that 42% of participants had oligodendroglioma and 32% were mixed gliomas, it is assumed that much of the benefit is driven by the co-deleted tumours responding to chemotherapy. The trial also does not tell us whether co-deleted patients would do just as well with up-front chemotherapy, reserving radiotherapy for later progression. The obvious advantage of such a strategy is that the potential neurocognitive sequelae of radiotherapy, especially in a young patient population, would be delayed. As such, we favour early chemotherapy in those with 1p/19q co-deletions. In Australia, temozolomide is restricted to recurrent grade 3 or grade 4 tumours and is not routinely available for LGG.
Rare brain tumours
Medulloblastoma is a rare cancer representing only 5% of all adult CNS malignancies.24,25 Much of the literature on adult medulloblastoma is in the form of case reports and small cohort studies. Management has largely been extrapolated and modified from the paediatric population, with no accepted standard of care world-wide. A recently published international patterns of care survey, mainly from Australia, reported that cranio-spinal irradiation was common, as was post-radiotherapy chemotherapy, but the regimens varied considerably – up to 10 different regimens were described – reflecting the uncertainty as to optimal management of this disease in adults.26
Ependymoma, a rare tumour involving the spinal cord more frequently than brain, represents only around 3% of CNS malignancies.27 While surgery with or without radiotherapy is the predominant management strategy for ependymoma, chemotherapy is considered for recurrent disease. However, to date, chemotherapy has not been shown to improve outcome in this disease, and clearly improvements are needed.
Despite their rarity, there are many successful clinical trials for CNS malignancies. Further, Australian centres have contributed significantly. These range from large international phase III studies to local phase I studies. A number of Australian sites are participating in early phase studies in which experimental drugs are attempting to inhibit molecular targets such as EGFR, EGFRv3, FGFR, PI3 Kinase and others. COGNO has conducted a number of Australian studies, including the recently completed CABARET study that recruited over 120 patients in 12 months across Australia.28
For rarer tumours such as ependymoma and medulloblastoma, the difficulty lies in the fact that it is unlikely that large scale randomised studies will ever be conducted, and trials, where available, would ideally need to be multi-centre and multi-national studies. To this end, the US-based Collaborative Ependymoma Research Network foundation has established two clinical trials for chemotherapy and targeted therapy in adults with ependymoma.29 It may be possible in the future for Australian centres to collaborate with overseas foundations in order to involve Australian patients in these research efforts.
A major challenge for neuro-oncology is applying limited objective evidence to routine clinical practice. Clinical information and direction often comes from retrospective post hoc subgroup analyses from clinical trials while at the same time, novel biomarkers come to light. Do we accept the limitations of retrospective review and change our practice, or should we await prospective confirmation? In some cases for example, there appears to be compelling benefit of chemotherapy for 1p/19q co-deleted oligodendrogliomas. In other diseases, such as management of medulloblastoma or ependymoma, we need to accept that there will never be robust randomised phase III trial data to support our management decisions.
Another challenge is the potential danger of ‘leaping ahead’ for patients with incurable aggressive tumours such as GBM, and attempting to incorporate drugs into clinical practice where evidence does not yet exist. It is understandable that even clinicians may clutch at straws if there is a small chance of benefit when faced with a patient in front of us. This is exemplified by the use of bevacizumab in GBM. After favourable single arm studies and non-comparative randomised phase II trials in the recurrent disease setting indicated benefit, at least anecdotally, many clinicians in the US began using the drug in de novo GBM. However, the subsequent AvaGlio and RTOG 0825 studies did not show an overall survival benefit when using bevacizumab in this context.13,14
Prognostic and predictive biomarkers should play a stronger role in the future to tailor and guide clinical practice, but in some ways they are still in their infancy. 1p/19q status is now routinely used to guide treatment decisions in grade 2 and 3 gliomas, whereas IDH mutations and MGMT methylation status have yet to reach a tipping point of guiding decisions.
The field of neuro-oncology has many of the same issues and hindrances as that of other rare tumours. First, rare tumours such as medulloblastoma are more common in the paediatric patient population, but the management of these patients cannot be extrapolated to the adult population due to fundamental differences in tumour biology and tolerability of therapies.30 Secondly, tumours must undergo expert neuropathology review. Indeed a review of ependymoma patients demonstrated that 14.6% were reclassified on expert neuropathology review.31 Third, there has been a reluctance to allow neuro-oncology patients access to generic phase I studies.
Finally, within Australia, access to drugs is not the same as in the US. Thus, evidence and recommendations from US studies and organisations such as the National Comprehensive Cancer Network Clinical Practice Guidelines may not always be relevant in the Australian context. This may be frustrating, but is occasionally overcome by compassionate drug access (e.g. via a hospital or a pharmaceutical company) if available evidence is deemed to warrant it for individual circumstances.
Brain tumours are rare in adults, but significant progress has occurred in recent years, changing the face of neuro-oncology in Australia and worldwide. The ongoing challenges are not simply because these tumours are rare, but also resistant to many therapies and in most cases incurable. However, continuing discoveries and clinical trials, as well as substantial work in collaboration and networking, will continue to facilitate progress in all aspects of CNS oncology, from diagnosis through to management and supportive care.