Gliomas are the most frequently occurring central nervous system (CNS) tumours and include astrocytomas, anaplastic astrocytomas and glioblastoma multiforme. They are severely disabling, producing symptoms of headaches, fits, confusion, personality changes and neurological deficits which impair the patients' quality of life. Despite advances in the diagnosis and treatment of gliomas the prognosis remains poor, with only 10-15% survival for patients with high-grade gliomas at 30+ months (Brada et al, 1998). Nevertheless, considerable variation exists with respect to post-operative survival. Identification of prognostic variables to assist in clinical management is thus of considerable value to both patients and clinicians.Historically, prognostic models for gliomas have used clinical variables and patient characteristics to determine factors that predict response to therapy and survival. In 1990, the Medical Research Council (MRC) working party developed a clinical scoring system that predicted a better prognosis for patients less than 45 years old at diagnosis, with longer duration of fits, debulking surgery and a better performance status (Report of the MRC Brain Tumour Working Party, 1990). Although adverse histological features including high proliferative activity (Bruner, 1994), the presence of necrosis (Barker et al, 1996) and a gemistocyte population of greater than 20% (Krouer et al, 1991) have been identified in univariate analysis of small series, these have not been evaluated in larger series and are not widely employed.The development of gliomas, in common with many tumours, is a multistep process involving the accumulation of several genetic events which involve the activation of oncogenes (Collins, 1995) and the loss of tumour suppressor genes (Jen et al, 1994;Louis and Gusella, 1995;von Deimling et al, 1995). The commonest genetic abnormality observed in gliomas is loss, often involving a whole homologue, of chromosome 10 (Bigner at al, 1990). This loss is associated with high-grade tumours, being rare in low-grade astrocytomas but present in some 60-85% of high-grade gliomas (Louis and Gusella, 1995). Recent studies have led to the identification of at least three tumour suppressor genes, PTEN (MMAC1), DMBT1 and LGI1 (Li et al, 1997;Mollenhauser et al, 1997;Steck et al, 1997;Chernova et al, 1998) on chromosome 10, which may be involved in glioma progression. The aim of this study was to determine whether loss of heterozygosity (LOH) for chromosome 10 is an important prognostic variable in high-grade gliomas and whether it provides additional predictive information to the previously defined clinical prognostic indices.
MATERIALS AND METHODS
PatientsBetween January 1994 and April 1997, blood samples and pathological blocks of tumour material were collected prospectively from 42 patients diagnosed with high-grade glioma. Blood tumour Loss of chromosome 10 is an independent prognostic factor in high-grade gliomas Summary Loss of heterozygosity (LOH) for chromosome 10 is the most frequent genetic abn...
