P14.81 Importance of early spectral variations during a longitudinal 60-month follow-up MRI and 1H-MRS in 135 treated glioblastoma patients

Abstract BACKGROUND To better understand glioblastomas tumor metabolism and metabolic responses to chemotherapy, radiotherapy and antiangiogenic treatments during a longitudinal 60-month follow-up MRI and 1H-MRS in 135 treated glioblastoma patients. MATERIAL AND METHODS 135 patients all biopsied and treated by surgery and STUPP protocol underwent MRI (Sagittal T1, FLAIR, T2*, Diffusion, Perfusion, CoronalT2 and 3DT1 post-Gadolinium) and 1H-MRS (PRESS with multiple TEs: 35ms, 144ms and 288ms on a 6 to 12cm3 volume) exams on 1.5T and 3T GEMS. 92/135 patients underwent resection. MRS data were processed under jMRUI yielding amplitudes, areas under curves and ratios of tCho/tCr, tNAA/tCr, Lip+Lact/tCr, Lact/tCr, Glc/tCr, Glx/tCr and Gln/tCr. Statistical analysis of longitudinal MRI perfusion and spectroscopic data (every 2–3 months above 30 months and 6–9 months over). RESULTS Spectroscopic profiles and ratios improve under STUPP protocol and then worsen depending on the percentage of resection. Biopsied patients: a progressive decrease in ratios of tCho/tCr (0–24 months), Lact/tCr (0–18 months) and Glc/tCr (0–18 months) and Glx/tCr (0–12 months) and then increase at respectively 24, 18, 28 and18 months. Resected patients: a progressive decrease in ratios of tCho/tCr (0–9 months), Lact/tCr (0–15 months) and Glc/tCr (0–15 months) and Glx/tCr (0 and 18 months) and then increase at respectively 32, 20, 18 and 24 months. MRI Gadolinium enhancing tumor and necrosis volumes, Diffusion and FLAIR volumes changed less compared to spectroscopic profiles and ratios which do change more. CONCLUSION 1H-MRS allows non-invasive follow-up of treated glioblastomas tumors. Despite inter-subjects variability, spectroscopic and metabolic changes often come well before clinical deterioration and sometimes before clinical improvement. Therefore, 1H-MRS could be more sensitive and could detect changes earlier than MRI and sometimes is predictive of survival and response to treatment. The analysis of spectral profiles of long survivors is interesting and could help to better understand tumor metabolism and therapeutic responses.