A prospective PET study of patients with glioblastoma multiforme Introduction The long-term prognosis for patients with gliobla- stoma multiforme (GBM) is poor (1). However, the use of intensive multimodality treatment, using surgical resection followed by adjuvant radio- and chemotherapy, may have significant impact on survival (2). In most clinical trials, the response of the tumor to adjuvant treatment is measured by monitoring tumor volume, using CT or MRI with and without contrast. Inability to distinguish between viable tumor cell mass, necrosis and scar tissue (3, 4), often impede the assessment of tumor response by structural criteria alone. Furthermore, concomit- ant use of steroids may influence the morphologi- cal appearance of the tumor (5, 6). Additional information of tumor response can be obtained by indirect measurements of tissue function, using positron emission tomography (PET) with 18 fluoro-deoxyglucose (FDG). The potential of this technique has been used to assess brain tumor malignancy (7), differentiate tumor recurrence from necrosis (8–10) and as a prognostic indicator (11). FDG uptake in tumor has been shown to correlate with cell density (12) and malignancy (7). It has also been shown that FDG-PET may be used as an indicator of prognosis in patients with primary brain tumors (11, 13), although evaluation of treatment out- come is still not reliable (14–16). Evaluating acute tumor response 24 h after chemotherapy with FDG-PET, has shown opposing results, since post-treatment increased glucose metabolism has Acta Neurol Scand 2006: 113: 412–418 DOI: 10.1111/j.1600-0404.2006.00628.x Copyright Ó Blackwell Munksgaard 2006 ACTA NEUROLOGICA SCANDINAVICA Andersen PB, Blinkenberg M, Lassen U, Kosteljanetz M, Wagner A, Poulsen HS, Sørensen PS, Paulson OB. A prospective PET study of patients with glioblastoma multiforme. Acta Neurol Scand 2006: 113: 412–418. Ó Blackwell Munksgaard 2006. Objective – To study the post-surgical metabolic and structural cerebral changes in patients with glioblastoma multiforme (GBM). Materials and methods – We examined ten patients prospectively with newly diagnosed GBM. All patients were primarily treated with surgery, followed by chemotherapy (carmustine, cisplatine and etoposide) and radiotherapy. Positron emission tomography (PET) was used to measure tumor- and cerebral metabolism. CT or MRI was used to estimate tumor volume by measurements of tumor area. Results – Tumor metabolism was not increased during chemotherapy (P ¼ 0.71), but increased during radiotherapy (P ¼ 0.01). CT/MRI showed similar results with no increase in tumor area during chemotherapy (P ¼ 0.33) but increase during radiotherapy (P ¼ 0.002). During the entire study, tumor metabolism and area increased evenly (P ¼ 0.01). Conclusions – Our study did not show a gain of PET compared with structural imaging in the prospective evaluation of GBM. We found a difference in metabolic increase and tumor growth between the two treatment regimens, although this finding has limited relevance due to the design of the study. P. B. Andersen 1,3,6 , M. Blinkenberg 1,3 , U. Lassen 4 , M. Kosteljanetz 2 , A. Wagner 5 , H. S. Poulsen 4 , P. S. Sørensen 1 , O. B. Paulson 1,3 1 Departments of Neurology and 2 Neurosurgery, 3 Neurobiological Research Unit, The Neuroscience Centre, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark; 4 Department of Oncology, Section for Neuro-Oncology, The Finsen Centre, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark; 5 Department of Neuroradiology; 6 The PET and Cyclotrone Unit, Centre of Diagnostic Investigations, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark Key words: brain tumor; chemotherapy; glioblastoma multiforme; neurooncology; PET, positron emission tomography; prospective study; radiotherapy Morten Blinkenberg, Department of Neurology N 2082, Copenhagen University Hospital, Rigshospitalet, 9 Ble- gdamsvej, DK-2100, Copenhagen, Denmark Tel.: +45 3545 2082 Fax: +45 3545 2626 e-mail: blink@dadlnet.dk The authors Preben B. Andersen and Morten Blinken- berg contributed equally to this work. Accepted for publication January 1, 2006 412