HYPOTHESIS AND SAMPLE SIZE The tumor delineated by FDG-PET is significantly different from the delineation achieved by MR T1 contrast weighted images in glioblastoma; expecting a standard error of 12.5 % (a confidence interval of 25%), with a confidence level set at 95%, a sample size of 15 patients would be accrued in the study.

Inclusion Criteria:

1. Post operative patients of documented high grade gliomas (grade III and IV)
2. Patients who have not undergone any previous irradiation to brain.
3. Performance Scale as defined by ECOG PS 0-3

Exclusion Criteria:

1. Patients who are diabetic.
2. Patients who are pregnant.
3. Any documented contrast allergy to the agents used for imaging.
4. Patients who are unable to comprehend or cooperate effectively for treatment planning.