Localisation of motor areas in brain tumour patients: a comparison of preoperative [F-18]FDG-PET and intraoperative cortical electrostimulation

Assessment of the exact spatial relation between tumour and adjacent functi onally relevant brain areas is a primary tool in the presurgical planning i n brain tumour patients. The purpose of this study was to compare a preoper ative fluorine-18 fluorodeoxyglucose positron emission tomography ([F-18]FD G PET) activation protocol in patients with tumours near the central area w ith the results of intraoperative direct cortical electrostimulation, and t o determine whether non-invasive preoperative PET imaging can provide resul ts equivalent to those achieved with the invasive neurosurgical "gold stand ard". In this prospective study, we examined 20 patients with various tumou rs of the central area, performing two PET scans (each 30 min after Lv. inj ection of 134-341 MBq [F-18]FDG) in each patient: (1) a resting baseline sc an and (2) an activation scan using a standardised motor task (finger tappi ng, foot stretching). Following PET/MRI realignment and normalisation to th e whole brain counts, parametric images of the activation versus the rest s tudy were calculated and pixels above categorical threshold values were pro jected to the individual MRI for bimodal assessment of morphology and funct ion (PET/MRI overlay). Intraoperative direct cortical electrostimulation wa s performed using a Viking IV probe (5 pulses, each of 100 mus) and documen ted using a dedicated neuro navigation system. Results were compared with t he preoperative PET findings. PET revealed significant activation of the co ntralateral primary motor cortex in 95% (19/20) of the brain tumour patient s (hand activation 13/13, foot activation 6/7), showing a mean increase in normalised [F-18]FDG uptake of 20.5%+/-5.2% (hand activation task) and 17.2 %+/-2.5% (foot activation task). Additionally detected activation of the ip silateral primary motor cortex was interpreted as a metabolic indication fo r interhemispheric compensational processes. Evaluation of the PET findings by cortical stimulation yielded a 94% sensitivity and a 95% specificity fo r identification of motor-associated brain areas. In conclusion, the findin gs indicate that a relatively simple and clinically available [F-18]FDG PET activation protocol enables a sufficiently precise assessment of the local relation between the intracranial tumour and the adjacent motor cortex are as and may facilitate the presurgical planning of tumour resection.