Prof Richard Grundy Research

Advanced Magnetic Resonance Imaging and Metabolic Studies of Low-Grade Gliomas of Childhood

A collaborative project between:
Children’s Brain Tumour Research Centre, Queen’s Medical Centre, Nottingham
Birmingham Children’s Hospital and Institute of Child Health
Royal Marsden Hospital and Institute of Cancer Research, Surrey.

Principal Investigator: Prof.Richard Grundy, Professor of Paediatric Neuro-Oncology and Cancer Biology, Honorary Consultant in Paediatric Oncology, University of Nottingham.

This project aims to develop a serial non-invasive method for the evaluation and prognostication in low grade gliomas of childhood based on Magnetic Resonance Spectroscopy.

Total Grant Funding Required: £117,789 over 3 years from October 2007.

Summary of Research in Layman’s Terms

The biology and natural history of low grade (pilocytic) glioma affecting the optic pathway is of particular interest.  Most commonly presenting in the first 5 years of life these tumours frequently cause considerable morbidity, particularly visual impairment and endocrine damage.

The natural history and clinical evolution of this disease is very difficult to predict.  Overall tumour growth is slow, but may be punctuated by episodes of more obvious tumour growth. Frequently there is a discrepancy between MRI appearances and changes in important clinical factors such a visual function. The decision over when to start therapy, or indeed to stop is difficult in the absence of any objective evidence of the clinical behaviour of the tumour. For the majority of patients there are currently few clinical predictive characteristics.  It would be of particular value if Magnetic Resonance Spectroscopy could identify biological characteristics which predicted quiescence or progression. In cases where chemotherapy or radiation therapy is used it would also be of value if we could detect the biological significance of residual imageable disease during and at the end of such treatment.  It would also be useful if FI could be used as a method of surveillance predicting changes in tumour behaviour. We aim to study these important questions in this study.

Background of proposed research

Low grade gliomas (LGG) are a group of pathologically diverse primary CNS tumours which present a unique challenge amongst childhood brain tumours. Clinical management decisions are made and refined on multiple occasions to take account of a complex set of clinical, imaging and pathological characteristics. Currently there are no clinical or neuroimaging features that predict clinical behaviour and outcome.

In some tumour subtypes, certain histological features, for example an assessment of the tumour cell proliferation index by immunohistochemistry for Ki67/MIB1 in optic nerve pilocytic astrocytomas, have been found useful in the assessment of the likelihood of tumour behaviour and recurrence. Much progress has been made in understanding other tumour groups by employing sophisticated in vitro molecular techniques on tumour tissue. However, studying the evolution of disease by these techniques is hampered by the requirement of serial tissue samples. The development of a non-invasive method of investigation which can be repeated on multiple occasions to give biologically and clinically useful information would be an important contribution to the understanding and management of these tumours.

Together low grade gliomas of childhood (WHO grades 1 and 2) represent between 25 and 35% of all childhood CNS tumours. Prognosis, treatment and outcome of these tumours depends on their site of origin and histological subtype. LGG of the optic pathway and of the hypothalamus and  thalamus arguably present the greatest clinical challenges, because they are associated with considerable morbidity and because the clinical evolution of this disease is very difficult to predict. Overall tumour growth is indolent, but may be punctuated by episodes of more obvious tumour growth which may or may not be associated with deterioration in vision or other clinical parameters.  The decision over when to initiate therapy (or indeed to stop) is difficult in the absence of any objective evidence of the clinical behaviour of the tumour.

This dilemma was first recognised by Oxenhandler and Sayers in 1978, with little advance since.  There is a significant association with Neurofibromatosis type I in which case the disease usually has a more benign clinical history and may be more responsive to treatment (Packer, ‘88, Listernick, ‘94). However, for the majority of patients without NF1 there are currently few clinical predictive characteristics. Overall treatment options are currently limited. The extent of surgery for tumours involving the optic chiasm and tracts are limited by the location of this tumour.  Chemotherapy is effective at ‘disease stabilization ‘, at least in the short term but is rarely curative . Radiotherapy is perhaps more effective, but in this predominantly young patient population is associated with considerable extra late morbidity (Pierce, 1990)  Thalamic LGG frequently have a more aggressive clinical behaviour than would be predicted based on the histopathological features.  It would be of particular value if we could identify’ biological/metabolic’ characteristics which indicated the likelihood of quiescence or progression of tumours in this location.

1H Magnetic Resonance Spectroscopy (MRS) is a non-invasive method for obtaining metabolic profiles (Muckherji, 1998). It is closely related to magnetic resonance imaging (MRI) and may be incorporated into routine MRI investigations. Work in adults with brain tumours has shown that MRS can non-invasively determine tumour histology (Pruel, 1996, 1998, Tate, 1998). It has also been shown that MRS can facilitate selection of the optimal biopsy site (Pruel, 1998, Burtscher, 2000) and detect tumour that is not discernable on MRI (Burtscher, 2000). Preliminary studies also suggest that MRS can predict the prognosis of childhood brain tumours (Warren 2000) and their response to treatment (Girard, 1998, Vaidya, 2003). A study of gliomas in children has been carried out by Lazereff et al. (1999). They performed pretreatment and post-treatment MRS in 10 children with biopsied or partially excised gliomas of which 8 were low grade. A decrease in the tumour choline level from pretreatment to post treatment MRS was found to correlate with response. The follow-up ranged between 6 and 40 months.

We are developing  MRS as a tool for investigating childhood brain tumours in 3 major UKCCSG centres  Birmingham Children’s Hospital (BCH), Queen’s Medical centre and Royal Marsden Hospital. At BCH we have carried out over 70 investigations in the past year and 15 patients  have been accrued at QMC in the last 9 months. The status of MRS is similarly advanced at the Royal Marsden Hospital/ Institute of Cancer Research  (RMH/ ICR), Sutton. During this period we have demonstrated that MRS can be reliably incorporated in the MRI scans which form part of the routine clinical care of these patients. Early analysis of this data indicates that MRS can give useful additional information to that from MRI and clinical indices and highlights the need for a more complete study. A web-accessible multi-user database for MRS data of childhood brain tumours is being developed at the University of Birmingham following funding from the Birmingham Children’s Hospital Research Foundation, The Alistair Wainwright and Joseph Foote fund. This database will be accessible real time to all 3 units.

Although it is desirable to develop non-invasive methods for the investigation of low grade gliomas it is important to link these studies with investigations on tumour tissue when it is available. In particular, the method of magic angle spinning MRS (MAS) may be used to give high resolution metabolic profiles on tumour tissue for comparison with the patient MRS studies (Barton 1999). The use of this technique has been shown to give important information not available from the in vivo studies. For example, the relative concentrations of glycerophosphocholine, choline and phosphocholine may be measured in vitro whereas only the concentration of the three metabolites combined may be determined in vivo. The glycerophoshocholine to phosphocholine ratio has been shown to be an important marker of tumour activity in vitro (Florian 1996). Importantly, one study of MAS in childhood brain tumours showed that the concentrations of major metabolites correlate well with those seen by in vivo MRS in the same patient verifying the validity of the method (Tzika 2002).  MAS also has the useful properties that only 10 to 20 mg of tissue are required and the tissue may be used for other studies after the experiment.